{
  "best_metric": 0.011049928143620491,
  "best_model_checkpoint": "./experiments/unet_run_2026-06-23_13-16\\checkpoint-12474",
  "epoch": 199.0,
  "eval_steps": 500,
  "global_step": 13134,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.07575757575757576,
      "grad_norm": 3.589477300643921,
      "learning_rate": 9.996212121212122e-05,
      "loss": 1.0979,
      "step": 5
    },
    {
      "epoch": 0.15151515151515152,
      "grad_norm": 5.41869592666626,
      "learning_rate": 9.992424242424244e-05,
      "loss": 0.957,
      "step": 10
    },
    {
      "epoch": 0.22727272727272727,
      "grad_norm": 2.4673285484313965,
      "learning_rate": 9.988636363636364e-05,
      "loss": 0.8332,
      "step": 15
    },
    {
      "epoch": 0.30303030303030304,
      "grad_norm": 1.9845205545425415,
      "learning_rate": 9.984848484848486e-05,
      "loss": 0.7463,
      "step": 20
    },
    {
      "epoch": 0.3787878787878788,
      "grad_norm": 1.967355728149414,
      "learning_rate": 9.981060606060606e-05,
      "loss": 0.6974,
      "step": 25
    },
    {
      "epoch": 0.45454545454545453,
      "grad_norm": 1.7775042057037354,
      "learning_rate": 9.977272727272728e-05,
      "loss": 0.6286,
      "step": 30
    },
    {
      "epoch": 0.5303030303030303,
      "grad_norm": 1.8440767526626587,
      "learning_rate": 9.973484848484848e-05,
      "loss": 0.595,
      "step": 35
    },
    {
      "epoch": 0.6060606060606061,
      "grad_norm": 1.6301528215408325,
      "learning_rate": 9.96969696969697e-05,
      "loss": 0.5452,
      "step": 40
    },
    {
      "epoch": 0.6818181818181818,
      "grad_norm": 1.555722951889038,
      "learning_rate": 9.965909090909091e-05,
      "loss": 0.5028,
      "step": 45
    },
    {
      "epoch": 0.7575757575757576,
      "grad_norm": 1.6868361234664917,
      "learning_rate": 9.962121212121213e-05,
      "loss": 0.478,
      "step": 50
    },
    {
      "epoch": 0.8333333333333334,
      "grad_norm": 1.5898619890213013,
      "learning_rate": 9.958333333333335e-05,
      "loss": 0.4445,
      "step": 55
    },
    {
      "epoch": 0.9090909090909091,
      "grad_norm": 1.579331636428833,
      "learning_rate": 9.954545454545455e-05,
      "loss": 0.4009,
      "step": 60
    },
    {
      "epoch": 0.9848484848484849,
      "grad_norm": 1.4321759939193726,
      "learning_rate": 9.950757575757577e-05,
      "loss": 0.377,
      "step": 65
    },
    {
      "epoch": 1.0,
      "eval_loss": 0.34798720479011536,
      "eval_mean_accuracy": 0.8179425734629177,
      "eval_mean_iou": 0.769930232898238,
      "eval_runtime": 44.9889,
      "eval_samples_per_second": 5.224,
      "eval_steps_per_second": 0.667,
      "step": 66
    },
    {
      "epoch": 1.0606060606060606,
      "grad_norm": 1.4558287858963013,
      "learning_rate": 9.946969696969698e-05,
      "loss": 0.3624,
      "step": 70
    },
    {
      "epoch": 1.1363636363636362,
      "grad_norm": 1.3527405261993408,
      "learning_rate": 9.943181818181819e-05,
      "loss": 0.3497,
      "step": 75
    },
    {
      "epoch": 1.2121212121212122,
      "grad_norm": 1.296321153640747,
      "learning_rate": 9.939393939393939e-05,
      "loss": 0.32,
      "step": 80
    },
    {
      "epoch": 1.2878787878787878,
      "grad_norm": 1.29216730594635,
      "learning_rate": 9.93560606060606e-05,
      "loss": 0.3077,
      "step": 85
    },
    {
      "epoch": 1.3636363636363638,
      "grad_norm": 1.2346165180206299,
      "learning_rate": 9.931818181818182e-05,
      "loss": 0.2887,
      "step": 90
    },
    {
      "epoch": 1.4393939393939394,
      "grad_norm": 1.6054928302764893,
      "learning_rate": 9.928030303030304e-05,
      "loss": 0.2893,
      "step": 95
    },
    {
      "epoch": 1.5151515151515151,
      "grad_norm": 1.1277427673339844,
      "learning_rate": 9.924242424242425e-05,
      "loss": 0.2787,
      "step": 100
    },
    {
      "epoch": 1.5909090909090908,
      "grad_norm": 1.133207082748413,
      "learning_rate": 9.920454545454546e-05,
      "loss": 0.2675,
      "step": 105
    },
    {
      "epoch": 1.6666666666666665,
      "grad_norm": 1.0519332885742188,
      "learning_rate": 9.916666666666667e-05,
      "loss": 0.2542,
      "step": 110
    },
    {
      "epoch": 1.7424242424242424,
      "grad_norm": 1.0028828382492065,
      "learning_rate": 9.912878787878789e-05,
      "loss": 0.2291,
      "step": 115
    },
    {
      "epoch": 1.8181818181818183,
      "grad_norm": 0.9600074291229248,
      "learning_rate": 9.909090909090911e-05,
      "loss": 0.2242,
      "step": 120
    },
    {
      "epoch": 1.893939393939394,
      "grad_norm": 1.0376956462860107,
      "learning_rate": 9.90530303030303e-05,
      "loss": 0.2283,
      "step": 125
    },
    {
      "epoch": 1.9696969696969697,
      "grad_norm": 0.9240972995758057,
      "learning_rate": 9.901515151515151e-05,
      "loss": 0.2173,
      "step": 130
    },
    {
      "epoch": 2.0,
      "eval_loss": 0.2091335505247116,
      "eval_mean_accuracy": 0.8132426533657254,
      "eval_mean_iou": 0.781493649673832,
      "eval_runtime": 34.3874,
      "eval_samples_per_second": 6.834,
      "eval_steps_per_second": 0.872,
      "step": 132
    },
    {
      "epoch": 2.0454545454545454,
      "grad_norm": 0.847649335861206,
      "learning_rate": 9.897727272727273e-05,
      "loss": 0.2048,
      "step": 135
    },
    {
      "epoch": 2.121212121212121,
      "grad_norm": 0.9049073457717896,
      "learning_rate": 9.893939393939395e-05,
      "loss": 0.1893,
      "step": 140
    },
    {
      "epoch": 2.196969696969697,
      "grad_norm": 0.7354316711425781,
      "learning_rate": 9.890151515151516e-05,
      "loss": 0.1856,
      "step": 145
    },
    {
      "epoch": 2.2727272727272725,
      "grad_norm": 0.7612099051475525,
      "learning_rate": 9.886363636363637e-05,
      "loss": 0.1837,
      "step": 150
    },
    {
      "epoch": 2.3484848484848486,
      "grad_norm": 0.8348954319953918,
      "learning_rate": 9.882575757575758e-05,
      "loss": 0.1694,
      "step": 155
    },
    {
      "epoch": 2.4242424242424243,
      "grad_norm": 0.8095713257789612,
      "learning_rate": 9.87878787878788e-05,
      "loss": 0.1699,
      "step": 160
    },
    {
      "epoch": 2.5,
      "grad_norm": 0.7507986426353455,
      "learning_rate": 9.875000000000002e-05,
      "loss": 0.1483,
      "step": 165
    },
    {
      "epoch": 2.5757575757575757,
      "grad_norm": 0.7362430691719055,
      "learning_rate": 9.871212121212122e-05,
      "loss": 0.1493,
      "step": 170
    },
    {
      "epoch": 2.6515151515151514,
      "grad_norm": 0.6315919160842896,
      "learning_rate": 9.867424242424242e-05,
      "loss": 0.1423,
      "step": 175
    },
    {
      "epoch": 2.7272727272727275,
      "grad_norm": 0.7730288505554199,
      "learning_rate": 9.863636363636364e-05,
      "loss": 0.1502,
      "step": 180
    },
    {
      "epoch": 2.8030303030303028,
      "grad_norm": 0.6501778960227966,
      "learning_rate": 9.859848484848485e-05,
      "loss": 0.1355,
      "step": 185
    },
    {
      "epoch": 2.878787878787879,
      "grad_norm": 0.5899284482002258,
      "learning_rate": 9.856060606060607e-05,
      "loss": 0.1364,
      "step": 190
    },
    {
      "epoch": 2.9545454545454546,
      "grad_norm": 0.6189233064651489,
      "learning_rate": 9.852272727272727e-05,
      "loss": 0.136,
      "step": 195
    },
    {
      "epoch": 3.0,
      "eval_loss": 0.12592849135398865,
      "eval_mean_accuracy": 0.8796883281531912,
      "eval_mean_iou": 0.8487147482918348,
      "eval_runtime": 34.8023,
      "eval_samples_per_second": 6.752,
      "eval_steps_per_second": 0.862,
      "step": 198
    },
    {
      "epoch": 3.0303030303030303,
      "grad_norm": 0.5857253074645996,
      "learning_rate": 9.848484848484849e-05,
      "loss": 0.126,
      "step": 200
    },
    {
      "epoch": 3.106060606060606,
      "grad_norm": 0.657252848148346,
      "learning_rate": 9.84469696969697e-05,
      "loss": 0.1241,
      "step": 205
    },
    {
      "epoch": 3.1818181818181817,
      "grad_norm": 0.5293025970458984,
      "learning_rate": 9.840909090909092e-05,
      "loss": 0.1115,
      "step": 210
    },
    {
      "epoch": 3.257575757575758,
      "grad_norm": 0.5324749946594238,
      "learning_rate": 9.837121212121213e-05,
      "loss": 0.1277,
      "step": 215
    },
    {
      "epoch": 3.3333333333333335,
      "grad_norm": 0.6226522922515869,
      "learning_rate": 9.833333333333333e-05,
      "loss": 0.105,
      "step": 220
    },
    {
      "epoch": 3.409090909090909,
      "grad_norm": 0.5890645384788513,
      "learning_rate": 9.829545454545455e-05,
      "loss": 0.1082,
      "step": 225
    },
    {
      "epoch": 3.484848484848485,
      "grad_norm": 0.5300428867340088,
      "learning_rate": 9.825757575757576e-05,
      "loss": 0.0997,
      "step": 230
    },
    {
      "epoch": 3.5606060606060606,
      "grad_norm": 0.6047218441963196,
      "learning_rate": 9.821969696969698e-05,
      "loss": 0.1047,
      "step": 235
    },
    {
      "epoch": 3.6363636363636362,
      "grad_norm": 0.49045872688293457,
      "learning_rate": 9.818181818181818e-05,
      "loss": 0.0953,
      "step": 240
    },
    {
      "epoch": 3.712121212121212,
      "grad_norm": 0.4327333867549896,
      "learning_rate": 9.81439393939394e-05,
      "loss": 0.0982,
      "step": 245
    },
    {
      "epoch": 3.787878787878788,
      "grad_norm": 0.4083159565925598,
      "learning_rate": 9.810606060606061e-05,
      "loss": 0.0929,
      "step": 250
    },
    {
      "epoch": 3.8636363636363638,
      "grad_norm": 0.5582062602043152,
      "learning_rate": 9.806818181818183e-05,
      "loss": 0.1002,
      "step": 255
    },
    {
      "epoch": 3.9393939393939394,
      "grad_norm": 0.4387126863002777,
      "learning_rate": 9.803030303030303e-05,
      "loss": 0.0867,
      "step": 260
    },
    {
      "epoch": 4.0,
      "eval_loss": 0.08807452768087387,
      "eval_mean_accuracy": 0.9002409391883904,
      "eval_mean_iou": 0.8723302409378504,
      "eval_runtime": 34.4927,
      "eval_samples_per_second": 6.813,
      "eval_steps_per_second": 0.87,
      "step": 264
    },
    {
      "epoch": 4.015151515151516,
      "grad_norm": 0.4800288677215576,
      "learning_rate": 9.799242424242425e-05,
      "loss": 0.089,
      "step": 265
    },
    {
      "epoch": 4.090909090909091,
      "grad_norm": 0.4667986035346985,
      "learning_rate": 9.795454545454545e-05,
      "loss": 0.0906,
      "step": 270
    },
    {
      "epoch": 4.166666666666667,
      "grad_norm": 0.3979324400424957,
      "learning_rate": 9.791666666666667e-05,
      "loss": 0.0856,
      "step": 275
    },
    {
      "epoch": 4.242424242424242,
      "grad_norm": 0.43097439408302307,
      "learning_rate": 9.787878787878789e-05,
      "loss": 0.0793,
      "step": 280
    },
    {
      "epoch": 4.318181818181818,
      "grad_norm": 0.47306445240974426,
      "learning_rate": 9.784090909090909e-05,
      "loss": 0.0776,
      "step": 285
    },
    {
      "epoch": 4.393939393939394,
      "grad_norm": 0.48418354988098145,
      "learning_rate": 9.78030303030303e-05,
      "loss": 0.0814,
      "step": 290
    },
    {
      "epoch": 4.46969696969697,
      "grad_norm": 0.6252722144126892,
      "learning_rate": 9.776515151515152e-05,
      "loss": 0.0805,
      "step": 295
    },
    {
      "epoch": 4.545454545454545,
      "grad_norm": 0.38845524191856384,
      "learning_rate": 9.772727272727274e-05,
      "loss": 0.0722,
      "step": 300
    },
    {
      "epoch": 4.621212121212121,
      "grad_norm": 0.32852452993392944,
      "learning_rate": 9.768939393939394e-05,
      "loss": 0.0728,
      "step": 305
    },
    {
      "epoch": 4.696969696969697,
      "grad_norm": 0.329818457365036,
      "learning_rate": 9.765151515151516e-05,
      "loss": 0.0699,
      "step": 310
    },
    {
      "epoch": 4.7727272727272725,
      "grad_norm": 0.3442506790161133,
      "learning_rate": 9.761363636363638e-05,
      "loss": 0.0684,
      "step": 315
    },
    {
      "epoch": 4.848484848484849,
      "grad_norm": 0.3667300343513489,
      "learning_rate": 9.757575757575758e-05,
      "loss": 0.0654,
      "step": 320
    },
    {
      "epoch": 4.924242424242424,
      "grad_norm": 0.3521306812763214,
      "learning_rate": 9.75378787878788e-05,
      "loss": 0.0644,
      "step": 325
    },
    {
      "epoch": 5.0,
      "grad_norm": 0.838910698890686,
      "learning_rate": 9.75e-05,
      "loss": 0.0791,
      "step": 330
    },
    {
      "epoch": 5.0,
      "eval_loss": 0.06843745708465576,
      "eval_mean_accuracy": 0.913246816813527,
      "eval_mean_iou": 0.8837213672874938,
      "eval_runtime": 34.9701,
      "eval_samples_per_second": 6.72,
      "eval_steps_per_second": 0.858,
      "step": 330
    },
    {
      "epoch": 5.075757575757576,
      "grad_norm": 0.33520424365997314,
      "learning_rate": 9.746212121212121e-05,
      "loss": 0.067,
      "step": 335
    },
    {
      "epoch": 5.151515151515151,
      "grad_norm": 0.3536280393600464,
      "learning_rate": 9.742424242424243e-05,
      "loss": 0.0658,
      "step": 340
    },
    {
      "epoch": 5.2272727272727275,
      "grad_norm": 0.3065834939479828,
      "learning_rate": 9.738636363636365e-05,
      "loss": 0.072,
      "step": 345
    },
    {
      "epoch": 5.303030303030303,
      "grad_norm": 0.2831350862979889,
      "learning_rate": 9.734848484848485e-05,
      "loss": 0.0614,
      "step": 350
    },
    {
      "epoch": 5.378787878787879,
      "grad_norm": 0.2720297574996948,
      "learning_rate": 9.731060606060607e-05,
      "loss": 0.059,
      "step": 355
    },
    {
      "epoch": 5.454545454545454,
      "grad_norm": 0.9245620369911194,
      "learning_rate": 9.727272727272728e-05,
      "loss": 0.0712,
      "step": 360
    },
    {
      "epoch": 5.53030303030303,
      "grad_norm": 0.36003217101097107,
      "learning_rate": 9.723484848484849e-05,
      "loss": 0.0614,
      "step": 365
    },
    {
      "epoch": 5.606060606060606,
      "grad_norm": 0.26046162843704224,
      "learning_rate": 9.71969696969697e-05,
      "loss": 0.0609,
      "step": 370
    },
    {
      "epoch": 5.681818181818182,
      "grad_norm": 0.2610915005207062,
      "learning_rate": 9.71590909090909e-05,
      "loss": 0.059,
      "step": 375
    },
    {
      "epoch": 5.757575757575758,
      "grad_norm": 0.3020758330821991,
      "learning_rate": 9.712121212121212e-05,
      "loss": 0.0567,
      "step": 380
    },
    {
      "epoch": 5.833333333333333,
      "grad_norm": 0.3303019404411316,
      "learning_rate": 9.708333333333334e-05,
      "loss": 0.0569,
      "step": 385
    },
    {
      "epoch": 5.909090909090909,
      "grad_norm": 0.3087840676307678,
      "learning_rate": 9.704545454545456e-05,
      "loss": 0.054,
      "step": 390
    },
    {
      "epoch": 5.984848484848484,
      "grad_norm": 0.22786583006381989,
      "learning_rate": 9.700757575757576e-05,
      "loss": 0.0522,
      "step": 395
    },
    {
      "epoch": 6.0,
      "eval_loss": 0.05420524254441261,
      "eval_mean_accuracy": 0.936309987345198,
      "eval_mean_iou": 0.8945142049705211,
      "eval_runtime": 35.0077,
      "eval_samples_per_second": 6.713,
      "eval_steps_per_second": 0.857,
      "step": 396
    },
    {
      "epoch": 6.0606060606060606,
      "grad_norm": 0.2567901909351349,
      "learning_rate": 9.696969696969698e-05,
      "loss": 0.0594,
      "step": 400
    },
    {
      "epoch": 6.136363636363637,
      "grad_norm": 0.25553104281425476,
      "learning_rate": 9.693181818181819e-05,
      "loss": 0.0534,
      "step": 405
    },
    {
      "epoch": 6.212121212121212,
      "grad_norm": 0.22771595418453217,
      "learning_rate": 9.689393939393941e-05,
      "loss": 0.0522,
      "step": 410
    },
    {
      "epoch": 6.287878787878788,
      "grad_norm": 0.2904994487762451,
      "learning_rate": 9.685606060606061e-05,
      "loss": 0.0557,
      "step": 415
    },
    {
      "epoch": 6.363636363636363,
      "grad_norm": 0.257795512676239,
      "learning_rate": 9.681818181818181e-05,
      "loss": 0.054,
      "step": 420
    },
    {
      "epoch": 6.4393939393939394,
      "grad_norm": 0.27261966466903687,
      "learning_rate": 9.678030303030303e-05,
      "loss": 0.0514,
      "step": 425
    },
    {
      "epoch": 6.515151515151516,
      "grad_norm": 0.2508927285671234,
      "learning_rate": 9.674242424242425e-05,
      "loss": 0.0463,
      "step": 430
    },
    {
      "epoch": 6.590909090909091,
      "grad_norm": 0.2328452318906784,
      "learning_rate": 9.670454545454546e-05,
      "loss": 0.0498,
      "step": 435
    },
    {
      "epoch": 6.666666666666667,
      "grad_norm": 0.24541780352592468,
      "learning_rate": 9.666666666666667e-05,
      "loss": 0.0472,
      "step": 440
    },
    {
      "epoch": 6.742424242424242,
      "grad_norm": 0.32659071683883667,
      "learning_rate": 9.662878787878788e-05,
      "loss": 0.0464,
      "step": 445
    },
    {
      "epoch": 6.818181818181818,
      "grad_norm": 0.2596575617790222,
      "learning_rate": 9.65909090909091e-05,
      "loss": 0.0452,
      "step": 450
    },
    {
      "epoch": 6.893939393939394,
      "grad_norm": 0.2107088267803192,
      "learning_rate": 9.655303030303032e-05,
      "loss": 0.0447,
      "step": 455
    },
    {
      "epoch": 6.96969696969697,
      "grad_norm": 0.2737276256084442,
      "learning_rate": 9.651515151515152e-05,
      "loss": 0.0447,
      "step": 460
    },
    {
      "epoch": 7.0,
      "eval_loss": 0.04422459006309509,
      "eval_mean_accuracy": 0.940060939542308,
      "eval_mean_iou": 0.8996565914639395,
      "eval_runtime": 34.4939,
      "eval_samples_per_second": 6.813,
      "eval_steps_per_second": 0.87,
      "step": 462
    },
    {
      "epoch": 7.045454545454546,
      "grad_norm": 0.18616357445716858,
      "learning_rate": 9.647727272727272e-05,
      "loss": 0.0466,
      "step": 465
    },
    {
      "epoch": 7.121212121212121,
      "grad_norm": 0.2652761936187744,
      "learning_rate": 9.643939393939394e-05,
      "loss": 0.0406,
      "step": 470
    },
    {
      "epoch": 7.196969696969697,
      "grad_norm": 0.3475441038608551,
      "learning_rate": 9.640151515151516e-05,
      "loss": 0.0413,
      "step": 475
    },
    {
      "epoch": 7.2727272727272725,
      "grad_norm": 0.36833781003952026,
      "learning_rate": 9.636363636363637e-05,
      "loss": 0.0469,
      "step": 480
    },
    {
      "epoch": 7.348484848484849,
      "grad_norm": 0.1868564486503601,
      "learning_rate": 9.632575757575758e-05,
      "loss": 0.0402,
      "step": 485
    },
    {
      "epoch": 7.424242424242424,
      "grad_norm": 0.19365091621875763,
      "learning_rate": 9.628787878787879e-05,
      "loss": 0.0386,
      "step": 490
    },
    {
      "epoch": 7.5,
      "grad_norm": 0.2226143479347229,
      "learning_rate": 9.625000000000001e-05,
      "loss": 0.0421,
      "step": 495
    },
    {
      "epoch": 7.575757575757576,
      "grad_norm": 0.21930691599845886,
      "learning_rate": 9.621212121212123e-05,
      "loss": 0.0386,
      "step": 500
    },
    {
      "epoch": 7.651515151515151,
      "grad_norm": 0.15792429447174072,
      "learning_rate": 9.617424242424243e-05,
      "loss": 0.041,
      "step": 505
    },
    {
      "epoch": 7.7272727272727275,
      "grad_norm": 0.17836716771125793,
      "learning_rate": 9.613636363636363e-05,
      "loss": 0.0406,
      "step": 510
    },
    {
      "epoch": 7.803030303030303,
      "grad_norm": 0.1674388349056244,
      "learning_rate": 9.609848484848485e-05,
      "loss": 0.0407,
      "step": 515
    },
    {
      "epoch": 7.878787878787879,
      "grad_norm": 0.2626052498817444,
      "learning_rate": 9.606060606060606e-05,
      "loss": 0.0447,
      "step": 520
    },
    {
      "epoch": 7.954545454545455,
      "grad_norm": 0.3617821931838989,
      "learning_rate": 9.602272727272728e-05,
      "loss": 0.0399,
      "step": 525
    },
    {
      "epoch": 8.0,
      "eval_loss": 0.03873525187373161,
      "eval_mean_accuracy": 0.9524271009450794,
      "eval_mean_iou": 0.8972008569882305,
      "eval_runtime": 35.6166,
      "eval_samples_per_second": 6.598,
      "eval_steps_per_second": 0.842,
      "step": 528
    },
    {
      "epoch": 8.030303030303031,
      "grad_norm": 0.18350818753242493,
      "learning_rate": 9.598484848484848e-05,
      "loss": 0.0374,
      "step": 530
    },
    {
      "epoch": 8.106060606060606,
      "grad_norm": 0.16380168497562408,
      "learning_rate": 9.59469696969697e-05,
      "loss": 0.0403,
      "step": 535
    },
    {
      "epoch": 8.181818181818182,
      "grad_norm": 0.1669972538948059,
      "learning_rate": 9.590909090909092e-05,
      "loss": 0.0365,
      "step": 540
    },
    {
      "epoch": 8.257575757575758,
      "grad_norm": 0.18097421526908875,
      "learning_rate": 9.587121212121213e-05,
      "loss": 0.0354,
      "step": 545
    },
    {
      "epoch": 8.333333333333334,
      "grad_norm": 0.2094983458518982,
      "learning_rate": 9.583333333333334e-05,
      "loss": 0.039,
      "step": 550
    },
    {
      "epoch": 8.409090909090908,
      "grad_norm": 0.15182794630527496,
      "learning_rate": 9.579545454545455e-05,
      "loss": 0.0413,
      "step": 555
    },
    {
      "epoch": 8.484848484848484,
      "grad_norm": 0.2121264487504959,
      "learning_rate": 9.575757575757576e-05,
      "loss": 0.0359,
      "step": 560
    },
    {
      "epoch": 8.56060606060606,
      "grad_norm": 0.1865435540676117,
      "learning_rate": 9.571969696969697e-05,
      "loss": 0.0402,
      "step": 565
    },
    {
      "epoch": 8.636363636363637,
      "grad_norm": 0.17947863042354584,
      "learning_rate": 9.568181818181819e-05,
      "loss": 0.0341,
      "step": 570
    },
    {
      "epoch": 8.712121212121213,
      "grad_norm": 0.13673025369644165,
      "learning_rate": 9.564393939393939e-05,
      "loss": 0.0341,
      "step": 575
    },
    {
      "epoch": 8.787878787878787,
      "grad_norm": 0.22620035707950592,
      "learning_rate": 9.560606060606061e-05,
      "loss": 0.0335,
      "step": 580
    },
    {
      "epoch": 8.863636363636363,
      "grad_norm": 0.13189756870269775,
      "learning_rate": 9.556818181818182e-05,
      "loss": 0.0352,
      "step": 585
    },
    {
      "epoch": 8.93939393939394,
      "grad_norm": 0.2993201017379761,
      "learning_rate": 9.553030303030304e-05,
      "loss": 0.0366,
      "step": 590
    },
    {
      "epoch": 9.0,
      "eval_loss": 0.03465564176440239,
      "eval_mean_accuracy": 0.9360416851533633,
      "eval_mean_iou": 0.9030759484857729,
      "eval_runtime": 34.8293,
      "eval_samples_per_second": 6.747,
      "eval_steps_per_second": 0.861,
      "step": 594
    },
    {
      "epoch": 9.015151515151516,
      "grad_norm": 0.13281479477882385,
      "learning_rate": 9.549242424242424e-05,
      "loss": 0.0308,
      "step": 595
    },
    {
      "epoch": 9.090909090909092,
      "grad_norm": 0.1456204354763031,
      "learning_rate": 9.545454545454546e-05,
      "loss": 0.0341,
      "step": 600
    },
    {
      "epoch": 9.166666666666666,
      "grad_norm": 0.13418981432914734,
      "learning_rate": 9.541666666666668e-05,
      "loss": 0.0288,
      "step": 605
    },
    {
      "epoch": 9.242424242424242,
      "grad_norm": 0.14538535475730896,
      "learning_rate": 9.537878787878788e-05,
      "loss": 0.0325,
      "step": 610
    },
    {
      "epoch": 9.318181818181818,
      "grad_norm": 0.16637976467609406,
      "learning_rate": 9.53409090909091e-05,
      "loss": 0.0375,
      "step": 615
    },
    {
      "epoch": 9.393939393939394,
      "grad_norm": 0.1578666716814041,
      "learning_rate": 9.53030303030303e-05,
      "loss": 0.0298,
      "step": 620
    },
    {
      "epoch": 9.469696969696969,
      "grad_norm": 0.13766665756702423,
      "learning_rate": 9.526515151515152e-05,
      "loss": 0.0292,
      "step": 625
    },
    {
      "epoch": 9.545454545454545,
      "grad_norm": 0.14352083206176758,
      "learning_rate": 9.522727272727273e-05,
      "loss": 0.0311,
      "step": 630
    },
    {
      "epoch": 9.621212121212121,
      "grad_norm": 0.163320854306221,
      "learning_rate": 9.518939393939395e-05,
      "loss": 0.0335,
      "step": 635
    },
    {
      "epoch": 9.696969696969697,
      "grad_norm": 0.16327431797981262,
      "learning_rate": 9.515151515151515e-05,
      "loss": 0.0378,
      "step": 640
    },
    {
      "epoch": 9.772727272727273,
      "grad_norm": 0.5533652901649475,
      "learning_rate": 9.511363636363637e-05,
      "loss": 0.0388,
      "step": 645
    },
    {
      "epoch": 9.848484848484848,
      "grad_norm": 0.17278915643692017,
      "learning_rate": 9.507575757575759e-05,
      "loss": 0.0333,
      "step": 650
    },
    {
      "epoch": 9.924242424242424,
      "grad_norm": 0.16369183361530304,
      "learning_rate": 9.50378787878788e-05,
      "loss": 0.0311,
      "step": 655
    },
    {
      "epoch": 10.0,
      "grad_norm": 1.009855031967163,
      "learning_rate": 9.5e-05,
      "loss": 0.0473,
      "step": 660
    },
    {
      "epoch": 10.0,
      "eval_loss": 0.030498143285512924,
      "eval_mean_accuracy": 0.9319052145053804,
      "eval_mean_iou": 0.9010954561929366,
      "eval_runtime": 35.0987,
      "eval_samples_per_second": 6.695,
      "eval_steps_per_second": 0.855,
      "step": 660
    },
    {
      "epoch": 10.075757575757576,
      "grad_norm": 0.2004818171262741,
      "learning_rate": 9.496212121212121e-05,
      "loss": 0.0346,
      "step": 665
    },
    {
      "epoch": 10.151515151515152,
      "grad_norm": 0.13994476199150085,
      "learning_rate": 9.492424242424242e-05,
      "loss": 0.028,
      "step": 670
    },
    {
      "epoch": 10.227272727272727,
      "grad_norm": 0.16897745430469513,
      "learning_rate": 9.488636363636364e-05,
      "loss": 0.0331,
      "step": 675
    },
    {
      "epoch": 10.303030303030303,
      "grad_norm": 0.23657144606113434,
      "learning_rate": 9.484848484848486e-05,
      "loss": 0.0324,
      "step": 680
    },
    {
      "epoch": 10.378787878787879,
      "grad_norm": 0.12473658472299576,
      "learning_rate": 9.481060606060606e-05,
      "loss": 0.0265,
      "step": 685
    },
    {
      "epoch": 10.454545454545455,
      "grad_norm": 0.12252913415431976,
      "learning_rate": 9.477272727272728e-05,
      "loss": 0.0299,
      "step": 690
    },
    {
      "epoch": 10.530303030303031,
      "grad_norm": 0.14586424827575684,
      "learning_rate": 9.47348484848485e-05,
      "loss": 0.0315,
      "step": 695
    },
    {
      "epoch": 10.606060606060606,
      "grad_norm": 0.18498015403747559,
      "learning_rate": 9.469696969696971e-05,
      "loss": 0.0302,
      "step": 700
    },
    {
      "epoch": 10.681818181818182,
      "grad_norm": 0.13446153700351715,
      "learning_rate": 9.465909090909091e-05,
      "loss": 0.0287,
      "step": 705
    },
    {
      "epoch": 10.757575757575758,
      "grad_norm": 0.29768991470336914,
      "learning_rate": 9.462121212121212e-05,
      "loss": 0.03,
      "step": 710
    },
    {
      "epoch": 10.833333333333334,
      "grad_norm": 0.2548729181289673,
      "learning_rate": 9.458333333333333e-05,
      "loss": 0.0328,
      "step": 715
    },
    {
      "epoch": 10.909090909090908,
      "grad_norm": 0.19455356895923615,
      "learning_rate": 9.454545454545455e-05,
      "loss": 0.0337,
      "step": 720
    },
    {
      "epoch": 10.984848484848484,
      "grad_norm": 0.2051912248134613,
      "learning_rate": 9.450757575757577e-05,
      "loss": 0.0291,
      "step": 725
    },
    {
      "epoch": 11.0,
      "eval_loss": 0.031183894723653793,
      "eval_mean_accuracy": 0.9099444740407038,
      "eval_mean_iou": 0.882907849723405,
      "eval_runtime": 37.8506,
      "eval_samples_per_second": 6.209,
      "eval_steps_per_second": 0.793,
      "step": 726
    },
    {
      "epoch": 11.06060606060606,
      "grad_norm": 0.23316840827465057,
      "learning_rate": 9.446969696969697e-05,
      "loss": 0.0301,
      "step": 730
    },
    {
      "epoch": 11.136363636363637,
      "grad_norm": 0.16540399193763733,
      "learning_rate": 9.443181818181819e-05,
      "loss": 0.0248,
      "step": 735
    },
    {
      "epoch": 11.212121212121213,
      "grad_norm": 0.16798929870128632,
      "learning_rate": 9.43939393939394e-05,
      "loss": 0.0288,
      "step": 740
    },
    {
      "epoch": 11.287878787878787,
      "grad_norm": 0.7785483598709106,
      "learning_rate": 9.435606060606062e-05,
      "loss": 0.0302,
      "step": 745
    },
    {
      "epoch": 11.363636363636363,
      "grad_norm": 0.3480140268802643,
      "learning_rate": 9.431818181818182e-05,
      "loss": 0.0304,
      "step": 750
    },
    {
      "epoch": 11.43939393939394,
      "grad_norm": 0.12131378054618835,
      "learning_rate": 9.428030303030302e-05,
      "loss": 0.0268,
      "step": 755
    },
    {
      "epoch": 11.515151515151516,
      "grad_norm": 0.3052375614643097,
      "learning_rate": 9.424242424242424e-05,
      "loss": 0.0296,
      "step": 760
    },
    {
      "epoch": 11.590909090909092,
      "grad_norm": 0.15812397003173828,
      "learning_rate": 9.420454545454546e-05,
      "loss": 0.0293,
      "step": 765
    },
    {
      "epoch": 11.666666666666666,
      "grad_norm": 0.0953594297170639,
      "learning_rate": 9.416666666666667e-05,
      "loss": 0.0254,
      "step": 770
    },
    {
      "epoch": 11.742424242424242,
      "grad_norm": 0.16207267343997955,
      "learning_rate": 9.412878787878788e-05,
      "loss": 0.0263,
      "step": 775
    },
    {
      "epoch": 11.818181818181818,
      "grad_norm": 1.0493265390396118,
      "learning_rate": 9.40909090909091e-05,
      "loss": 0.0302,
      "step": 780
    },
    {
      "epoch": 11.893939393939394,
      "grad_norm": 0.19405654072761536,
      "learning_rate": 9.405303030303031e-05,
      "loss": 0.0322,
      "step": 785
    },
    {
      "epoch": 11.969696969696969,
      "grad_norm": 0.19813752174377441,
      "learning_rate": 9.401515151515153e-05,
      "loss": 0.0361,
      "step": 790
    },
    {
      "epoch": 12.0,
      "eval_loss": 0.03096848353743553,
      "eval_mean_accuracy": 0.953395163058209,
      "eval_mean_iou": 0.8931990691438353,
      "eval_runtime": 40.4297,
      "eval_samples_per_second": 5.813,
      "eval_steps_per_second": 0.742,
      "step": 792
    },
    {
      "epoch": 12.045454545454545,
      "grad_norm": 0.1197110041975975,
      "learning_rate": 9.397727272727273e-05,
      "loss": 0.0334,
      "step": 795
    },
    {
      "epoch": 12.121212121212121,
      "grad_norm": 0.17101149260997772,
      "learning_rate": 9.393939393939395e-05,
      "loss": 0.0342,
      "step": 800
    },
    {
      "epoch": 12.196969696969697,
      "grad_norm": 0.12377778440713882,
      "learning_rate": 9.390151515151515e-05,
      "loss": 0.0295,
      "step": 805
    },
    {
      "epoch": 12.272727272727273,
      "grad_norm": 0.2139781266450882,
      "learning_rate": 9.386363636363637e-05,
      "loss": 0.0257,
      "step": 810
    },
    {
      "epoch": 12.348484848484848,
      "grad_norm": 0.1744828224182129,
      "learning_rate": 9.382575757575758e-05,
      "loss": 0.0272,
      "step": 815
    },
    {
      "epoch": 12.424242424242424,
      "grad_norm": 0.1008741557598114,
      "learning_rate": 9.378787878787879e-05,
      "loss": 0.032,
      "step": 820
    },
    {
      "epoch": 12.5,
      "grad_norm": 1.2705918550491333,
      "learning_rate": 9.375e-05,
      "loss": 0.0417,
      "step": 825
    },
    {
      "epoch": 12.575757575757576,
      "grad_norm": 0.1071762666106224,
      "learning_rate": 9.371212121212122e-05,
      "loss": 0.0313,
      "step": 830
    },
    {
      "epoch": 12.651515151515152,
      "grad_norm": 0.10107818990945816,
      "learning_rate": 9.367424242424244e-05,
      "loss": 0.0264,
      "step": 835
    },
    {
      "epoch": 12.727272727272727,
      "grad_norm": 0.11939676851034164,
      "learning_rate": 9.363636363636364e-05,
      "loss": 0.025,
      "step": 840
    },
    {
      "epoch": 12.803030303030303,
      "grad_norm": 0.16331548988819122,
      "learning_rate": 9.359848484848485e-05,
      "loss": 0.029,
      "step": 845
    },
    {
      "epoch": 12.878787878787879,
      "grad_norm": 0.08585627377033234,
      "learning_rate": 9.356060606060606e-05,
      "loss": 0.0259,
      "step": 850
    },
    {
      "epoch": 12.954545454545455,
      "grad_norm": 0.1356777399778366,
      "learning_rate": 9.352272727272727e-05,
      "loss": 0.0226,
      "step": 855
    },
    {
      "epoch": 13.0,
      "eval_loss": 0.026158301159739494,
      "eval_mean_accuracy": 0.9463841238229115,
      "eval_mean_iou": 0.9034798569395027,
      "eval_runtime": 41.1391,
      "eval_samples_per_second": 5.712,
      "eval_steps_per_second": 0.729,
      "step": 858
    },
    {
      "epoch": 13.030303030303031,
      "grad_norm": 0.12823344767093658,
      "learning_rate": 9.348484848484849e-05,
      "loss": 0.0254,
      "step": 860
    },
    {
      "epoch": 13.106060606060606,
      "grad_norm": 0.2597331404685974,
      "learning_rate": 9.34469696969697e-05,
      "loss": 0.0285,
      "step": 865
    },
    {
      "epoch": 13.181818181818182,
      "grad_norm": 0.168666809797287,
      "learning_rate": 9.340909090909091e-05,
      "loss": 0.0214,
      "step": 870
    },
    {
      "epoch": 13.257575757575758,
      "grad_norm": 0.08039192110300064,
      "learning_rate": 9.337121212121213e-05,
      "loss": 0.0251,
      "step": 875
    },
    {
      "epoch": 13.333333333333334,
      "grad_norm": 0.1234969049692154,
      "learning_rate": 9.333333333333334e-05,
      "loss": 0.0332,
      "step": 880
    },
    {
      "epoch": 13.409090909090908,
      "grad_norm": 0.12861916422843933,
      "learning_rate": 9.329545454545455e-05,
      "loss": 0.0379,
      "step": 885
    },
    {
      "epoch": 13.484848484848484,
      "grad_norm": 0.2895907163619995,
      "learning_rate": 9.325757575757576e-05,
      "loss": 0.0285,
      "step": 890
    },
    {
      "epoch": 13.56060606060606,
      "grad_norm": 0.10298188030719757,
      "learning_rate": 9.321969696969698e-05,
      "loss": 0.0272,
      "step": 895
    },
    {
      "epoch": 13.636363636363637,
      "grad_norm": 0.09611810743808746,
      "learning_rate": 9.318181818181818e-05,
      "loss": 0.0256,
      "step": 900
    },
    {
      "epoch": 13.712121212121213,
      "grad_norm": 0.08387937396764755,
      "learning_rate": 9.31439393939394e-05,
      "loss": 0.0208,
      "step": 905
    },
    {
      "epoch": 13.787878787878787,
      "grad_norm": 0.09015592932701111,
      "learning_rate": 9.31060606060606e-05,
      "loss": 0.0234,
      "step": 910
    },
    {
      "epoch": 13.863636363636363,
      "grad_norm": 0.09478239715099335,
      "learning_rate": 9.306818181818182e-05,
      "loss": 0.0229,
      "step": 915
    },
    {
      "epoch": 13.93939393939394,
      "grad_norm": 0.09258238226175308,
      "learning_rate": 9.303030303030303e-05,
      "loss": 0.0257,
      "step": 920
    },
    {
      "epoch": 14.0,
      "eval_loss": 0.02480032481253147,
      "eval_mean_accuracy": 0.9166888958980787,
      "eval_mean_iou": 0.8914179029195939,
      "eval_runtime": 34.5013,
      "eval_samples_per_second": 6.811,
      "eval_steps_per_second": 0.87,
      "step": 924
    },
    {
      "epoch": 14.015151515151516,
      "grad_norm": 0.08832405507564545,
      "learning_rate": 9.299242424242425e-05,
      "loss": 0.0317,
      "step": 925
    },
    {
      "epoch": 14.090909090909092,
      "grad_norm": 0.1963007003068924,
      "learning_rate": 9.295454545454545e-05,
      "loss": 0.0231,
      "step": 930
    },
    {
      "epoch": 14.166666666666666,
      "grad_norm": 0.08205869793891907,
      "learning_rate": 9.291666666666667e-05,
      "loss": 0.0208,
      "step": 935
    },
    {
      "epoch": 14.242424242424242,
      "grad_norm": 0.1925150454044342,
      "learning_rate": 9.287878787878789e-05,
      "loss": 0.024,
      "step": 940
    },
    {
      "epoch": 14.318181818181818,
      "grad_norm": 0.0778929591178894,
      "learning_rate": 9.28409090909091e-05,
      "loss": 0.0207,
      "step": 945
    },
    {
      "epoch": 14.393939393939394,
      "grad_norm": 0.10945238918066025,
      "learning_rate": 9.280303030303031e-05,
      "loss": 0.0254,
      "step": 950
    },
    {
      "epoch": 14.469696969696969,
      "grad_norm": 0.1330975741147995,
      "learning_rate": 9.276515151515151e-05,
      "loss": 0.0265,
      "step": 955
    },
    {
      "epoch": 14.545454545454545,
      "grad_norm": 0.07871849834918976,
      "learning_rate": 9.272727272727273e-05,
      "loss": 0.0239,
      "step": 960
    },
    {
      "epoch": 14.621212121212121,
      "grad_norm": 0.08424016833305359,
      "learning_rate": 9.268939393939394e-05,
      "loss": 0.0265,
      "step": 965
    },
    {
      "epoch": 14.696969696969697,
      "grad_norm": 0.10369917750358582,
      "learning_rate": 9.265151515151516e-05,
      "loss": 0.0224,
      "step": 970
    },
    {
      "epoch": 14.772727272727273,
      "grad_norm": 0.07275547087192535,
      "learning_rate": 9.261363636363636e-05,
      "loss": 0.0199,
      "step": 975
    },
    {
      "epoch": 14.848484848484848,
      "grad_norm": 0.07103205472230911,
      "learning_rate": 9.257575757575758e-05,
      "loss": 0.0213,
      "step": 980
    },
    {
      "epoch": 14.924242424242424,
      "grad_norm": 0.09189906716346741,
      "learning_rate": 9.25378787878788e-05,
      "loss": 0.0303,
      "step": 985
    },
    {
      "epoch": 15.0,
      "grad_norm": 0.4084390699863434,
      "learning_rate": 9.250000000000001e-05,
      "loss": 0.0246,
      "step": 990
    },
    {
      "epoch": 15.0,
      "eval_loss": 0.02351529523730278,
      "eval_mean_accuracy": 0.949832504172823,
      "eval_mean_iou": 0.9049995276347355,
      "eval_runtime": 34.7275,
      "eval_samples_per_second": 6.767,
      "eval_steps_per_second": 0.864,
      "step": 990
    },
    {
      "epoch": 15.075757575757576,
      "grad_norm": 0.14191284775733948,
      "learning_rate": 9.246212121212122e-05,
      "loss": 0.0203,
      "step": 995
    },
    {
      "epoch": 15.151515151515152,
      "grad_norm": 0.07639510929584503,
      "learning_rate": 9.242424242424242e-05,
      "loss": 0.0183,
      "step": 1000
    },
    {
      "epoch": 15.227272727272727,
      "grad_norm": 0.09424243867397308,
      "learning_rate": 9.238636363636363e-05,
      "loss": 0.0205,
      "step": 1005
    },
    {
      "epoch": 15.303030303030303,
      "grad_norm": 0.14415790140628815,
      "learning_rate": 9.234848484848485e-05,
      "loss": 0.0214,
      "step": 1010
    },
    {
      "epoch": 15.378787878787879,
      "grad_norm": 0.1036006435751915,
      "learning_rate": 9.231060606060607e-05,
      "loss": 0.0189,
      "step": 1015
    },
    {
      "epoch": 15.454545454545455,
      "grad_norm": 0.09542055428028107,
      "learning_rate": 9.227272727272727e-05,
      "loss": 0.0243,
      "step": 1020
    },
    {
      "epoch": 15.530303030303031,
      "grad_norm": 0.09930290281772614,
      "learning_rate": 9.223484848484849e-05,
      "loss": 0.0236,
      "step": 1025
    },
    {
      "epoch": 15.606060606060606,
      "grad_norm": 0.08781050145626068,
      "learning_rate": 9.21969696969697e-05,
      "loss": 0.0207,
      "step": 1030
    },
    {
      "epoch": 15.681818181818182,
      "grad_norm": 0.2105681598186493,
      "learning_rate": 9.215909090909092e-05,
      "loss": 0.025,
      "step": 1035
    },
    {
      "epoch": 15.757575757575758,
      "grad_norm": 0.07945144921541214,
      "learning_rate": 9.212121212121214e-05,
      "loss": 0.0214,
      "step": 1040
    },
    {
      "epoch": 15.833333333333334,
      "grad_norm": 0.0685647502541542,
      "learning_rate": 9.208333333333333e-05,
      "loss": 0.0267,
      "step": 1045
    },
    {
      "epoch": 15.909090909090908,
      "grad_norm": 0.1310282200574875,
      "learning_rate": 9.204545454545454e-05,
      "loss": 0.0211,
      "step": 1050
    },
    {
      "epoch": 15.984848484848484,
      "grad_norm": 0.0733848363161087,
      "learning_rate": 9.200757575757576e-05,
      "loss": 0.0237,
      "step": 1055
    },
    {
      "epoch": 16.0,
      "eval_loss": 0.02089761011302471,
      "eval_mean_accuracy": 0.9331587680755228,
      "eval_mean_iou": 0.9058876201409776,
      "eval_runtime": 41.3015,
      "eval_samples_per_second": 5.69,
      "eval_steps_per_second": 0.726,
      "step": 1056
    },
    {
      "epoch": 16.060606060606062,
      "grad_norm": 0.07626417279243469,
      "learning_rate": 9.196969696969698e-05,
      "loss": 0.0255,
      "step": 1060
    },
    {
      "epoch": 16.136363636363637,
      "grad_norm": 0.2288917601108551,
      "learning_rate": 9.193181818181818e-05,
      "loss": 0.0206,
      "step": 1065
    },
    {
      "epoch": 16.21212121212121,
      "grad_norm": 0.11269503086805344,
      "learning_rate": 9.18939393939394e-05,
      "loss": 0.0224,
      "step": 1070
    },
    {
      "epoch": 16.28787878787879,
      "grad_norm": 0.10464727133512497,
      "learning_rate": 9.185606060606061e-05,
      "loss": 0.0183,
      "step": 1075
    },
    {
      "epoch": 16.363636363636363,
      "grad_norm": 0.1360609382390976,
      "learning_rate": 9.181818181818183e-05,
      "loss": 0.0223,
      "step": 1080
    },
    {
      "epoch": 16.439393939393938,
      "grad_norm": 0.08922774344682693,
      "learning_rate": 9.178030303030305e-05,
      "loss": 0.0208,
      "step": 1085
    },
    {
      "epoch": 16.515151515151516,
      "grad_norm": 0.24164314568042755,
      "learning_rate": 9.174242424242425e-05,
      "loss": 0.024,
      "step": 1090
    },
    {
      "epoch": 16.59090909090909,
      "grad_norm": 0.20555198192596436,
      "learning_rate": 9.170454545454545e-05,
      "loss": 0.0193,
      "step": 1095
    },
    {
      "epoch": 16.666666666666668,
      "grad_norm": 0.319273442029953,
      "learning_rate": 9.166666666666667e-05,
      "loss": 0.0283,
      "step": 1100
    },
    {
      "epoch": 16.742424242424242,
      "grad_norm": 0.17492833733558655,
      "learning_rate": 9.162878787878788e-05,
      "loss": 0.0191,
      "step": 1105
    },
    {
      "epoch": 16.818181818181817,
      "grad_norm": 0.3059009313583374,
      "learning_rate": 9.159090909090909e-05,
      "loss": 0.0211,
      "step": 1110
    },
    {
      "epoch": 16.893939393939394,
      "grad_norm": 0.06272600591182709,
      "learning_rate": 9.15530303030303e-05,
      "loss": 0.0185,
      "step": 1115
    },
    {
      "epoch": 16.96969696969697,
      "grad_norm": 0.08431482315063477,
      "learning_rate": 9.151515151515152e-05,
      "loss": 0.0168,
      "step": 1120
    },
    {
      "epoch": 17.0,
      "eval_loss": 0.020880362018942833,
      "eval_mean_accuracy": 0.9414531365821945,
      "eval_mean_iou": 0.9085385747749858,
      "eval_runtime": 40.3233,
      "eval_samples_per_second": 5.828,
      "eval_steps_per_second": 0.744,
      "step": 1122
    },
    {
      "epoch": 17.045454545454547,
      "grad_norm": 0.11327819526195526,
      "learning_rate": 9.147727272727274e-05,
      "loss": 0.0252,
      "step": 1125
    },
    {
      "epoch": 17.12121212121212,
      "grad_norm": 0.13701438903808594,
      "learning_rate": 9.143939393939395e-05,
      "loss": 0.0216,
      "step": 1130
    },
    {
      "epoch": 17.196969696969695,
      "grad_norm": 0.22047576308250427,
      "learning_rate": 9.140151515151516e-05,
      "loss": 0.0206,
      "step": 1135
    },
    {
      "epoch": 17.272727272727273,
      "grad_norm": 0.08911259472370148,
      "learning_rate": 9.136363636363637e-05,
      "loss": 0.0221,
      "step": 1140
    },
    {
      "epoch": 17.348484848484848,
      "grad_norm": 0.2525509297847748,
      "learning_rate": 9.132575757575758e-05,
      "loss": 0.0209,
      "step": 1145
    },
    {
      "epoch": 17.424242424242426,
      "grad_norm": 0.066762775182724,
      "learning_rate": 9.128787878787879e-05,
      "loss": 0.0204,
      "step": 1150
    },
    {
      "epoch": 17.5,
      "grad_norm": 0.19102181494235992,
      "learning_rate": 9.125e-05,
      "loss": 0.0236,
      "step": 1155
    },
    {
      "epoch": 17.575757575757574,
      "grad_norm": 0.13234372437000275,
      "learning_rate": 9.121212121212121e-05,
      "loss": 0.018,
      "step": 1160
    },
    {
      "epoch": 17.651515151515152,
      "grad_norm": 0.09476885199546814,
      "learning_rate": 9.117424242424243e-05,
      "loss": 0.017,
      "step": 1165
    },
    {
      "epoch": 17.727272727272727,
      "grad_norm": 0.07637146860361099,
      "learning_rate": 9.113636363636365e-05,
      "loss": 0.0178,
      "step": 1170
    },
    {
      "epoch": 17.803030303030305,
      "grad_norm": 0.07766649127006531,
      "learning_rate": 9.109848484848486e-05,
      "loss": 0.0184,
      "step": 1175
    },
    {
      "epoch": 17.87878787878788,
      "grad_norm": 0.05786791816353798,
      "learning_rate": 9.106060606060606e-05,
      "loss": 0.0165,
      "step": 1180
    },
    {
      "epoch": 17.954545454545453,
      "grad_norm": 0.12654520571231842,
      "learning_rate": 9.102272727272728e-05,
      "loss": 0.0193,
      "step": 1185
    },
    {
      "epoch": 18.0,
      "eval_loss": 0.018953416496515274,
      "eval_mean_accuracy": 0.9479388526835519,
      "eval_mean_iou": 0.9136012174199729,
      "eval_runtime": 37.7471,
      "eval_samples_per_second": 6.226,
      "eval_steps_per_second": 0.795,
      "step": 1188
    },
    {
      "epoch": 18.03030303030303,
      "grad_norm": 0.05534211918711662,
      "learning_rate": 9.098484848484848e-05,
      "loss": 0.02,
      "step": 1190
    },
    {
      "epoch": 18.106060606060606,
      "grad_norm": 0.08186236768960953,
      "learning_rate": 9.09469696969697e-05,
      "loss": 0.0175,
      "step": 1195
    },
    {
      "epoch": 18.181818181818183,
      "grad_norm": 0.061108190566301346,
      "learning_rate": 9.090909090909092e-05,
      "loss": 0.0191,
      "step": 1200
    },
    {
      "epoch": 18.257575757575758,
      "grad_norm": 0.09618055820465088,
      "learning_rate": 9.087121212121212e-05,
      "loss": 0.0197,
      "step": 1205
    },
    {
      "epoch": 18.333333333333332,
      "grad_norm": 0.07263849675655365,
      "learning_rate": 9.083333333333334e-05,
      "loss": 0.0156,
      "step": 1210
    },
    {
      "epoch": 18.40909090909091,
      "grad_norm": 0.08618713915348053,
      "learning_rate": 9.079545454545455e-05,
      "loss": 0.0216,
      "step": 1215
    },
    {
      "epoch": 18.484848484848484,
      "grad_norm": 0.07267585396766663,
      "learning_rate": 9.075757575757577e-05,
      "loss": 0.0205,
      "step": 1220
    },
    {
      "epoch": 18.560606060606062,
      "grad_norm": 0.15266315639019012,
      "learning_rate": 9.071969696969697e-05,
      "loss": 0.023,
      "step": 1225
    },
    {
      "epoch": 18.636363636363637,
      "grad_norm": 0.055605363100767136,
      "learning_rate": 9.068181818181819e-05,
      "loss": 0.0154,
      "step": 1230
    },
    {
      "epoch": 18.71212121212121,
      "grad_norm": 0.05512472614645958,
      "learning_rate": 9.06439393939394e-05,
      "loss": 0.0177,
      "step": 1235
    },
    {
      "epoch": 18.78787878787879,
      "grad_norm": 0.12221260368824005,
      "learning_rate": 9.060606060606061e-05,
      "loss": 0.0181,
      "step": 1240
    },
    {
      "epoch": 18.863636363636363,
      "grad_norm": 0.37372562289237976,
      "learning_rate": 9.056818181818183e-05,
      "loss": 0.0198,
      "step": 1245
    },
    {
      "epoch": 18.939393939393938,
      "grad_norm": 0.16616711020469666,
      "learning_rate": 9.053030303030303e-05,
      "loss": 0.0164,
      "step": 1250
    },
    {
      "epoch": 19.0,
      "eval_loss": 0.01951894909143448,
      "eval_mean_accuracy": 0.9524658568188715,
      "eval_mean_iou": 0.9115952350230673,
      "eval_runtime": 40.7492,
      "eval_samples_per_second": 5.767,
      "eval_steps_per_second": 0.736,
      "step": 1254
    },
    {
      "epoch": 19.015151515151516,
      "grad_norm": 0.0696030706167221,
      "learning_rate": 9.049242424242424e-05,
      "loss": 0.0166,
      "step": 1255
    },
    {
      "epoch": 19.09090909090909,
      "grad_norm": 0.07677122950553894,
      "learning_rate": 9.045454545454546e-05,
      "loss": 0.02,
      "step": 1260
    },
    {
      "epoch": 19.166666666666668,
      "grad_norm": 0.2776522636413574,
      "learning_rate": 9.041666666666668e-05,
      "loss": 0.0198,
      "step": 1265
    },
    {
      "epoch": 19.242424242424242,
      "grad_norm": 0.05005476251244545,
      "learning_rate": 9.037878787878788e-05,
      "loss": 0.0167,
      "step": 1270
    },
    {
      "epoch": 19.318181818181817,
      "grad_norm": 0.059852272272109985,
      "learning_rate": 9.03409090909091e-05,
      "loss": 0.0192,
      "step": 1275
    },
    {
      "epoch": 19.393939393939394,
      "grad_norm": 0.12680846452713013,
      "learning_rate": 9.030303030303031e-05,
      "loss": 0.0187,
      "step": 1280
    },
    {
      "epoch": 19.46969696969697,
      "grad_norm": 0.14521756768226624,
      "learning_rate": 9.026515151515153e-05,
      "loss": 0.0201,
      "step": 1285
    },
    {
      "epoch": 19.545454545454547,
      "grad_norm": 0.053130924701690674,
      "learning_rate": 9.022727272727273e-05,
      "loss": 0.0152,
      "step": 1290
    },
    {
      "epoch": 19.62121212121212,
      "grad_norm": 0.06579020619392395,
      "learning_rate": 9.018939393939394e-05,
      "loss": 0.0163,
      "step": 1295
    },
    {
      "epoch": 19.696969696969695,
      "grad_norm": 0.07096754759550095,
      "learning_rate": 9.015151515151515e-05,
      "loss": 0.0177,
      "step": 1300
    },
    {
      "epoch": 19.772727272727273,
      "grad_norm": 0.1248403787612915,
      "learning_rate": 9.011363636363637e-05,
      "loss": 0.018,
      "step": 1305
    },
    {
      "epoch": 19.848484848484848,
      "grad_norm": 0.059574246406555176,
      "learning_rate": 9.007575757575759e-05,
      "loss": 0.0157,
      "step": 1310
    },
    {
      "epoch": 19.924242424242426,
      "grad_norm": 0.1070672795176506,
      "learning_rate": 9.003787878787879e-05,
      "loss": 0.0211,
      "step": 1315
    },
    {
      "epoch": 20.0,
      "grad_norm": 0.1226222887635231,
      "learning_rate": 9e-05,
      "loss": 0.0178,
      "step": 1320
    },
    {
      "epoch": 20.0,
      "eval_loss": 0.018179375678300858,
      "eval_mean_accuracy": 0.9432248453051383,
      "eval_mean_iou": 0.9112924368958693,
      "eval_runtime": 40.401,
      "eval_samples_per_second": 5.817,
      "eval_steps_per_second": 0.743,
      "step": 1320
    },
    {
      "epoch": 20.075757575757574,
      "grad_norm": 0.1072898581624031,
      "learning_rate": 8.996212121212122e-05,
      "loss": 0.018,
      "step": 1325
    },
    {
      "epoch": 20.151515151515152,
      "grad_norm": 0.08781715482473373,
      "learning_rate": 8.992424242424244e-05,
      "loss": 0.0184,
      "step": 1330
    },
    {
      "epoch": 20.227272727272727,
      "grad_norm": 0.06843460351228714,
      "learning_rate": 8.988636363636364e-05,
      "loss": 0.0188,
      "step": 1335
    },
    {
      "epoch": 20.303030303030305,
      "grad_norm": 0.10172174870967865,
      "learning_rate": 8.984848484848484e-05,
      "loss": 0.0155,
      "step": 1340
    },
    {
      "epoch": 20.37878787878788,
      "grad_norm": 0.04984050244092941,
      "learning_rate": 8.981060606060606e-05,
      "loss": 0.0152,
      "step": 1345
    },
    {
      "epoch": 20.454545454545453,
      "grad_norm": 0.1292862594127655,
      "learning_rate": 8.977272727272728e-05,
      "loss": 0.0162,
      "step": 1350
    },
    {
      "epoch": 20.53030303030303,
      "grad_norm": 0.10163886845111847,
      "learning_rate": 8.97348484848485e-05,
      "loss": 0.0165,
      "step": 1355
    },
    {
      "epoch": 20.606060606060606,
      "grad_norm": 0.05588515102863312,
      "learning_rate": 8.96969696969697e-05,
      "loss": 0.0235,
      "step": 1360
    },
    {
      "epoch": 20.681818181818183,
      "grad_norm": 0.08275488764047623,
      "learning_rate": 8.965909090909091e-05,
      "loss": 0.0204,
      "step": 1365
    },
    {
      "epoch": 20.757575757575758,
      "grad_norm": 0.10210832208395004,
      "learning_rate": 8.962121212121213e-05,
      "loss": 0.0168,
      "step": 1370
    },
    {
      "epoch": 20.833333333333332,
      "grad_norm": 0.08400724828243256,
      "learning_rate": 8.958333333333335e-05,
      "loss": 0.017,
      "step": 1375
    },
    {
      "epoch": 20.90909090909091,
      "grad_norm": 0.05280748009681702,
      "learning_rate": 8.954545454545455e-05,
      "loss": 0.0175,
      "step": 1380
    },
    {
      "epoch": 20.984848484848484,
      "grad_norm": 0.110740527510643,
      "learning_rate": 8.950757575757575e-05,
      "loss": 0.0265,
      "step": 1385
    },
    {
      "epoch": 21.0,
      "eval_loss": 0.01816415600478649,
      "eval_mean_accuracy": 0.9433752074266861,
      "eval_mean_iou": 0.9117099712387807,
      "eval_runtime": 39.2116,
      "eval_samples_per_second": 5.993,
      "eval_steps_per_second": 0.765,
      "step": 1386
    },
    {
      "epoch": 21.060606060606062,
      "grad_norm": 0.1256209909915924,
      "learning_rate": 8.946969696969697e-05,
      "loss": 0.0178,
      "step": 1390
    },
    {
      "epoch": 21.136363636363637,
      "grad_norm": 0.16505251824855804,
      "learning_rate": 8.943181818181819e-05,
      "loss": 0.0189,
      "step": 1395
    },
    {
      "epoch": 21.21212121212121,
      "grad_norm": 0.0947386622428894,
      "learning_rate": 8.93939393939394e-05,
      "loss": 0.0173,
      "step": 1400
    },
    {
      "epoch": 21.28787878787879,
      "grad_norm": 0.060093361884355545,
      "learning_rate": 8.93560606060606e-05,
      "loss": 0.0187,
      "step": 1405
    },
    {
      "epoch": 21.363636363636363,
      "grad_norm": 0.09483636915683746,
      "learning_rate": 8.931818181818182e-05,
      "loss": 0.0146,
      "step": 1410
    },
    {
      "epoch": 21.439393939393938,
      "grad_norm": 0.10417105257511139,
      "learning_rate": 8.928030303030304e-05,
      "loss": 0.0169,
      "step": 1415
    },
    {
      "epoch": 21.515151515151516,
      "grad_norm": 0.1067221537232399,
      "learning_rate": 8.924242424242426e-05,
      "loss": 0.0186,
      "step": 1420
    },
    {
      "epoch": 21.59090909090909,
      "grad_norm": 0.038923099637031555,
      "learning_rate": 8.920454545454546e-05,
      "loss": 0.0161,
      "step": 1425
    },
    {
      "epoch": 21.666666666666668,
      "grad_norm": 0.08629253506660461,
      "learning_rate": 8.916666666666667e-05,
      "loss": 0.0163,
      "step": 1430
    },
    {
      "epoch": 21.742424242424242,
      "grad_norm": 0.24443939328193665,
      "learning_rate": 8.912878787878788e-05,
      "loss": 0.0186,
      "step": 1435
    },
    {
      "epoch": 21.818181818181817,
      "grad_norm": 0.06895927339792252,
      "learning_rate": 8.90909090909091e-05,
      "loss": 0.0165,
      "step": 1440
    },
    {
      "epoch": 21.893939393939394,
      "grad_norm": 0.07047417759895325,
      "learning_rate": 8.905303030303031e-05,
      "loss": 0.0191,
      "step": 1445
    },
    {
      "epoch": 21.96969696969697,
      "grad_norm": 0.07517726719379425,
      "learning_rate": 8.901515151515151e-05,
      "loss": 0.0149,
      "step": 1450
    },
    {
      "epoch": 22.0,
      "eval_loss": 0.020833691582083702,
      "eval_mean_accuracy": 0.8975814960033816,
      "eval_mean_iou": 0.8740122945223798,
      "eval_runtime": 41.9313,
      "eval_samples_per_second": 5.604,
      "eval_steps_per_second": 0.715,
      "step": 1452
    },
    {
      "epoch": 22.045454545454547,
      "grad_norm": 0.1332666128873825,
      "learning_rate": 8.897727272727273e-05,
      "loss": 0.0183,
      "step": 1455
    },
    {
      "epoch": 22.12121212121212,
      "grad_norm": 0.11585947126150131,
      "learning_rate": 8.893939393939395e-05,
      "loss": 0.0164,
      "step": 1460
    },
    {
      "epoch": 22.196969696969695,
      "grad_norm": 0.09697961062192917,
      "learning_rate": 8.890151515151516e-05,
      "loss": 0.0159,
      "step": 1465
    },
    {
      "epoch": 22.272727272727273,
      "grad_norm": 0.0729738399386406,
      "learning_rate": 8.886363636363637e-05,
      "loss": 0.0154,
      "step": 1470
    },
    {
      "epoch": 22.348484848484848,
      "grad_norm": 0.07646304368972778,
      "learning_rate": 8.882575757575758e-05,
      "loss": 0.017,
      "step": 1475
    },
    {
      "epoch": 22.424242424242426,
      "grad_norm": 0.10247813910245895,
      "learning_rate": 8.87878787878788e-05,
      "loss": 0.0146,
      "step": 1480
    },
    {
      "epoch": 22.5,
      "grad_norm": 0.0826369896531105,
      "learning_rate": 8.875e-05,
      "loss": 0.0145,
      "step": 1485
    },
    {
      "epoch": 22.575757575757574,
      "grad_norm": 0.061552003026008606,
      "learning_rate": 8.871212121212122e-05,
      "loss": 0.0144,
      "step": 1490
    },
    {
      "epoch": 22.651515151515152,
      "grad_norm": 0.20094624161720276,
      "learning_rate": 8.867424242424242e-05,
      "loss": 0.0173,
      "step": 1495
    },
    {
      "epoch": 22.727272727272727,
      "grad_norm": 0.049575645476579666,
      "learning_rate": 8.863636363636364e-05,
      "loss": 0.0164,
      "step": 1500
    },
    {
      "epoch": 22.803030303030305,
      "grad_norm": 0.10276893526315689,
      "learning_rate": 8.859848484848485e-05,
      "loss": 0.02,
      "step": 1505
    },
    {
      "epoch": 22.87878787878788,
      "grad_norm": 0.0979294702410698,
      "learning_rate": 8.856060606060607e-05,
      "loss": 0.0179,
      "step": 1510
    },
    {
      "epoch": 22.954545454545453,
      "grad_norm": 0.06229182332754135,
      "learning_rate": 8.852272727272727e-05,
      "loss": 0.0189,
      "step": 1515
    },
    {
      "epoch": 23.0,
      "eval_loss": 0.017257515341043472,
      "eval_mean_accuracy": 0.9403299187712085,
      "eval_mean_iou": 0.910537943036256,
      "eval_runtime": 41.0221,
      "eval_samples_per_second": 5.729,
      "eval_steps_per_second": 0.731,
      "step": 1518
    },
    {
      "epoch": 23.03030303030303,
      "grad_norm": 0.05739479139447212,
      "learning_rate": 8.848484848484849e-05,
      "loss": 0.0161,
      "step": 1520
    },
    {
      "epoch": 23.106060606060606,
      "grad_norm": 0.05012909322977066,
      "learning_rate": 8.844696969696971e-05,
      "loss": 0.0191,
      "step": 1525
    },
    {
      "epoch": 23.181818181818183,
      "grad_norm": 0.1779683381319046,
      "learning_rate": 8.840909090909091e-05,
      "loss": 0.0175,
      "step": 1530
    },
    {
      "epoch": 23.257575757575758,
      "grad_norm": 0.09027104079723358,
      "learning_rate": 8.837121212121213e-05,
      "loss": 0.0159,
      "step": 1535
    },
    {
      "epoch": 23.333333333333332,
      "grad_norm": 0.053633686155080795,
      "learning_rate": 8.833333333333333e-05,
      "loss": 0.0149,
      "step": 1540
    },
    {
      "epoch": 23.40909090909091,
      "grad_norm": 0.05222697928547859,
      "learning_rate": 8.829545454545455e-05,
      "loss": 0.0153,
      "step": 1545
    },
    {
      "epoch": 23.484848484848484,
      "grad_norm": 0.0841483622789383,
      "learning_rate": 8.825757575757576e-05,
      "loss": 0.0159,
      "step": 1550
    },
    {
      "epoch": 23.560606060606062,
      "grad_norm": 0.1662817895412445,
      "learning_rate": 8.821969696969698e-05,
      "loss": 0.0156,
      "step": 1555
    },
    {
      "epoch": 23.636363636363637,
      "grad_norm": 0.07585467398166656,
      "learning_rate": 8.818181818181818e-05,
      "loss": 0.0137,
      "step": 1560
    },
    {
      "epoch": 23.71212121212121,
      "grad_norm": 0.0699695274233818,
      "learning_rate": 8.81439393939394e-05,
      "loss": 0.0152,
      "step": 1565
    },
    {
      "epoch": 23.78787878787879,
      "grad_norm": 0.17828363180160522,
      "learning_rate": 8.810606060606062e-05,
      "loss": 0.0177,
      "step": 1570
    },
    {
      "epoch": 23.863636363636363,
      "grad_norm": 0.23298045992851257,
      "learning_rate": 8.806818181818183e-05,
      "loss": 0.0188,
      "step": 1575
    },
    {
      "epoch": 23.939393939393938,
      "grad_norm": 0.09302160888910294,
      "learning_rate": 8.803030303030304e-05,
      "loss": 0.0187,
      "step": 1580
    },
    {
      "epoch": 24.0,
      "eval_loss": 0.01727343164384365,
      "eval_mean_accuracy": 0.9292851681759875,
      "eval_mean_iou": 0.9045289934001666,
      "eval_runtime": 37.9462,
      "eval_samples_per_second": 6.193,
      "eval_steps_per_second": 0.791,
      "step": 1584
    },
    {
      "epoch": 24.015151515151516,
      "grad_norm": 0.06782419979572296,
      "learning_rate": 8.799242424242424e-05,
      "loss": 0.0227,
      "step": 1585
    },
    {
      "epoch": 24.09090909090909,
      "grad_norm": 0.10343431681394577,
      "learning_rate": 8.795454545454545e-05,
      "loss": 0.016,
      "step": 1590
    },
    {
      "epoch": 24.166666666666668,
      "grad_norm": 0.09992711246013641,
      "learning_rate": 8.791666666666667e-05,
      "loss": 0.0161,
      "step": 1595
    },
    {
      "epoch": 24.242424242424242,
      "grad_norm": 0.10084862262010574,
      "learning_rate": 8.787878787878789e-05,
      "loss": 0.0132,
      "step": 1600
    },
    {
      "epoch": 24.318181818181817,
      "grad_norm": 0.09875880181789398,
      "learning_rate": 8.784090909090909e-05,
      "loss": 0.015,
      "step": 1605
    },
    {
      "epoch": 24.393939393939394,
      "grad_norm": 0.07086075097322464,
      "learning_rate": 8.780303030303031e-05,
      "loss": 0.0138,
      "step": 1610
    },
    {
      "epoch": 24.46969696969697,
      "grad_norm": 0.04054461047053337,
      "learning_rate": 8.776515151515152e-05,
      "loss": 0.0152,
      "step": 1615
    },
    {
      "epoch": 24.545454545454547,
      "grad_norm": 0.04059217497706413,
      "learning_rate": 8.772727272727274e-05,
      "loss": 0.0138,
      "step": 1620
    },
    {
      "epoch": 24.62121212121212,
      "grad_norm": 0.17548631131649017,
      "learning_rate": 8.768939393939394e-05,
      "loss": 0.0169,
      "step": 1625
    },
    {
      "epoch": 24.696969696969695,
      "grad_norm": 0.06841861456632614,
      "learning_rate": 8.765151515151515e-05,
      "loss": 0.0214,
      "step": 1630
    },
    {
      "epoch": 24.772727272727273,
      "grad_norm": 0.10320588946342468,
      "learning_rate": 8.761363636363636e-05,
      "loss": 0.0209,
      "step": 1635
    },
    {
      "epoch": 24.848484848484848,
      "grad_norm": 0.12154887616634369,
      "learning_rate": 8.757575757575758e-05,
      "loss": 0.0145,
      "step": 1640
    },
    {
      "epoch": 24.924242424242426,
      "grad_norm": 0.06344634294509888,
      "learning_rate": 8.75378787878788e-05,
      "loss": 0.0143,
      "step": 1645
    },
    {
      "epoch": 25.0,
      "grad_norm": 0.12611378729343414,
      "learning_rate": 8.75e-05,
      "loss": 0.0177,
      "step": 1650
    },
    {
      "epoch": 25.0,
      "eval_loss": 0.016218528151512146,
      "eval_mean_accuracy": 0.9509601400138873,
      "eval_mean_iou": 0.9155473781894271,
      "eval_runtime": 37.6623,
      "eval_samples_per_second": 6.24,
      "eval_steps_per_second": 0.797,
      "step": 1650
    },
    {
      "epoch": 25.075757575757574,
      "grad_norm": 0.0751732736825943,
      "learning_rate": 8.746212121212122e-05,
      "loss": 0.0156,
      "step": 1655
    },
    {
      "epoch": 25.151515151515152,
      "grad_norm": 0.04189491644501686,
      "learning_rate": 8.742424242424243e-05,
      "loss": 0.0146,
      "step": 1660
    },
    {
      "epoch": 25.227272727272727,
      "grad_norm": 0.05579284951090813,
      "learning_rate": 8.738636363636365e-05,
      "loss": 0.0118,
      "step": 1665
    },
    {
      "epoch": 25.303030303030305,
      "grad_norm": 0.17117440700531006,
      "learning_rate": 8.734848484848485e-05,
      "loss": 0.0207,
      "step": 1670
    },
    {
      "epoch": 25.37878787878788,
      "grad_norm": 0.06103392690420151,
      "learning_rate": 8.731060606060605e-05,
      "loss": 0.0165,
      "step": 1675
    },
    {
      "epoch": 25.454545454545453,
      "grad_norm": 0.056527379900217056,
      "learning_rate": 8.727272727272727e-05,
      "loss": 0.012,
      "step": 1680
    },
    {
      "epoch": 25.53030303030303,
      "grad_norm": 0.04721764102578163,
      "learning_rate": 8.723484848484849e-05,
      "loss": 0.0192,
      "step": 1685
    },
    {
      "epoch": 25.606060606060606,
      "grad_norm": 0.09804808348417282,
      "learning_rate": 8.71969696969697e-05,
      "loss": 0.0167,
      "step": 1690
    },
    {
      "epoch": 25.681818181818183,
      "grad_norm": 0.05927478149533272,
      "learning_rate": 8.715909090909091e-05,
      "loss": 0.0227,
      "step": 1695
    },
    {
      "epoch": 25.757575757575758,
      "grad_norm": 0.17171381413936615,
      "learning_rate": 8.712121212121212e-05,
      "loss": 0.0174,
      "step": 1700
    },
    {
      "epoch": 25.833333333333332,
      "grad_norm": 0.06737685948610306,
      "learning_rate": 8.708333333333334e-05,
      "loss": 0.0177,
      "step": 1705
    },
    {
      "epoch": 25.90909090909091,
      "grad_norm": 0.13116347789764404,
      "learning_rate": 8.704545454545456e-05,
      "loss": 0.0168,
      "step": 1710
    },
    {
      "epoch": 25.984848484848484,
      "grad_norm": 0.09560756385326385,
      "learning_rate": 8.700757575757576e-05,
      "loss": 0.0191,
      "step": 1715
    },
    {
      "epoch": 26.0,
      "eval_loss": 0.018060876056551933,
      "eval_mean_accuracy": 0.9644576157367912,
      "eval_mean_iou": 0.9071648091231127,
      "eval_runtime": 38.8503,
      "eval_samples_per_second": 6.049,
      "eval_steps_per_second": 0.772,
      "step": 1716
    },
    {
      "epoch": 26.060606060606062,
      "grad_norm": 0.07194330543279648,
      "learning_rate": 8.696969696969698e-05,
      "loss": 0.0166,
      "step": 1720
    },
    {
      "epoch": 26.136363636363637,
      "grad_norm": 0.10216555744409561,
      "learning_rate": 8.693181818181818e-05,
      "loss": 0.02,
      "step": 1725
    },
    {
      "epoch": 26.21212121212121,
      "grad_norm": 0.0822141170501709,
      "learning_rate": 8.68939393939394e-05,
      "loss": 0.0154,
      "step": 1730
    },
    {
      "epoch": 26.28787878787879,
      "grad_norm": 0.15010997653007507,
      "learning_rate": 8.685606060606061e-05,
      "loss": 0.0148,
      "step": 1735
    },
    {
      "epoch": 26.363636363636363,
      "grad_norm": 0.1780836135149002,
      "learning_rate": 8.681818181818182e-05,
      "loss": 0.0155,
      "step": 1740
    },
    {
      "epoch": 26.439393939393938,
      "grad_norm": 0.36100277304649353,
      "learning_rate": 8.678030303030303e-05,
      "loss": 0.0183,
      "step": 1745
    },
    {
      "epoch": 26.515151515151516,
      "grad_norm": 0.45620912313461304,
      "learning_rate": 8.674242424242425e-05,
      "loss": 0.0257,
      "step": 1750
    },
    {
      "epoch": 26.59090909090909,
      "grad_norm": 0.08473169058561325,
      "learning_rate": 8.670454545454547e-05,
      "loss": 0.0173,
      "step": 1755
    },
    {
      "epoch": 26.666666666666668,
      "grad_norm": 0.11675431579351425,
      "learning_rate": 8.666666666666667e-05,
      "loss": 0.0147,
      "step": 1760
    },
    {
      "epoch": 26.742424242424242,
      "grad_norm": 0.09614520519971848,
      "learning_rate": 8.662878787878788e-05,
      "loss": 0.0159,
      "step": 1765
    },
    {
      "epoch": 26.818181818181817,
      "grad_norm": 0.10838396102190018,
      "learning_rate": 8.65909090909091e-05,
      "loss": 0.0158,
      "step": 1770
    },
    {
      "epoch": 26.893939393939394,
      "grad_norm": 0.06302326917648315,
      "learning_rate": 8.65530303030303e-05,
      "loss": 0.0163,
      "step": 1775
    },
    {
      "epoch": 26.96969696969697,
      "grad_norm": 0.07336388528347015,
      "learning_rate": 8.651515151515152e-05,
      "loss": 0.0165,
      "step": 1780
    },
    {
      "epoch": 27.0,
      "eval_loss": 0.017215073108673096,
      "eval_mean_accuracy": 0.9264071364673326,
      "eval_mean_iou": 0.9004306182260974,
      "eval_runtime": 38.3757,
      "eval_samples_per_second": 6.124,
      "eval_steps_per_second": 0.782,
      "step": 1782
    },
    {
      "epoch": 27.045454545454547,
      "grad_norm": 0.06732988357543945,
      "learning_rate": 8.647727272727272e-05,
      "loss": 0.0151,
      "step": 1785
    },
    {
      "epoch": 27.12121212121212,
      "grad_norm": 0.2578263282775879,
      "learning_rate": 8.643939393939394e-05,
      "loss": 0.0204,
      "step": 1790
    },
    {
      "epoch": 27.196969696969695,
      "grad_norm": 0.08113482594490051,
      "learning_rate": 8.640151515151516e-05,
      "loss": 0.0162,
      "step": 1795
    },
    {
      "epoch": 27.272727272727273,
      "grad_norm": 0.060990288853645325,
      "learning_rate": 8.636363636363637e-05,
      "loss": 0.0133,
      "step": 1800
    },
    {
      "epoch": 27.348484848484848,
      "grad_norm": 0.06865828484296799,
      "learning_rate": 8.632575757575758e-05,
      "loss": 0.0131,
      "step": 1805
    },
    {
      "epoch": 27.424242424242426,
      "grad_norm": 0.15351566672325134,
      "learning_rate": 8.628787878787879e-05,
      "loss": 0.0177,
      "step": 1810
    },
    {
      "epoch": 27.5,
      "grad_norm": 0.055379584431648254,
      "learning_rate": 8.625000000000001e-05,
      "loss": 0.0146,
      "step": 1815
    },
    {
      "epoch": 27.575757575757574,
      "grad_norm": 0.04952799901366234,
      "learning_rate": 8.621212121212121e-05,
      "loss": 0.0145,
      "step": 1820
    },
    {
      "epoch": 27.651515151515152,
      "grad_norm": 0.07691798359155655,
      "learning_rate": 8.617424242424243e-05,
      "loss": 0.0191,
      "step": 1825
    },
    {
      "epoch": 27.727272727272727,
      "grad_norm": 0.05929436534643173,
      "learning_rate": 8.613636363636363e-05,
      "loss": 0.0166,
      "step": 1830
    },
    {
      "epoch": 27.803030303030305,
      "grad_norm": 0.10661505162715912,
      "learning_rate": 8.609848484848485e-05,
      "loss": 0.0129,
      "step": 1835
    },
    {
      "epoch": 27.87878787878788,
      "grad_norm": 0.13317100703716278,
      "learning_rate": 8.606060606060606e-05,
      "loss": 0.0159,
      "step": 1840
    },
    {
      "epoch": 27.954545454545453,
      "grad_norm": 0.1373067945241928,
      "learning_rate": 8.602272727272728e-05,
      "loss": 0.0141,
      "step": 1845
    },
    {
      "epoch": 28.0,
      "eval_loss": 0.016448181122541428,
      "eval_mean_accuracy": 0.9376718472677584,
      "eval_mean_iou": 0.9093956123573079,
      "eval_runtime": 36.872,
      "eval_samples_per_second": 6.373,
      "eval_steps_per_second": 0.814,
      "step": 1848
    },
    {
      "epoch": 28.03030303030303,
      "grad_norm": 0.04432457685470581,
      "learning_rate": 8.598484848484848e-05,
      "loss": 0.0132,
      "step": 1850
    },
    {
      "epoch": 28.106060606060606,
      "grad_norm": 0.09766856580972672,
      "learning_rate": 8.59469696969697e-05,
      "loss": 0.0119,
      "step": 1855
    },
    {
      "epoch": 28.181818181818183,
      "grad_norm": 0.05427827686071396,
      "learning_rate": 8.590909090909092e-05,
      "loss": 0.0146,
      "step": 1860
    },
    {
      "epoch": 28.257575757575758,
      "grad_norm": 0.08492875844240189,
      "learning_rate": 8.587121212121213e-05,
      "loss": 0.0143,
      "step": 1865
    },
    {
      "epoch": 28.333333333333332,
      "grad_norm": 0.15752869844436646,
      "learning_rate": 8.583333333333334e-05,
      "loss": 0.015,
      "step": 1870
    },
    {
      "epoch": 28.40909090909091,
      "grad_norm": 0.03909919410943985,
      "learning_rate": 8.579545454545454e-05,
      "loss": 0.0107,
      "step": 1875
    },
    {
      "epoch": 28.484848484848484,
      "grad_norm": 0.050600919872522354,
      "learning_rate": 8.575757575757576e-05,
      "loss": 0.0165,
      "step": 1880
    },
    {
      "epoch": 28.560606060606062,
      "grad_norm": 0.11105132102966309,
      "learning_rate": 8.571969696969697e-05,
      "loss": 0.0154,
      "step": 1885
    },
    {
      "epoch": 28.636363636363637,
      "grad_norm": 0.03404098004102707,
      "learning_rate": 8.568181818181819e-05,
      "loss": 0.0141,
      "step": 1890
    },
    {
      "epoch": 28.71212121212121,
      "grad_norm": 0.09969612956047058,
      "learning_rate": 8.564393939393939e-05,
      "loss": 0.017,
      "step": 1895
    },
    {
      "epoch": 28.78787878787879,
      "grad_norm": 0.1037430390715599,
      "learning_rate": 8.560606060606061e-05,
      "loss": 0.0166,
      "step": 1900
    },
    {
      "epoch": 28.863636363636363,
      "grad_norm": 0.05573015660047531,
      "learning_rate": 8.556818181818183e-05,
      "loss": 0.0142,
      "step": 1905
    },
    {
      "epoch": 28.939393939393938,
      "grad_norm": 0.056287191808223724,
      "learning_rate": 8.553030303030304e-05,
      "loss": 0.0161,
      "step": 1910
    },
    {
      "epoch": 29.0,
      "eval_loss": 0.015243682079017162,
      "eval_mean_accuracy": 0.9468581657780586,
      "eval_mean_iou": 0.9160524775493769,
      "eval_runtime": 36.3893,
      "eval_samples_per_second": 6.458,
      "eval_steps_per_second": 0.824,
      "step": 1914
    },
    {
      "epoch": 29.015151515151516,
      "grad_norm": 0.06723536550998688,
      "learning_rate": 8.549242424242425e-05,
      "loss": 0.0136,
      "step": 1915
    },
    {
      "epoch": 29.09090909090909,
      "grad_norm": 0.0491645373404026,
      "learning_rate": 8.545454545454545e-05,
      "loss": 0.0158,
      "step": 1920
    },
    {
      "epoch": 29.166666666666668,
      "grad_norm": 0.10214636474847794,
      "learning_rate": 8.541666666666666e-05,
      "loss": 0.0128,
      "step": 1925
    },
    {
      "epoch": 29.242424242424242,
      "grad_norm": 0.09459321200847626,
      "learning_rate": 8.537878787878788e-05,
      "loss": 0.0125,
      "step": 1930
    },
    {
      "epoch": 29.318181818181817,
      "grad_norm": 0.08826824277639389,
      "learning_rate": 8.53409090909091e-05,
      "loss": 0.0137,
      "step": 1935
    },
    {
      "epoch": 29.393939393939394,
      "grad_norm": 0.07649490237236023,
      "learning_rate": 8.53030303030303e-05,
      "loss": 0.0182,
      "step": 1940
    },
    {
      "epoch": 29.46969696969697,
      "grad_norm": 0.1806381344795227,
      "learning_rate": 8.526515151515152e-05,
      "loss": 0.0117,
      "step": 1945
    },
    {
      "epoch": 29.545454545454547,
      "grad_norm": 0.09376899152994156,
      "learning_rate": 8.522727272727273e-05,
      "loss": 0.0177,
      "step": 1950
    },
    {
      "epoch": 29.62121212121212,
      "grad_norm": 0.041973233222961426,
      "learning_rate": 8.518939393939395e-05,
      "loss": 0.0141,
      "step": 1955
    },
    {
      "epoch": 29.696969696969695,
      "grad_norm": 0.06091102957725525,
      "learning_rate": 8.515151515151515e-05,
      "loss": 0.0132,
      "step": 1960
    },
    {
      "epoch": 29.772727272727273,
      "grad_norm": 0.05774535983800888,
      "learning_rate": 8.511363636363637e-05,
      "loss": 0.0171,
      "step": 1965
    },
    {
      "epoch": 29.848484848484848,
      "grad_norm": 0.08068455755710602,
      "learning_rate": 8.507575757575757e-05,
      "loss": 0.0169,
      "step": 1970
    },
    {
      "epoch": 29.924242424242426,
      "grad_norm": 0.0695386677980423,
      "learning_rate": 8.503787878787879e-05,
      "loss": 0.0164,
      "step": 1975
    },
    {
      "epoch": 30.0,
      "grad_norm": 0.11324222385883331,
      "learning_rate": 8.5e-05,
      "loss": 0.0114,
      "step": 1980
    },
    {
      "epoch": 30.0,
      "eval_loss": 0.015367408283054829,
      "eval_mean_accuracy": 0.9604017201799336,
      "eval_mean_iou": 0.9164267790956439,
      "eval_runtime": 38.4197,
      "eval_samples_per_second": 6.117,
      "eval_steps_per_second": 0.781,
      "step": 1980
    },
    {
      "epoch": 30.075757575757574,
      "grad_norm": 0.12909524142742157,
      "learning_rate": 8.496212121212121e-05,
      "loss": 0.0123,
      "step": 1985
    },
    {
      "epoch": 30.151515151515152,
      "grad_norm": 0.07254344969987869,
      "learning_rate": 8.492424242424243e-05,
      "loss": 0.0139,
      "step": 1990
    },
    {
      "epoch": 30.227272727272727,
      "grad_norm": 0.37066739797592163,
      "learning_rate": 8.488636363636364e-05,
      "loss": 0.0159,
      "step": 1995
    },
    {
      "epoch": 30.303030303030305,
      "grad_norm": 0.07680170983076096,
      "learning_rate": 8.484848484848486e-05,
      "loss": 0.0105,
      "step": 2000
    },
    {
      "epoch": 30.37878787878788,
      "grad_norm": 0.04912244528532028,
      "learning_rate": 8.481060606060606e-05,
      "loss": 0.0159,
      "step": 2005
    },
    {
      "epoch": 30.454545454545453,
      "grad_norm": 0.12234355509281158,
      "learning_rate": 8.477272727272728e-05,
      "loss": 0.016,
      "step": 2010
    },
    {
      "epoch": 30.53030303030303,
      "grad_norm": 0.14877626299858093,
      "learning_rate": 8.473484848484848e-05,
      "loss": 0.0174,
      "step": 2015
    },
    {
      "epoch": 30.606060606060606,
      "grad_norm": 0.07860980182886124,
      "learning_rate": 8.46969696969697e-05,
      "loss": 0.0159,
      "step": 2020
    },
    {
      "epoch": 30.681818181818183,
      "grad_norm": 0.06150322034955025,
      "learning_rate": 8.465909090909091e-05,
      "loss": 0.0143,
      "step": 2025
    },
    {
      "epoch": 30.757575757575758,
      "grad_norm": 0.06585820019245148,
      "learning_rate": 8.462121212121212e-05,
      "loss": 0.0171,
      "step": 2030
    },
    {
      "epoch": 30.833333333333332,
      "grad_norm": 0.1390763521194458,
      "learning_rate": 8.458333333333333e-05,
      "loss": 0.0125,
      "step": 2035
    },
    {
      "epoch": 30.90909090909091,
      "grad_norm": 0.1276780515909195,
      "learning_rate": 8.454545454545455e-05,
      "loss": 0.0145,
      "step": 2040
    },
    {
      "epoch": 30.984848484848484,
      "grad_norm": 0.05084895342588425,
      "learning_rate": 8.450757575757577e-05,
      "loss": 0.0138,
      "step": 2045
    },
    {
      "epoch": 31.0,
      "eval_loss": 0.015100602060556412,
      "eval_mean_accuracy": 0.9339994718170246,
      "eval_mean_iou": 0.9099176068234552,
      "eval_runtime": 42.9174,
      "eval_samples_per_second": 5.476,
      "eval_steps_per_second": 0.699,
      "step": 2046
    },
    {
      "epoch": 31.060606060606062,
      "grad_norm": 0.034428905695676804,
      "learning_rate": 8.446969696969697e-05,
      "loss": 0.0114,
      "step": 2050
    },
    {
      "epoch": 31.136363636363637,
      "grad_norm": 0.07030927389860153,
      "learning_rate": 8.443181818181819e-05,
      "loss": 0.0148,
      "step": 2055
    },
    {
      "epoch": 31.21212121212121,
      "grad_norm": 0.0496549978852272,
      "learning_rate": 8.43939393939394e-05,
      "loss": 0.0123,
      "step": 2060
    },
    {
      "epoch": 31.28787878787879,
      "grad_norm": 0.030932331457734108,
      "learning_rate": 8.43560606060606e-05,
      "loss": 0.0102,
      "step": 2065
    },
    {
      "epoch": 31.363636363636363,
      "grad_norm": 0.07468841224908829,
      "learning_rate": 8.431818181818182e-05,
      "loss": 0.0134,
      "step": 2070
    },
    {
      "epoch": 31.439393939393938,
      "grad_norm": 0.1080772802233696,
      "learning_rate": 8.428030303030303e-05,
      "loss": 0.0129,
      "step": 2075
    },
    {
      "epoch": 31.515151515151516,
      "grad_norm": 0.050275810062885284,
      "learning_rate": 8.424242424242424e-05,
      "loss": 0.0118,
      "step": 2080
    },
    {
      "epoch": 31.59090909090909,
      "grad_norm": 0.1436341553926468,
      "learning_rate": 8.420454545454546e-05,
      "loss": 0.0153,
      "step": 2085
    },
    {
      "epoch": 31.666666666666668,
      "grad_norm": 0.06254042685031891,
      "learning_rate": 8.416666666666668e-05,
      "loss": 0.015,
      "step": 2090
    },
    {
      "epoch": 31.742424242424242,
      "grad_norm": 0.06055235117673874,
      "learning_rate": 8.412878787878788e-05,
      "loss": 0.0109,
      "step": 2095
    },
    {
      "epoch": 31.818181818181817,
      "grad_norm": 0.11876318603754044,
      "learning_rate": 8.40909090909091e-05,
      "loss": 0.0174,
      "step": 2100
    },
    {
      "epoch": 31.893939393939394,
      "grad_norm": 0.11994854360818863,
      "learning_rate": 8.405303030303031e-05,
      "loss": 0.0172,
      "step": 2105
    },
    {
      "epoch": 31.96969696969697,
      "grad_norm": 0.05131750926375389,
      "learning_rate": 8.401515151515153e-05,
      "loss": 0.0144,
      "step": 2110
    },
    {
      "epoch": 32.0,
      "eval_loss": 0.014232022687792778,
      "eval_mean_accuracy": 0.9513312344016893,
      "eval_mean_iou": 0.9192330335908245,
      "eval_runtime": 38.4672,
      "eval_samples_per_second": 6.109,
      "eval_steps_per_second": 0.78,
      "step": 2112
    },
    {
      "epoch": 32.04545454545455,
      "grad_norm": 0.06063136085867882,
      "learning_rate": 8.397727272727273e-05,
      "loss": 0.0186,
      "step": 2115
    },
    {
      "epoch": 32.121212121212125,
      "grad_norm": 0.054627787321805954,
      "learning_rate": 8.393939393939393e-05,
      "loss": 0.0124,
      "step": 2120
    },
    {
      "epoch": 32.196969696969695,
      "grad_norm": 0.09246259927749634,
      "learning_rate": 8.390151515151515e-05,
      "loss": 0.0167,
      "step": 2125
    },
    {
      "epoch": 32.27272727272727,
      "grad_norm": 0.07662643492221832,
      "learning_rate": 8.386363636363637e-05,
      "loss": 0.0126,
      "step": 2130
    },
    {
      "epoch": 32.34848484848485,
      "grad_norm": 0.10725986957550049,
      "learning_rate": 8.382575757575758e-05,
      "loss": 0.0145,
      "step": 2135
    },
    {
      "epoch": 32.42424242424242,
      "grad_norm": 0.07634241878986359,
      "learning_rate": 8.378787878787879e-05,
      "loss": 0.0123,
      "step": 2140
    },
    {
      "epoch": 32.5,
      "grad_norm": 0.06595125049352646,
      "learning_rate": 8.375e-05,
      "loss": 0.0162,
      "step": 2145
    },
    {
      "epoch": 32.57575757575758,
      "grad_norm": 0.06710188090801239,
      "learning_rate": 8.371212121212122e-05,
      "loss": 0.0157,
      "step": 2150
    },
    {
      "epoch": 32.65151515151515,
      "grad_norm": 0.12186404317617416,
      "learning_rate": 8.367424242424244e-05,
      "loss": 0.0132,
      "step": 2155
    },
    {
      "epoch": 32.72727272727273,
      "grad_norm": 0.05659103766083717,
      "learning_rate": 8.363636363636364e-05,
      "loss": 0.0122,
      "step": 2160
    },
    {
      "epoch": 32.803030303030305,
      "grad_norm": 0.2276543378829956,
      "learning_rate": 8.359848484848484e-05,
      "loss": 0.0129,
      "step": 2165
    },
    {
      "epoch": 32.878787878787875,
      "grad_norm": 0.03798941522836685,
      "learning_rate": 8.356060606060606e-05,
      "loss": 0.0169,
      "step": 2170
    },
    {
      "epoch": 32.95454545454545,
      "grad_norm": 0.03684765100479126,
      "learning_rate": 8.352272727272727e-05,
      "loss": 0.0161,
      "step": 2175
    },
    {
      "epoch": 33.0,
      "eval_loss": 0.014978514984250069,
      "eval_mean_accuracy": 0.9446832629526215,
      "eval_mean_iou": 0.9142589616071185,
      "eval_runtime": 38.4894,
      "eval_samples_per_second": 6.106,
      "eval_steps_per_second": 0.779,
      "step": 2178
    },
    {
      "epoch": 33.03030303030303,
      "grad_norm": 0.07790496200323105,
      "learning_rate": 8.348484848484849e-05,
      "loss": 0.0149,
      "step": 2180
    },
    {
      "epoch": 33.10606060606061,
      "grad_norm": 0.041151996701955795,
      "learning_rate": 8.34469696969697e-05,
      "loss": 0.0157,
      "step": 2185
    },
    {
      "epoch": 33.18181818181818,
      "grad_norm": 0.07710423320531845,
      "learning_rate": 8.340909090909091e-05,
      "loss": 0.0149,
      "step": 2190
    },
    {
      "epoch": 33.25757575757576,
      "grad_norm": 0.05968889966607094,
      "learning_rate": 8.337121212121213e-05,
      "loss": 0.0132,
      "step": 2195
    },
    {
      "epoch": 33.333333333333336,
      "grad_norm": 0.1608990877866745,
      "learning_rate": 8.333333333333334e-05,
      "loss": 0.0135,
      "step": 2200
    },
    {
      "epoch": 33.40909090909091,
      "grad_norm": 0.03847300261259079,
      "learning_rate": 8.329545454545456e-05,
      "loss": 0.0128,
      "step": 2205
    },
    {
      "epoch": 33.484848484848484,
      "grad_norm": 0.03341185301542282,
      "learning_rate": 8.325757575757575e-05,
      "loss": 0.0148,
      "step": 2210
    },
    {
      "epoch": 33.56060606060606,
      "grad_norm": 0.0915418490767479,
      "learning_rate": 8.321969696969697e-05,
      "loss": 0.0132,
      "step": 2215
    },
    {
      "epoch": 33.63636363636363,
      "grad_norm": 0.03685728460550308,
      "learning_rate": 8.318181818181818e-05,
      "loss": 0.0139,
      "step": 2220
    },
    {
      "epoch": 33.71212121212121,
      "grad_norm": 0.04617101326584816,
      "learning_rate": 8.31439393939394e-05,
      "loss": 0.0122,
      "step": 2225
    },
    {
      "epoch": 33.78787878787879,
      "grad_norm": 0.058006107807159424,
      "learning_rate": 8.310606060606062e-05,
      "loss": 0.0197,
      "step": 2230
    },
    {
      "epoch": 33.86363636363637,
      "grad_norm": 0.12553590536117554,
      "learning_rate": 8.306818181818182e-05,
      "loss": 0.0126,
      "step": 2235
    },
    {
      "epoch": 33.93939393939394,
      "grad_norm": 0.08858802169561386,
      "learning_rate": 8.303030303030304e-05,
      "loss": 0.0115,
      "step": 2240
    },
    {
      "epoch": 34.0,
      "eval_loss": 0.014615623280405998,
      "eval_mean_accuracy": 0.9434780272896767,
      "eval_mean_iou": 0.915723673694996,
      "eval_runtime": 37.8721,
      "eval_samples_per_second": 6.205,
      "eval_steps_per_second": 0.792,
      "step": 2244
    },
    {
      "epoch": 34.015151515151516,
      "grad_norm": 0.056122202426195145,
      "learning_rate": 8.299242424242425e-05,
      "loss": 0.0135,
      "step": 2245
    },
    {
      "epoch": 34.09090909090909,
      "grad_norm": 0.037949059158563614,
      "learning_rate": 8.295454545454547e-05,
      "loss": 0.0123,
      "step": 2250
    },
    {
      "epoch": 34.166666666666664,
      "grad_norm": 0.07207425683736801,
      "learning_rate": 8.291666666666667e-05,
      "loss": 0.0193,
      "step": 2255
    },
    {
      "epoch": 34.24242424242424,
      "grad_norm": 0.13813644647598267,
      "learning_rate": 8.287878787878787e-05,
      "loss": 0.0153,
      "step": 2260
    },
    {
      "epoch": 34.31818181818182,
      "grad_norm": 0.10792592912912369,
      "learning_rate": 8.284090909090909e-05,
      "loss": 0.0135,
      "step": 2265
    },
    {
      "epoch": 34.39393939393939,
      "grad_norm": 0.08793012797832489,
      "learning_rate": 8.280303030303031e-05,
      "loss": 0.0136,
      "step": 2270
    },
    {
      "epoch": 34.46969696969697,
      "grad_norm": 0.07147427648305893,
      "learning_rate": 8.276515151515152e-05,
      "loss": 0.0134,
      "step": 2275
    },
    {
      "epoch": 34.54545454545455,
      "grad_norm": 0.04749901965260506,
      "learning_rate": 8.272727272727273e-05,
      "loss": 0.0108,
      "step": 2280
    },
    {
      "epoch": 34.621212121212125,
      "grad_norm": 0.10276518762111664,
      "learning_rate": 8.268939393939394e-05,
      "loss": 0.0174,
      "step": 2285
    },
    {
      "epoch": 34.696969696969695,
      "grad_norm": 0.11331098526716232,
      "learning_rate": 8.265151515151516e-05,
      "loss": 0.0129,
      "step": 2290
    },
    {
      "epoch": 34.77272727272727,
      "grad_norm": 0.047895096242427826,
      "learning_rate": 8.261363636363638e-05,
      "loss": 0.0115,
      "step": 2295
    },
    {
      "epoch": 34.84848484848485,
      "grad_norm": 0.17029358446598053,
      "learning_rate": 8.257575757575758e-05,
      "loss": 0.0117,
      "step": 2300
    },
    {
      "epoch": 34.92424242424242,
      "grad_norm": 0.09644016623497009,
      "learning_rate": 8.25378787878788e-05,
      "loss": 0.0163,
      "step": 2305
    },
    {
      "epoch": 35.0,
      "grad_norm": 0.09039471298456192,
      "learning_rate": 8.25e-05,
      "loss": 0.0138,
      "step": 2310
    },
    {
      "epoch": 35.0,
      "eval_loss": 0.014718534424901009,
      "eval_mean_accuracy": 0.944126916779635,
      "eval_mean_iou": 0.912400165109001,
      "eval_runtime": 37.7765,
      "eval_samples_per_second": 6.221,
      "eval_steps_per_second": 0.794,
      "step": 2310
    },
    {
      "epoch": 35.07575757575758,
      "grad_norm": 0.053617559373378754,
      "learning_rate": 8.246212121212122e-05,
      "loss": 0.0142,
      "step": 2315
    },
    {
      "epoch": 35.15151515151515,
      "grad_norm": 0.05597515404224396,
      "learning_rate": 8.242424242424243e-05,
      "loss": 0.0129,
      "step": 2320
    },
    {
      "epoch": 35.22727272727273,
      "grad_norm": 0.19057418406009674,
      "learning_rate": 8.238636363636364e-05,
      "loss": 0.0151,
      "step": 2325
    },
    {
      "epoch": 35.303030303030305,
      "grad_norm": 0.044556934386491776,
      "learning_rate": 8.234848484848485e-05,
      "loss": 0.0111,
      "step": 2330
    },
    {
      "epoch": 35.378787878787875,
      "grad_norm": 0.08347728848457336,
      "learning_rate": 8.231060606060607e-05,
      "loss": 0.0151,
      "step": 2335
    },
    {
      "epoch": 35.45454545454545,
      "grad_norm": 0.20699027180671692,
      "learning_rate": 8.227272727272729e-05,
      "loss": 0.0134,
      "step": 2340
    },
    {
      "epoch": 35.53030303030303,
      "grad_norm": 0.055762626230716705,
      "learning_rate": 8.223484848484849e-05,
      "loss": 0.0123,
      "step": 2345
    },
    {
      "epoch": 35.60606060606061,
      "grad_norm": 0.13136908411979675,
      "learning_rate": 8.21969696969697e-05,
      "loss": 0.0164,
      "step": 2350
    },
    {
      "epoch": 35.68181818181818,
      "grad_norm": 0.05497013404965401,
      "learning_rate": 8.215909090909091e-05,
      "loss": 0.0155,
      "step": 2355
    },
    {
      "epoch": 35.75757575757576,
      "grad_norm": 0.09149844944477081,
      "learning_rate": 8.212121212121212e-05,
      "loss": 0.0108,
      "step": 2360
    },
    {
      "epoch": 35.833333333333336,
      "grad_norm": 0.0727037638425827,
      "learning_rate": 8.208333333333334e-05,
      "loss": 0.0131,
      "step": 2365
    },
    {
      "epoch": 35.90909090909091,
      "grad_norm": 0.06407946348190308,
      "learning_rate": 8.204545454545454e-05,
      "loss": 0.0146,
      "step": 2370
    },
    {
      "epoch": 35.984848484848484,
      "grad_norm": 0.0770241990685463,
      "learning_rate": 8.200757575757576e-05,
      "loss": 0.0128,
      "step": 2375
    },
    {
      "epoch": 36.0,
      "eval_loss": 0.014062165282666683,
      "eval_mean_accuracy": 0.958079278547082,
      "eval_mean_iou": 0.9195207225183076,
      "eval_runtime": 36.112,
      "eval_samples_per_second": 6.508,
      "eval_steps_per_second": 0.831,
      "step": 2376
    },
    {
      "epoch": 36.06060606060606,
      "grad_norm": 0.12653273344039917,
      "learning_rate": 8.196969696969698e-05,
      "loss": 0.0125,
      "step": 2380
    },
    {
      "epoch": 36.13636363636363,
      "grad_norm": 0.0538422055542469,
      "learning_rate": 8.19318181818182e-05,
      "loss": 0.0111,
      "step": 2385
    },
    {
      "epoch": 36.21212121212121,
      "grad_norm": 0.05743101239204407,
      "learning_rate": 8.18939393939394e-05,
      "loss": 0.0173,
      "step": 2390
    },
    {
      "epoch": 36.28787878787879,
      "grad_norm": 0.10392845422029495,
      "learning_rate": 8.185606060606061e-05,
      "loss": 0.0103,
      "step": 2395
    },
    {
      "epoch": 36.36363636363637,
      "grad_norm": 0.06989451497793198,
      "learning_rate": 8.181818181818183e-05,
      "loss": 0.0099,
      "step": 2400
    },
    {
      "epoch": 36.43939393939394,
      "grad_norm": 0.12108205258846283,
      "learning_rate": 8.178030303030303e-05,
      "loss": 0.0171,
      "step": 2405
    },
    {
      "epoch": 36.515151515151516,
      "grad_norm": 0.03485065698623657,
      "learning_rate": 8.174242424242425e-05,
      "loss": 0.0123,
      "step": 2410
    },
    {
      "epoch": 36.59090909090909,
      "grad_norm": 0.06650546193122864,
      "learning_rate": 8.170454545454545e-05,
      "loss": 0.0175,
      "step": 2415
    },
    {
      "epoch": 36.666666666666664,
      "grad_norm": 0.039845433086156845,
      "learning_rate": 8.166666666666667e-05,
      "loss": 0.012,
      "step": 2420
    },
    {
      "epoch": 36.74242424242424,
      "grad_norm": 0.08973627537488937,
      "learning_rate": 8.162878787878789e-05,
      "loss": 0.0112,
      "step": 2425
    },
    {
      "epoch": 36.81818181818182,
      "grad_norm": 0.0863916277885437,
      "learning_rate": 8.15909090909091e-05,
      "loss": 0.0145,
      "step": 2430
    },
    {
      "epoch": 36.89393939393939,
      "grad_norm": 0.06171298399567604,
      "learning_rate": 8.15530303030303e-05,
      "loss": 0.017,
      "step": 2435
    },
    {
      "epoch": 36.96969696969697,
      "grad_norm": 0.06107330694794655,
      "learning_rate": 8.151515151515152e-05,
      "loss": 0.0121,
      "step": 2440
    },
    {
      "epoch": 37.0,
      "eval_loss": 0.013979670591652393,
      "eval_mean_accuracy": 0.9526889369341363,
      "eval_mean_iou": 0.9183460350045268,
      "eval_runtime": 35.4578,
      "eval_samples_per_second": 6.628,
      "eval_steps_per_second": 0.846,
      "step": 2442
    },
    {
      "epoch": 37.04545454545455,
      "grad_norm": 0.05428370088338852,
      "learning_rate": 8.147727272727274e-05,
      "loss": 0.014,
      "step": 2445
    },
    {
      "epoch": 37.121212121212125,
      "grad_norm": 0.06593769788742065,
      "learning_rate": 8.143939393939395e-05,
      "loss": 0.0117,
      "step": 2450
    },
    {
      "epoch": 37.196969696969695,
      "grad_norm": 0.071235291659832,
      "learning_rate": 8.140151515151516e-05,
      "loss": 0.0126,
      "step": 2455
    },
    {
      "epoch": 37.27272727272727,
      "grad_norm": 0.04574131593108177,
      "learning_rate": 8.136363636363636e-05,
      "loss": 0.0146,
      "step": 2460
    },
    {
      "epoch": 37.34848484848485,
      "grad_norm": 0.07325722277164459,
      "learning_rate": 8.132575757575758e-05,
      "loss": 0.0118,
      "step": 2465
    },
    {
      "epoch": 37.42424242424242,
      "grad_norm": 0.10241012275218964,
      "learning_rate": 8.12878787878788e-05,
      "loss": 0.0163,
      "step": 2470
    },
    {
      "epoch": 37.5,
      "grad_norm": 0.10306882113218307,
      "learning_rate": 8.125000000000001e-05,
      "loss": 0.0145,
      "step": 2475
    },
    {
      "epoch": 37.57575757575758,
      "grad_norm": 0.09917788952589035,
      "learning_rate": 8.121212121212121e-05,
      "loss": 0.0142,
      "step": 2480
    },
    {
      "epoch": 37.65151515151515,
      "grad_norm": 0.0961432009935379,
      "learning_rate": 8.117424242424243e-05,
      "loss": 0.0149,
      "step": 2485
    },
    {
      "epoch": 37.72727272727273,
      "grad_norm": 0.03723779320716858,
      "learning_rate": 8.113636363636365e-05,
      "loss": 0.0139,
      "step": 2490
    },
    {
      "epoch": 37.803030303030305,
      "grad_norm": 0.06299424916505814,
      "learning_rate": 8.109848484848486e-05,
      "loss": 0.0123,
      "step": 2495
    },
    {
      "epoch": 37.878787878787875,
      "grad_norm": 0.07499869167804718,
      "learning_rate": 8.106060606060607e-05,
      "loss": 0.0129,
      "step": 2500
    },
    {
      "epoch": 37.95454545454545,
      "grad_norm": 0.1560441106557846,
      "learning_rate": 8.102272727272727e-05,
      "loss": 0.0113,
      "step": 2505
    },
    {
      "epoch": 38.0,
      "eval_loss": 0.015773385763168335,
      "eval_mean_accuracy": 0.9271945872884162,
      "eval_mean_iou": 0.9031807862279778,
      "eval_runtime": 41.8183,
      "eval_samples_per_second": 5.62,
      "eval_steps_per_second": 0.717,
      "step": 2508
    },
    {
      "epoch": 38.03030303030303,
      "grad_norm": 0.0485360324382782,
      "learning_rate": 8.098484848484848e-05,
      "loss": 0.0178,
      "step": 2510
    },
    {
      "epoch": 38.10606060606061,
      "grad_norm": 0.038664255291223526,
      "learning_rate": 8.09469696969697e-05,
      "loss": 0.0097,
      "step": 2515
    },
    {
      "epoch": 38.18181818181818,
      "grad_norm": 0.12002712488174438,
      "learning_rate": 8.090909090909092e-05,
      "loss": 0.0129,
      "step": 2520
    },
    {
      "epoch": 38.25757575757576,
      "grad_norm": 0.07412465661764145,
      "learning_rate": 8.087121212121212e-05,
      "loss": 0.0113,
      "step": 2525
    },
    {
      "epoch": 38.333333333333336,
      "grad_norm": 0.04746988043189049,
      "learning_rate": 8.083333333333334e-05,
      "loss": 0.0115,
      "step": 2530
    },
    {
      "epoch": 38.40909090909091,
      "grad_norm": 0.06748125702142715,
      "learning_rate": 8.079545454545455e-05,
      "loss": 0.0148,
      "step": 2535
    },
    {
      "epoch": 38.484848484848484,
      "grad_norm": 0.07398603111505508,
      "learning_rate": 8.075757575757577e-05,
      "loss": 0.0126,
      "step": 2540
    },
    {
      "epoch": 38.56060606060606,
      "grad_norm": 0.03598150610923767,
      "learning_rate": 8.071969696969697e-05,
      "loss": 0.0113,
      "step": 2545
    },
    {
      "epoch": 38.63636363636363,
      "grad_norm": 0.06754065304994583,
      "learning_rate": 8.068181818181818e-05,
      "loss": 0.0132,
      "step": 2550
    },
    {
      "epoch": 38.71212121212121,
      "grad_norm": 0.09649275243282318,
      "learning_rate": 8.064393939393939e-05,
      "loss": 0.0134,
      "step": 2555
    },
    {
      "epoch": 38.78787878787879,
      "grad_norm": 0.13337717950344086,
      "learning_rate": 8.060606060606061e-05,
      "loss": 0.0153,
      "step": 2560
    },
    {
      "epoch": 38.86363636363637,
      "grad_norm": 0.04376828670501709,
      "learning_rate": 8.056818181818183e-05,
      "loss": 0.0132,
      "step": 2565
    },
    {
      "epoch": 38.93939393939394,
      "grad_norm": 0.08427128195762634,
      "learning_rate": 8.053030303030303e-05,
      "loss": 0.016,
      "step": 2570
    },
    {
      "epoch": 39.0,
      "eval_loss": 0.014303994365036488,
      "eval_mean_accuracy": 0.9630122293011035,
      "eval_mean_iou": 0.9188319659961146,
      "eval_runtime": 39.9802,
      "eval_samples_per_second": 5.878,
      "eval_steps_per_second": 0.75,
      "step": 2574
    },
    {
      "epoch": 39.015151515151516,
      "grad_norm": 0.04401514306664467,
      "learning_rate": 8.049242424242425e-05,
      "loss": 0.0157,
      "step": 2575
    },
    {
      "epoch": 39.09090909090909,
      "grad_norm": 0.07215376198291779,
      "learning_rate": 8.045454545454546e-05,
      "loss": 0.0098,
      "step": 2580
    },
    {
      "epoch": 39.166666666666664,
      "grad_norm": 0.05644891411066055,
      "learning_rate": 8.041666666666668e-05,
      "loss": 0.0169,
      "step": 2585
    },
    {
      "epoch": 39.24242424242424,
      "grad_norm": 0.038284022361040115,
      "learning_rate": 8.037878787878788e-05,
      "loss": 0.0125,
      "step": 2590
    },
    {
      "epoch": 39.31818181818182,
      "grad_norm": 0.05088666081428528,
      "learning_rate": 8.03409090909091e-05,
      "loss": 0.0139,
      "step": 2595
    },
    {
      "epoch": 39.39393939393939,
      "grad_norm": 0.18770332634449005,
      "learning_rate": 8.03030303030303e-05,
      "loss": 0.0128,
      "step": 2600
    },
    {
      "epoch": 39.46969696969697,
      "grad_norm": 0.04921300709247589,
      "learning_rate": 8.026515151515152e-05,
      "loss": 0.0123,
      "step": 2605
    },
    {
      "epoch": 39.54545454545455,
      "grad_norm": 0.0448882132768631,
      "learning_rate": 8.022727272727273e-05,
      "loss": 0.0142,
      "step": 2610
    },
    {
      "epoch": 39.621212121212125,
      "grad_norm": 0.06699865311384201,
      "learning_rate": 8.018939393939394e-05,
      "loss": 0.0116,
      "step": 2615
    },
    {
      "epoch": 39.696969696969695,
      "grad_norm": 0.051503922790288925,
      "learning_rate": 8.015151515151515e-05,
      "loss": 0.0112,
      "step": 2620
    },
    {
      "epoch": 39.77272727272727,
      "grad_norm": 0.03496067598462105,
      "learning_rate": 8.011363636363637e-05,
      "loss": 0.0151,
      "step": 2625
    },
    {
      "epoch": 39.84848484848485,
      "grad_norm": 0.06014701724052429,
      "learning_rate": 8.007575757575759e-05,
      "loss": 0.0126,
      "step": 2630
    },
    {
      "epoch": 39.92424242424242,
      "grad_norm": 0.03566575422883034,
      "learning_rate": 8.003787878787879e-05,
      "loss": 0.0114,
      "step": 2635
    },
    {
      "epoch": 40.0,
      "grad_norm": 0.15319675207138062,
      "learning_rate": 8e-05,
      "loss": 0.0125,
      "step": 2640
    },
    {
      "epoch": 40.0,
      "eval_loss": 0.014355950057506561,
      "eval_mean_accuracy": 0.9408709767181943,
      "eval_mean_iou": 0.914492648030509,
      "eval_runtime": 36.3631,
      "eval_samples_per_second": 6.463,
      "eval_steps_per_second": 0.825,
      "step": 2640
    },
    {
      "epoch": 40.07575757575758,
      "grad_norm": 0.0465068481862545,
      "learning_rate": 7.996212121212121e-05,
      "loss": 0.0112,
      "step": 2645
    },
    {
      "epoch": 40.15151515151515,
      "grad_norm": 0.05283168703317642,
      "learning_rate": 7.992424242424243e-05,
      "loss": 0.0105,
      "step": 2650
    },
    {
      "epoch": 40.22727272727273,
      "grad_norm": 0.04744649678468704,
      "learning_rate": 7.988636363636364e-05,
      "loss": 0.0115,
      "step": 2655
    },
    {
      "epoch": 40.303030303030305,
      "grad_norm": 0.11857732385396957,
      "learning_rate": 7.984848484848485e-05,
      "loss": 0.0108,
      "step": 2660
    },
    {
      "epoch": 40.378787878787875,
      "grad_norm": 0.0855904221534729,
      "learning_rate": 7.981060606060606e-05,
      "loss": 0.0107,
      "step": 2665
    },
    {
      "epoch": 40.45454545454545,
      "grad_norm": 0.050729889422655106,
      "learning_rate": 7.977272727272728e-05,
      "loss": 0.0118,
      "step": 2670
    },
    {
      "epoch": 40.53030303030303,
      "grad_norm": 0.08516892045736313,
      "learning_rate": 7.97348484848485e-05,
      "loss": 0.0167,
      "step": 2675
    },
    {
      "epoch": 40.60606060606061,
      "grad_norm": 0.1474185287952423,
      "learning_rate": 7.96969696969697e-05,
      "loss": 0.0132,
      "step": 2680
    },
    {
      "epoch": 40.68181818181818,
      "grad_norm": 0.047357555478811264,
      "learning_rate": 7.965909090909091e-05,
      "loss": 0.0139,
      "step": 2685
    },
    {
      "epoch": 40.75757575757576,
      "grad_norm": 0.13994964957237244,
      "learning_rate": 7.962121212121213e-05,
      "loss": 0.0153,
      "step": 2690
    },
    {
      "epoch": 40.833333333333336,
      "grad_norm": 0.049286775290966034,
      "learning_rate": 7.958333333333333e-05,
      "loss": 0.0122,
      "step": 2695
    },
    {
      "epoch": 40.90909090909091,
      "grad_norm": 0.06891585141420364,
      "learning_rate": 7.954545454545455e-05,
      "loss": 0.0262,
      "step": 2700
    },
    {
      "epoch": 40.984848484848484,
      "grad_norm": 0.3179267644882202,
      "learning_rate": 7.950757575757575e-05,
      "loss": 0.0178,
      "step": 2705
    },
    {
      "epoch": 41.0,
      "eval_loss": 0.014667695388197899,
      "eval_mean_accuracy": 0.9547284888161127,
      "eval_mean_iou": 0.9152563030460374,
      "eval_runtime": 39.4797,
      "eval_samples_per_second": 5.952,
      "eval_steps_per_second": 0.76,
      "step": 2706
    },
    {
      "epoch": 41.06060606060606,
      "grad_norm": 1.7570785284042358,
      "learning_rate": 7.946969696969697e-05,
      "loss": 0.0159,
      "step": 2710
    },
    {
      "epoch": 41.13636363636363,
      "grad_norm": 0.26008275151252747,
      "learning_rate": 7.943181818181819e-05,
      "loss": 0.0164,
      "step": 2715
    },
    {
      "epoch": 41.21212121212121,
      "grad_norm": 0.10009404271841049,
      "learning_rate": 7.93939393939394e-05,
      "loss": 0.0129,
      "step": 2720
    },
    {
      "epoch": 41.28787878787879,
      "grad_norm": 0.253768652677536,
      "learning_rate": 7.93560606060606e-05,
      "loss": 0.0193,
      "step": 2725
    },
    {
      "epoch": 41.36363636363637,
      "grad_norm": 0.07129409909248352,
      "learning_rate": 7.931818181818182e-05,
      "loss": 0.0159,
      "step": 2730
    },
    {
      "epoch": 41.43939393939394,
      "grad_norm": 0.09366777539253235,
      "learning_rate": 7.928030303030304e-05,
      "loss": 0.0135,
      "step": 2735
    },
    {
      "epoch": 41.515151515151516,
      "grad_norm": 0.27111494541168213,
      "learning_rate": 7.924242424242426e-05,
      "loss": 0.0108,
      "step": 2740
    },
    {
      "epoch": 41.59090909090909,
      "grad_norm": 0.14143340289592743,
      "learning_rate": 7.920454545454546e-05,
      "loss": 0.0188,
      "step": 2745
    },
    {
      "epoch": 41.666666666666664,
      "grad_norm": 0.09004312753677368,
      "learning_rate": 7.916666666666666e-05,
      "loss": 0.0136,
      "step": 2750
    },
    {
      "epoch": 41.74242424242424,
      "grad_norm": 0.11193764209747314,
      "learning_rate": 7.912878787878788e-05,
      "loss": 0.0135,
      "step": 2755
    },
    {
      "epoch": 41.81818181818182,
      "grad_norm": 1.8563555479049683,
      "learning_rate": 7.90909090909091e-05,
      "loss": 0.0208,
      "step": 2760
    },
    {
      "epoch": 41.89393939393939,
      "grad_norm": 0.03630666434764862,
      "learning_rate": 7.905303030303031e-05,
      "loss": 0.0141,
      "step": 2765
    },
    {
      "epoch": 41.96969696969697,
      "grad_norm": 0.07599017024040222,
      "learning_rate": 7.901515151515151e-05,
      "loss": 0.014,
      "step": 2770
    },
    {
      "epoch": 42.0,
      "eval_loss": 0.014991636388003826,
      "eval_mean_accuracy": 0.9494945701784347,
      "eval_mean_iou": 0.9137321427167358,
      "eval_runtime": 42.0221,
      "eval_samples_per_second": 5.592,
      "eval_steps_per_second": 0.714,
      "step": 2772
    },
    {
      "epoch": 42.04545454545455,
      "grad_norm": 0.06971406936645508,
      "learning_rate": 7.897727272727273e-05,
      "loss": 0.0137,
      "step": 2775
    },
    {
      "epoch": 42.121212121212125,
      "grad_norm": 0.30499935150146484,
      "learning_rate": 7.893939393939395e-05,
      "loss": 0.0185,
      "step": 2780
    },
    {
      "epoch": 42.196969696969695,
      "grad_norm": 0.06412819027900696,
      "learning_rate": 7.890151515151516e-05,
      "loss": 0.0138,
      "step": 2785
    },
    {
      "epoch": 42.27272727272727,
      "grad_norm": 0.3736109137535095,
      "learning_rate": 7.886363636363637e-05,
      "loss": 0.0161,
      "step": 2790
    },
    {
      "epoch": 42.34848484848485,
      "grad_norm": 0.07162470370531082,
      "learning_rate": 7.882575757575757e-05,
      "loss": 0.0113,
      "step": 2795
    },
    {
      "epoch": 42.42424242424242,
      "grad_norm": 0.05985365808010101,
      "learning_rate": 7.878787878787879e-05,
      "loss": 0.0142,
      "step": 2800
    },
    {
      "epoch": 42.5,
      "grad_norm": 0.253432035446167,
      "learning_rate": 7.875e-05,
      "loss": 0.0119,
      "step": 2805
    },
    {
      "epoch": 42.57575757575758,
      "grad_norm": 0.11161007732152939,
      "learning_rate": 7.871212121212122e-05,
      "loss": 0.0145,
      "step": 2810
    },
    {
      "epoch": 42.65151515151515,
      "grad_norm": 0.08115492016077042,
      "learning_rate": 7.867424242424242e-05,
      "loss": 0.0132,
      "step": 2815
    },
    {
      "epoch": 42.72727272727273,
      "grad_norm": 0.16690608859062195,
      "learning_rate": 7.863636363636364e-05,
      "loss": 0.0161,
      "step": 2820
    },
    {
      "epoch": 42.803030303030305,
      "grad_norm": 0.05905652046203613,
      "learning_rate": 7.859848484848486e-05,
      "loss": 0.0143,
      "step": 2825
    },
    {
      "epoch": 42.878787878787875,
      "grad_norm": 0.0785427838563919,
      "learning_rate": 7.856060606060607e-05,
      "loss": 0.0144,
      "step": 2830
    },
    {
      "epoch": 42.95454545454545,
      "grad_norm": 0.06213955208659172,
      "learning_rate": 7.852272727272728e-05,
      "loss": 0.0212,
      "step": 2835
    },
    {
      "epoch": 43.0,
      "eval_loss": 0.014552533626556396,
      "eval_mean_accuracy": 0.9509986163655171,
      "eval_mean_iou": 0.9159469813361092,
      "eval_runtime": 38.9568,
      "eval_samples_per_second": 6.032,
      "eval_steps_per_second": 0.77,
      "step": 2838
    },
    {
      "epoch": 43.03030303030303,
      "grad_norm": 0.14538201689720154,
      "learning_rate": 7.848484848484848e-05,
      "loss": 0.0154,
      "step": 2840
    },
    {
      "epoch": 43.10606060606061,
      "grad_norm": 0.1533840298652649,
      "learning_rate": 7.84469696969697e-05,
      "loss": 0.0125,
      "step": 2845
    },
    {
      "epoch": 43.18181818181818,
      "grad_norm": 0.1251472532749176,
      "learning_rate": 7.840909090909091e-05,
      "loss": 0.0139,
      "step": 2850
    },
    {
      "epoch": 43.25757575757576,
      "grad_norm": 0.11311665922403336,
      "learning_rate": 7.837121212121213e-05,
      "loss": 0.0107,
      "step": 2855
    },
    {
      "epoch": 43.333333333333336,
      "grad_norm": 0.052770040929317474,
      "learning_rate": 7.833333333333333e-05,
      "loss": 0.0128,
      "step": 2860
    },
    {
      "epoch": 43.40909090909091,
      "grad_norm": 0.06932497769594193,
      "learning_rate": 7.829545454545455e-05,
      "loss": 0.0149,
      "step": 2865
    },
    {
      "epoch": 43.484848484848484,
      "grad_norm": 0.056223709136247635,
      "learning_rate": 7.825757575757576e-05,
      "loss": 0.0141,
      "step": 2870
    },
    {
      "epoch": 43.56060606060606,
      "grad_norm": 0.06966979801654816,
      "learning_rate": 7.821969696969698e-05,
      "loss": 0.0129,
      "step": 2875
    },
    {
      "epoch": 43.63636363636363,
      "grad_norm": 0.12024807184934616,
      "learning_rate": 7.818181818181818e-05,
      "loss": 0.0133,
      "step": 2880
    },
    {
      "epoch": 43.71212121212121,
      "grad_norm": 0.07419823855161667,
      "learning_rate": 7.81439393939394e-05,
      "loss": 0.014,
      "step": 2885
    },
    {
      "epoch": 43.78787878787879,
      "grad_norm": 0.08092731237411499,
      "learning_rate": 7.81060606060606e-05,
      "loss": 0.0144,
      "step": 2890
    },
    {
      "epoch": 43.86363636363637,
      "grad_norm": 0.03666728362441063,
      "learning_rate": 7.806818181818182e-05,
      "loss": 0.0143,
      "step": 2895
    },
    {
      "epoch": 43.93939393939394,
      "grad_norm": 0.05808990076184273,
      "learning_rate": 7.803030303030304e-05,
      "loss": 0.0158,
      "step": 2900
    },
    {
      "epoch": 44.0,
      "eval_loss": 0.01385864894837141,
      "eval_mean_accuracy": 0.9527755723915611,
      "eval_mean_iou": 0.9178972684359848,
      "eval_runtime": 37.475,
      "eval_samples_per_second": 6.271,
      "eval_steps_per_second": 0.801,
      "step": 2904
    },
    {
      "epoch": 44.015151515151516,
      "grad_norm": 0.05168011039495468,
      "learning_rate": 7.799242424242424e-05,
      "loss": 0.0122,
      "step": 2905
    },
    {
      "epoch": 44.09090909090909,
      "grad_norm": 0.03633609414100647,
      "learning_rate": 7.795454545454546e-05,
      "loss": 0.0091,
      "step": 2910
    },
    {
      "epoch": 44.166666666666664,
      "grad_norm": 0.12624555826187134,
      "learning_rate": 7.791666666666667e-05,
      "loss": 0.0143,
      "step": 2915
    },
    {
      "epoch": 44.24242424242424,
      "grad_norm": 0.1368200182914734,
      "learning_rate": 7.787878787878789e-05,
      "loss": 0.0108,
      "step": 2920
    },
    {
      "epoch": 44.31818181818182,
      "grad_norm": 0.05255744606256485,
      "learning_rate": 7.784090909090909e-05,
      "loss": 0.0132,
      "step": 2925
    },
    {
      "epoch": 44.39393939393939,
      "grad_norm": 0.10443238914012909,
      "learning_rate": 7.780303030303031e-05,
      "loss": 0.0124,
      "step": 2930
    },
    {
      "epoch": 44.46969696969697,
      "grad_norm": 0.042250387370586395,
      "learning_rate": 7.776515151515152e-05,
      "loss": 0.0146,
      "step": 2935
    },
    {
      "epoch": 44.54545454545455,
      "grad_norm": 0.052224092185497284,
      "learning_rate": 7.772727272727273e-05,
      "loss": 0.0153,
      "step": 2940
    },
    {
      "epoch": 44.621212121212125,
      "grad_norm": 0.16634100675582886,
      "learning_rate": 7.768939393939394e-05,
      "loss": 0.018,
      "step": 2945
    },
    {
      "epoch": 44.696969696969695,
      "grad_norm": 0.06778372079133987,
      "learning_rate": 7.765151515151515e-05,
      "loss": 0.0117,
      "step": 2950
    },
    {
      "epoch": 44.77272727272727,
      "grad_norm": 0.09640312939882278,
      "learning_rate": 7.761363636363636e-05,
      "loss": 0.0154,
      "step": 2955
    },
    {
      "epoch": 44.84848484848485,
      "grad_norm": 0.04297403618693352,
      "learning_rate": 7.757575757575758e-05,
      "loss": 0.0125,
      "step": 2960
    },
    {
      "epoch": 44.92424242424242,
      "grad_norm": 0.09395831823348999,
      "learning_rate": 7.75378787878788e-05,
      "loss": 0.0122,
      "step": 2965
    },
    {
      "epoch": 45.0,
      "grad_norm": 0.11690888553857803,
      "learning_rate": 7.75e-05,
      "loss": 0.0105,
      "step": 2970
    },
    {
      "epoch": 45.0,
      "eval_loss": 0.013977193273603916,
      "eval_mean_accuracy": 0.9404354087479972,
      "eval_mean_iou": 0.9144762454194502,
      "eval_runtime": 37.8216,
      "eval_samples_per_second": 6.213,
      "eval_steps_per_second": 0.793,
      "step": 2970
    },
    {
      "epoch": 45.07575757575758,
      "grad_norm": 0.07428726553916931,
      "learning_rate": 7.746212121212122e-05,
      "loss": 0.0139,
      "step": 2975
    },
    {
      "epoch": 45.15151515151515,
      "grad_norm": 0.042812131345272064,
      "learning_rate": 7.742424242424243e-05,
      "loss": 0.0094,
      "step": 2980
    },
    {
      "epoch": 45.22727272727273,
      "grad_norm": 0.09956932067871094,
      "learning_rate": 7.738636363636364e-05,
      "loss": 0.014,
      "step": 2985
    },
    {
      "epoch": 45.303030303030305,
      "grad_norm": 0.0540524423122406,
      "learning_rate": 7.734848484848485e-05,
      "loss": 0.0137,
      "step": 2990
    },
    {
      "epoch": 45.378787878787875,
      "grad_norm": 0.10315060615539551,
      "learning_rate": 7.731060606060606e-05,
      "loss": 0.011,
      "step": 2995
    },
    {
      "epoch": 45.45454545454545,
      "grad_norm": 0.08139504492282867,
      "learning_rate": 7.727272727272727e-05,
      "loss": 0.0152,
      "step": 3000
    },
    {
      "epoch": 45.53030303030303,
      "grad_norm": 0.07393460720777512,
      "learning_rate": 7.723484848484849e-05,
      "loss": 0.0126,
      "step": 3005
    },
    {
      "epoch": 45.60606060606061,
      "grad_norm": 0.039620134979486465,
      "learning_rate": 7.71969696969697e-05,
      "loss": 0.0108,
      "step": 3010
    },
    {
      "epoch": 45.68181818181818,
      "grad_norm": 0.03623414784669876,
      "learning_rate": 7.715909090909091e-05,
      "loss": 0.0116,
      "step": 3015
    },
    {
      "epoch": 45.75757575757576,
      "grad_norm": 0.09395652264356613,
      "learning_rate": 7.712121212121212e-05,
      "loss": 0.0136,
      "step": 3020
    },
    {
      "epoch": 45.833333333333336,
      "grad_norm": 0.05183081328868866,
      "learning_rate": 7.708333333333334e-05,
      "loss": 0.0093,
      "step": 3025
    },
    {
      "epoch": 45.90909090909091,
      "grad_norm": 0.03278835117816925,
      "learning_rate": 7.704545454545456e-05,
      "loss": 0.014,
      "step": 3030
    },
    {
      "epoch": 45.984848484848484,
      "grad_norm": 0.08692388236522675,
      "learning_rate": 7.700757575757576e-05,
      "loss": 0.0152,
      "step": 3035
    },
    {
      "epoch": 46.0,
      "eval_loss": 0.013542413711547852,
      "eval_mean_accuracy": 0.9444993175052597,
      "eval_mean_iou": 0.9163730968281026,
      "eval_runtime": 39.5385,
      "eval_samples_per_second": 5.944,
      "eval_steps_per_second": 0.759,
      "step": 3036
    },
    {
      "epoch": 46.06060606060606,
      "grad_norm": 0.10967297852039337,
      "learning_rate": 7.696969696969696e-05,
      "loss": 0.018,
      "step": 3040
    },
    {
      "epoch": 46.13636363636363,
      "grad_norm": 0.13117849826812744,
      "learning_rate": 7.693181818181818e-05,
      "loss": 0.013,
      "step": 3045
    },
    {
      "epoch": 46.21212121212121,
      "grad_norm": 0.03816286474466324,
      "learning_rate": 7.68939393939394e-05,
      "loss": 0.0139,
      "step": 3050
    },
    {
      "epoch": 46.28787878787879,
      "grad_norm": 0.04810022562742233,
      "learning_rate": 7.685606060606061e-05,
      "loss": 0.0124,
      "step": 3055
    },
    {
      "epoch": 46.36363636363637,
      "grad_norm": 0.04847066104412079,
      "learning_rate": 7.681818181818182e-05,
      "loss": 0.0147,
      "step": 3060
    },
    {
      "epoch": 46.43939393939394,
      "grad_norm": 0.08611077070236206,
      "learning_rate": 7.678030303030303e-05,
      "loss": 0.0113,
      "step": 3065
    },
    {
      "epoch": 46.515151515151516,
      "grad_norm": 0.05069834738969803,
      "learning_rate": 7.674242424242425e-05,
      "loss": 0.0122,
      "step": 3070
    },
    {
      "epoch": 46.59090909090909,
      "grad_norm": 0.07546529918909073,
      "learning_rate": 7.670454545454547e-05,
      "loss": 0.013,
      "step": 3075
    },
    {
      "epoch": 46.666666666666664,
      "grad_norm": 0.09206637740135193,
      "learning_rate": 7.666666666666667e-05,
      "loss": 0.0135,
      "step": 3080
    },
    {
      "epoch": 46.74242424242424,
      "grad_norm": 0.06238163635134697,
      "learning_rate": 7.662878787878787e-05,
      "loss": 0.0092,
      "step": 3085
    },
    {
      "epoch": 46.81818181818182,
      "grad_norm": 0.0814301148056984,
      "learning_rate": 7.659090909090909e-05,
      "loss": 0.0111,
      "step": 3090
    },
    {
      "epoch": 46.89393939393939,
      "grad_norm": 0.12262319028377533,
      "learning_rate": 7.65530303030303e-05,
      "loss": 0.0125,
      "step": 3095
    },
    {
      "epoch": 46.96969696969697,
      "grad_norm": 0.20744876563549042,
      "learning_rate": 7.651515151515152e-05,
      "loss": 0.0233,
      "step": 3100
    },
    {
      "epoch": 47.0,
      "eval_loss": 0.01413767971098423,
      "eval_mean_accuracy": 0.948157938331392,
      "eval_mean_iou": 0.9171055514739753,
      "eval_runtime": 39.8776,
      "eval_samples_per_second": 5.893,
      "eval_steps_per_second": 0.752,
      "step": 3102
    },
    {
      "epoch": 47.04545454545455,
      "grad_norm": 0.08193027228116989,
      "learning_rate": 7.647727272727272e-05,
      "loss": 0.018,
      "step": 3105
    },
    {
      "epoch": 47.121212121212125,
      "grad_norm": 0.27056941390037537,
      "learning_rate": 7.643939393939394e-05,
      "loss": 0.0135,
      "step": 3110
    },
    {
      "epoch": 47.196969696969695,
      "grad_norm": 0.057886213064193726,
      "learning_rate": 7.640151515151516e-05,
      "loss": 0.0122,
      "step": 3115
    },
    {
      "epoch": 47.27272727272727,
      "grad_norm": 0.04713014140725136,
      "learning_rate": 7.636363636363637e-05,
      "loss": 0.0134,
      "step": 3120
    },
    {
      "epoch": 47.34848484848485,
      "grad_norm": 0.09342219680547714,
      "learning_rate": 7.632575757575758e-05,
      "loss": 0.0126,
      "step": 3125
    },
    {
      "epoch": 47.42424242424242,
      "grad_norm": 0.030409380793571472,
      "learning_rate": 7.62878787878788e-05,
      "loss": 0.0164,
      "step": 3130
    },
    {
      "epoch": 47.5,
      "grad_norm": 0.16618812084197998,
      "learning_rate": 7.625e-05,
      "loss": 0.015,
      "step": 3135
    },
    {
      "epoch": 47.57575757575758,
      "grad_norm": 0.09027186036109924,
      "learning_rate": 7.621212121212121e-05,
      "loss": 0.0107,
      "step": 3140
    },
    {
      "epoch": 47.65151515151515,
      "grad_norm": 0.02703385055065155,
      "learning_rate": 7.617424242424243e-05,
      "loss": 0.0139,
      "step": 3145
    },
    {
      "epoch": 47.72727272727273,
      "grad_norm": 0.0698995441198349,
      "learning_rate": 7.613636363636363e-05,
      "loss": 0.012,
      "step": 3150
    },
    {
      "epoch": 47.803030303030305,
      "grad_norm": 0.1583048403263092,
      "learning_rate": 7.609848484848485e-05,
      "loss": 0.0137,
      "step": 3155
    },
    {
      "epoch": 47.878787878787875,
      "grad_norm": 0.0941573902964592,
      "learning_rate": 7.606060606060607e-05,
      "loss": 0.011,
      "step": 3160
    },
    {
      "epoch": 47.95454545454545,
      "grad_norm": 0.09458589553833008,
      "learning_rate": 7.602272727272728e-05,
      "loss": 0.0141,
      "step": 3165
    },
    {
      "epoch": 48.0,
      "eval_loss": 0.013407318852841854,
      "eval_mean_accuracy": 0.9545140328913648,
      "eval_mean_iou": 0.9198645138020445,
      "eval_runtime": 42.3635,
      "eval_samples_per_second": 5.547,
      "eval_steps_per_second": 0.708,
      "step": 3168
    },
    {
      "epoch": 48.03030303030303,
      "grad_norm": 0.08197533339262009,
      "learning_rate": 7.598484848484849e-05,
      "loss": 0.013,
      "step": 3170
    },
    {
      "epoch": 48.10606060606061,
      "grad_norm": 0.03099500946700573,
      "learning_rate": 7.59469696969697e-05,
      "loss": 0.0127,
      "step": 3175
    },
    {
      "epoch": 48.18181818181818,
      "grad_norm": 0.05799512192606926,
      "learning_rate": 7.59090909090909e-05,
      "loss": 0.0103,
      "step": 3180
    },
    {
      "epoch": 48.25757575757576,
      "grad_norm": 0.08802284300327301,
      "learning_rate": 7.587121212121212e-05,
      "loss": 0.0165,
      "step": 3185
    },
    {
      "epoch": 48.333333333333336,
      "grad_norm": 0.16304674744606018,
      "learning_rate": 7.583333333333334e-05,
      "loss": 0.0126,
      "step": 3190
    },
    {
      "epoch": 48.40909090909091,
      "grad_norm": 0.1502985805273056,
      "learning_rate": 7.579545454545454e-05,
      "loss": 0.013,
      "step": 3195
    },
    {
      "epoch": 48.484848484848484,
      "grad_norm": 0.057666897773742676,
      "learning_rate": 7.575757575757576e-05,
      "loss": 0.0104,
      "step": 3200
    },
    {
      "epoch": 48.56060606060606,
      "grad_norm": 0.03507477790117264,
      "learning_rate": 7.571969696969697e-05,
      "loss": 0.0124,
      "step": 3205
    },
    {
      "epoch": 48.63636363636363,
      "grad_norm": 0.05581803247332573,
      "learning_rate": 7.568181818181819e-05,
      "loss": 0.0158,
      "step": 3210
    },
    {
      "epoch": 48.71212121212121,
      "grad_norm": 0.15561729669570923,
      "learning_rate": 7.56439393939394e-05,
      "loss": 0.0162,
      "step": 3215
    },
    {
      "epoch": 48.78787878787879,
      "grad_norm": 0.10783232748508453,
      "learning_rate": 7.560606060606061e-05,
      "loss": 0.0136,
      "step": 3220
    },
    {
      "epoch": 48.86363636363637,
      "grad_norm": 0.08476749062538147,
      "learning_rate": 7.556818181818183e-05,
      "loss": 0.0117,
      "step": 3225
    },
    {
      "epoch": 48.93939393939394,
      "grad_norm": 0.07583211362361908,
      "learning_rate": 7.553030303030303e-05,
      "loss": 0.0113,
      "step": 3230
    },
    {
      "epoch": 49.0,
      "eval_loss": 0.013585193082690239,
      "eval_mean_accuracy": 0.9569897308920382,
      "eval_mean_iou": 0.9196594685573602,
      "eval_runtime": 36.7017,
      "eval_samples_per_second": 6.403,
      "eval_steps_per_second": 0.817,
      "step": 3234
    },
    {
      "epoch": 49.015151515151516,
      "grad_norm": 0.07685961574316025,
      "learning_rate": 7.549242424242425e-05,
      "loss": 0.0145,
      "step": 3235
    },
    {
      "epoch": 49.09090909090909,
      "grad_norm": 0.04889879375696182,
      "learning_rate": 7.545454545454545e-05,
      "loss": 0.0139,
      "step": 3240
    },
    {
      "epoch": 49.166666666666664,
      "grad_norm": 0.04259079694747925,
      "learning_rate": 7.541666666666667e-05,
      "loss": 0.0099,
      "step": 3245
    },
    {
      "epoch": 49.24242424242424,
      "grad_norm": 0.04860289767384529,
      "learning_rate": 7.537878787878788e-05,
      "loss": 0.013,
      "step": 3250
    },
    {
      "epoch": 49.31818181818182,
      "grad_norm": 0.04036470502614975,
      "learning_rate": 7.53409090909091e-05,
      "loss": 0.0145,
      "step": 3255
    },
    {
      "epoch": 49.39393939393939,
      "grad_norm": 0.13199807703495026,
      "learning_rate": 7.530303030303032e-05,
      "loss": 0.012,
      "step": 3260
    },
    {
      "epoch": 49.46969696969697,
      "grad_norm": 0.1777004897594452,
      "learning_rate": 7.526515151515152e-05,
      "loss": 0.0161,
      "step": 3265
    },
    {
      "epoch": 49.54545454545455,
      "grad_norm": 0.031824395060539246,
      "learning_rate": 7.522727272727273e-05,
      "loss": 0.0092,
      "step": 3270
    },
    {
      "epoch": 49.621212121212125,
      "grad_norm": 0.0718945860862732,
      "learning_rate": 7.518939393939395e-05,
      "loss": 0.0138,
      "step": 3275
    },
    {
      "epoch": 49.696969696969695,
      "grad_norm": 0.039252229034900665,
      "learning_rate": 7.515151515151515e-05,
      "loss": 0.0133,
      "step": 3280
    },
    {
      "epoch": 49.77272727272727,
      "grad_norm": 0.08561480790376663,
      "learning_rate": 7.511363636363636e-05,
      "loss": 0.0107,
      "step": 3285
    },
    {
      "epoch": 49.84848484848485,
      "grad_norm": 0.22405771911144257,
      "learning_rate": 7.507575757575757e-05,
      "loss": 0.0128,
      "step": 3290
    },
    {
      "epoch": 49.92424242424242,
      "grad_norm": 0.0903986319899559,
      "learning_rate": 7.503787878787879e-05,
      "loss": 0.012,
      "step": 3295
    },
    {
      "epoch": 50.0,
      "grad_norm": 0.29394635558128357,
      "learning_rate": 7.500000000000001e-05,
      "loss": 0.0141,
      "step": 3300
    },
    {
      "epoch": 50.0,
      "eval_loss": 0.013239020481705666,
      "eval_mean_accuracy": 0.948916308860805,
      "eval_mean_iou": 0.919184769920378,
      "eval_runtime": 36.5287,
      "eval_samples_per_second": 6.433,
      "eval_steps_per_second": 0.821,
      "step": 3300
    },
    {
      "epoch": 50.07575757575758,
      "grad_norm": 0.054516732692718506,
      "learning_rate": 7.496212121212122e-05,
      "loss": 0.0102,
      "step": 3305
    },
    {
      "epoch": 50.15151515151515,
      "grad_norm": 0.06817471235990524,
      "learning_rate": 7.492424242424243e-05,
      "loss": 0.0118,
      "step": 3310
    },
    {
      "epoch": 50.22727272727273,
      "grad_norm": 0.05059773474931717,
      "learning_rate": 7.488636363636364e-05,
      "loss": 0.0121,
      "step": 3315
    },
    {
      "epoch": 50.303030303030305,
      "grad_norm": 0.029846513643860817,
      "learning_rate": 7.484848484848486e-05,
      "loss": 0.0121,
      "step": 3320
    },
    {
      "epoch": 50.378787878787875,
      "grad_norm": 0.046205855906009674,
      "learning_rate": 7.481060606060606e-05,
      "loss": 0.0115,
      "step": 3325
    },
    {
      "epoch": 50.45454545454545,
      "grad_norm": 0.024707447737455368,
      "learning_rate": 7.477272727272727e-05,
      "loss": 0.0094,
      "step": 3330
    },
    {
      "epoch": 50.53030303030303,
      "grad_norm": 0.02938036061823368,
      "learning_rate": 7.473484848484848e-05,
      "loss": 0.0102,
      "step": 3335
    },
    {
      "epoch": 50.60606060606061,
      "grad_norm": 0.05755782872438431,
      "learning_rate": 7.46969696969697e-05,
      "loss": 0.0164,
      "step": 3340
    },
    {
      "epoch": 50.68181818181818,
      "grad_norm": 0.045329637825489044,
      "learning_rate": 7.465909090909092e-05,
      "loss": 0.0108,
      "step": 3345
    },
    {
      "epoch": 50.75757575757576,
      "grad_norm": 0.07368195801973343,
      "learning_rate": 7.462121212121213e-05,
      "loss": 0.0137,
      "step": 3350
    },
    {
      "epoch": 50.833333333333336,
      "grad_norm": 0.06482062488794327,
      "learning_rate": 7.458333333333333e-05,
      "loss": 0.0169,
      "step": 3355
    },
    {
      "epoch": 50.90909090909091,
      "grad_norm": 0.11054930090904236,
      "learning_rate": 7.454545454545455e-05,
      "loss": 0.0136,
      "step": 3360
    },
    {
      "epoch": 50.984848484848484,
      "grad_norm": 0.08624602109193802,
      "learning_rate": 7.450757575757577e-05,
      "loss": 0.0142,
      "step": 3365
    },
    {
      "epoch": 51.0,
      "eval_loss": 0.013167026452720165,
      "eval_mean_accuracy": 0.9472601118732853,
      "eval_mean_iou": 0.9185300185502054,
      "eval_runtime": 35.2488,
      "eval_samples_per_second": 6.667,
      "eval_steps_per_second": 0.851,
      "step": 3366
    },
    {
      "epoch": 51.06060606060606,
      "grad_norm": 0.05143900588154793,
      "learning_rate": 7.446969696969698e-05,
      "loss": 0.0157,
      "step": 3370
    },
    {
      "epoch": 51.13636363636363,
      "grad_norm": 0.04773813858628273,
      "learning_rate": 7.443181818181817e-05,
      "loss": 0.0143,
      "step": 3375
    },
    {
      "epoch": 51.21212121212121,
      "grad_norm": 0.04963133484125137,
      "learning_rate": 7.439393939393939e-05,
      "loss": 0.0103,
      "step": 3380
    },
    {
      "epoch": 51.28787878787879,
      "grad_norm": 0.05506278946995735,
      "learning_rate": 7.435606060606061e-05,
      "loss": 0.012,
      "step": 3385
    },
    {
      "epoch": 51.36363636363637,
      "grad_norm": 0.06806264817714691,
      "learning_rate": 7.431818181818182e-05,
      "loss": 0.0285,
      "step": 3390
    },
    {
      "epoch": 51.43939393939394,
      "grad_norm": 0.025749368593096733,
      "learning_rate": 7.428030303030304e-05,
      "loss": 0.0127,
      "step": 3395
    },
    {
      "epoch": 51.515151515151516,
      "grad_norm": 0.1990114003419876,
      "learning_rate": 7.424242424242424e-05,
      "loss": 0.0117,
      "step": 3400
    },
    {
      "epoch": 51.59090909090909,
      "grad_norm": 0.13872888684272766,
      "learning_rate": 7.420454545454546e-05,
      "loss": 0.0687,
      "step": 3405
    },
    {
      "epoch": 51.666666666666664,
      "grad_norm": 0.0722043514251709,
      "learning_rate": 7.416666666666668e-05,
      "loss": 0.0194,
      "step": 3410
    },
    {
      "epoch": 51.74242424242424,
      "grad_norm": 0.036345645785331726,
      "learning_rate": 7.412878787878789e-05,
      "loss": 0.0283,
      "step": 3415
    },
    {
      "epoch": 51.81818181818182,
      "grad_norm": 0.05883508175611496,
      "learning_rate": 7.40909090909091e-05,
      "loss": 0.0175,
      "step": 3420
    },
    {
      "epoch": 51.89393939393939,
      "grad_norm": 0.12866470217704773,
      "learning_rate": 7.40530303030303e-05,
      "loss": 0.0125,
      "step": 3425
    },
    {
      "epoch": 51.96969696969697,
      "grad_norm": 0.07310930639505386,
      "learning_rate": 7.401515151515152e-05,
      "loss": 0.012,
      "step": 3430
    },
    {
      "epoch": 52.0,
      "eval_loss": 0.015447704121470451,
      "eval_mean_accuracy": 0.9607018323894033,
      "eval_mean_iou": 0.9130043906987909,
      "eval_runtime": 41.3421,
      "eval_samples_per_second": 5.684,
      "eval_steps_per_second": 0.726,
      "step": 3432
    },
    {
      "epoch": 52.04545454545455,
      "grad_norm": 0.10036005079746246,
      "learning_rate": 7.397727272727273e-05,
      "loss": 0.0103,
      "step": 3435
    },
    {
      "epoch": 52.121212121212125,
      "grad_norm": 0.11839846521615982,
      "learning_rate": 7.393939393939395e-05,
      "loss": 0.0154,
      "step": 3440
    },
    {
      "epoch": 52.196969696969695,
      "grad_norm": 0.05460628867149353,
      "learning_rate": 7.390151515151515e-05,
      "loss": 0.0161,
      "step": 3445
    },
    {
      "epoch": 52.27272727272727,
      "grad_norm": 0.17195387184619904,
      "learning_rate": 7.386363636363637e-05,
      "loss": 0.0161,
      "step": 3450
    },
    {
      "epoch": 52.34848484848485,
      "grad_norm": 0.11281295120716095,
      "learning_rate": 7.382575757575758e-05,
      "loss": 0.0237,
      "step": 3455
    },
    {
      "epoch": 52.42424242424242,
      "grad_norm": 0.09658186882734299,
      "learning_rate": 7.37878787878788e-05,
      "loss": 0.0104,
      "step": 3460
    },
    {
      "epoch": 52.5,
      "grad_norm": 0.04379212111234665,
      "learning_rate": 7.375e-05,
      "loss": 0.0133,
      "step": 3465
    },
    {
      "epoch": 52.57575757575758,
      "grad_norm": 0.07511848956346512,
      "learning_rate": 7.37121212121212e-05,
      "loss": 0.0147,
      "step": 3470
    },
    {
      "epoch": 52.65151515151515,
      "grad_norm": 0.06669140607118607,
      "learning_rate": 7.367424242424242e-05,
      "loss": 0.0136,
      "step": 3475
    },
    {
      "epoch": 52.72727272727273,
      "grad_norm": 1.4316908121109009,
      "learning_rate": 7.363636363636364e-05,
      "loss": 0.0148,
      "step": 3480
    },
    {
      "epoch": 52.803030303030305,
      "grad_norm": 0.05004597455263138,
      "learning_rate": 7.359848484848486e-05,
      "loss": 0.0139,
      "step": 3485
    },
    {
      "epoch": 52.878787878787875,
      "grad_norm": 0.05055457353591919,
      "learning_rate": 7.356060606060606e-05,
      "loss": 0.0168,
      "step": 3490
    },
    {
      "epoch": 52.95454545454545,
      "grad_norm": 0.21981512010097504,
      "learning_rate": 7.352272727272728e-05,
      "loss": 0.0133,
      "step": 3495
    },
    {
      "epoch": 53.0,
      "eval_loss": 0.014284273609519005,
      "eval_mean_accuracy": 0.9476361224957444,
      "eval_mean_iou": 0.9156907043774533,
      "eval_runtime": 42.798,
      "eval_samples_per_second": 5.491,
      "eval_steps_per_second": 0.701,
      "step": 3498
    },
    {
      "epoch": 53.03030303030303,
      "grad_norm": 0.05711522325873375,
      "learning_rate": 7.348484848484849e-05,
      "loss": 0.0103,
      "step": 3500
    },
    {
      "epoch": 53.10606060606061,
      "grad_norm": 0.04860633239150047,
      "learning_rate": 7.344696969696971e-05,
      "loss": 0.0112,
      "step": 3505
    },
    {
      "epoch": 53.18181818181818,
      "grad_norm": 0.12113384902477264,
      "learning_rate": 7.340909090909091e-05,
      "loss": 0.0136,
      "step": 3510
    },
    {
      "epoch": 53.25757575757576,
      "grad_norm": 0.06631729751825333,
      "learning_rate": 7.337121212121213e-05,
      "loss": 0.0099,
      "step": 3515
    },
    {
      "epoch": 53.333333333333336,
      "grad_norm": 0.1377585530281067,
      "learning_rate": 7.333333333333333e-05,
      "loss": 0.0142,
      "step": 3520
    },
    {
      "epoch": 53.40909090909091,
      "grad_norm": 0.06398643553256989,
      "learning_rate": 7.329545454545455e-05,
      "loss": 0.0111,
      "step": 3525
    },
    {
      "epoch": 53.484848484848484,
      "grad_norm": 0.10173145681619644,
      "learning_rate": 7.325757575757576e-05,
      "loss": 0.0134,
      "step": 3530
    },
    {
      "epoch": 53.56060606060606,
      "grad_norm": 0.19841890037059784,
      "learning_rate": 7.321969696969697e-05,
      "loss": 0.0169,
      "step": 3535
    },
    {
      "epoch": 53.63636363636363,
      "grad_norm": 0.09498052299022675,
      "learning_rate": 7.318181818181818e-05,
      "loss": 0.0131,
      "step": 3540
    },
    {
      "epoch": 53.71212121212121,
      "grad_norm": 0.09234619140625,
      "learning_rate": 7.31439393939394e-05,
      "loss": 0.0143,
      "step": 3545
    },
    {
      "epoch": 53.78787878787879,
      "grad_norm": 0.163950115442276,
      "learning_rate": 7.310606060606062e-05,
      "loss": 0.0151,
      "step": 3550
    },
    {
      "epoch": 53.86363636363637,
      "grad_norm": 0.7671867609024048,
      "learning_rate": 7.306818181818182e-05,
      "loss": 0.0161,
      "step": 3555
    },
    {
      "epoch": 53.93939393939394,
      "grad_norm": 0.028596986085176468,
      "learning_rate": 7.303030303030304e-05,
      "loss": 0.0089,
      "step": 3560
    },
    {
      "epoch": 54.0,
      "eval_loss": 0.013690474443137646,
      "eval_mean_accuracy": 0.9477638247186114,
      "eval_mean_iou": 0.9171720989120438,
      "eval_runtime": 36.9675,
      "eval_samples_per_second": 6.357,
      "eval_steps_per_second": 0.812,
      "step": 3564
    },
    {
      "epoch": 54.015151515151516,
      "grad_norm": 0.04326793551445007,
      "learning_rate": 7.299242424242425e-05,
      "loss": 0.0122,
      "step": 3565
    },
    {
      "epoch": 54.09090909090909,
      "grad_norm": 0.056759435683488846,
      "learning_rate": 7.295454545454546e-05,
      "loss": 0.0104,
      "step": 3570
    },
    {
      "epoch": 54.166666666666664,
      "grad_norm": 0.02118513733148575,
      "learning_rate": 7.291666666666667e-05,
      "loss": 0.0119,
      "step": 3575
    },
    {
      "epoch": 54.24242424242424,
      "grad_norm": 0.03992696478962898,
      "learning_rate": 7.287878787878788e-05,
      "loss": 0.0137,
      "step": 3580
    },
    {
      "epoch": 54.31818181818182,
      "grad_norm": 0.05451955646276474,
      "learning_rate": 7.284090909090909e-05,
      "loss": 0.0147,
      "step": 3585
    },
    {
      "epoch": 54.39393939393939,
      "grad_norm": 0.1193266436457634,
      "learning_rate": 7.280303030303031e-05,
      "loss": 0.0124,
      "step": 3590
    },
    {
      "epoch": 54.46969696969697,
      "grad_norm": 0.07213295996189117,
      "learning_rate": 7.276515151515153e-05,
      "loss": 0.015,
      "step": 3595
    },
    {
      "epoch": 54.54545454545455,
      "grad_norm": 0.07287770509719849,
      "learning_rate": 7.272727272727273e-05,
      "loss": 0.0144,
      "step": 3600
    },
    {
      "epoch": 54.621212121212125,
      "grad_norm": 0.06693650782108307,
      "learning_rate": 7.268939393939394e-05,
      "loss": 0.0121,
      "step": 3605
    },
    {
      "epoch": 54.696969696969695,
      "grad_norm": 0.024828793480992317,
      "learning_rate": 7.265151515151516e-05,
      "loss": 0.0144,
      "step": 3610
    },
    {
      "epoch": 54.77272727272727,
      "grad_norm": 0.035814329981803894,
      "learning_rate": 7.261363636363638e-05,
      "loss": 0.0113,
      "step": 3615
    },
    {
      "epoch": 54.84848484848485,
      "grad_norm": 0.08699552714824677,
      "learning_rate": 7.257575757575758e-05,
      "loss": 0.0127,
      "step": 3620
    },
    {
      "epoch": 54.92424242424242,
      "grad_norm": 0.042220618575811386,
      "learning_rate": 7.253787878787878e-05,
      "loss": 0.0114,
      "step": 3625
    },
    {
      "epoch": 55.0,
      "grad_norm": 0.03382052853703499,
      "learning_rate": 7.25e-05,
      "loss": 0.0115,
      "step": 3630
    },
    {
      "epoch": 55.0,
      "eval_loss": 0.013277423568069935,
      "eval_mean_accuracy": 0.9454677261881846,
      "eval_mean_iou": 0.9172010772990417,
      "eval_runtime": 37.135,
      "eval_samples_per_second": 6.328,
      "eval_steps_per_second": 0.808,
      "step": 3630
    },
    {
      "epoch": 55.07575757575758,
      "grad_norm": 0.054507989436388016,
      "learning_rate": 7.246212121212122e-05,
      "loss": 0.0132,
      "step": 3635
    },
    {
      "epoch": 55.15151515151515,
      "grad_norm": 0.04291966184973717,
      "learning_rate": 7.242424242424243e-05,
      "loss": 0.0148,
      "step": 3640
    },
    {
      "epoch": 55.22727272727273,
      "grad_norm": 0.11435730755329132,
      "learning_rate": 7.238636363636364e-05,
      "loss": 0.0123,
      "step": 3645
    },
    {
      "epoch": 55.303030303030305,
      "grad_norm": 0.050386372953653336,
      "learning_rate": 7.234848484848485e-05,
      "loss": 0.0094,
      "step": 3650
    },
    {
      "epoch": 55.378787878787875,
      "grad_norm": 0.07213056087493896,
      "learning_rate": 7.231060606060607e-05,
      "loss": 0.0131,
      "step": 3655
    },
    {
      "epoch": 55.45454545454545,
      "grad_norm": 0.0995408445596695,
      "learning_rate": 7.227272727272729e-05,
      "loss": 0.0121,
      "step": 3660
    },
    {
      "epoch": 55.53030303030303,
      "grad_norm": 0.07010676711797714,
      "learning_rate": 7.223484848484849e-05,
      "loss": 0.0133,
      "step": 3665
    },
    {
      "epoch": 55.60606060606061,
      "grad_norm": 0.09495700150728226,
      "learning_rate": 7.219696969696969e-05,
      "loss": 0.0149,
      "step": 3670
    },
    {
      "epoch": 55.68181818181818,
      "grad_norm": 0.04553232714533806,
      "learning_rate": 7.215909090909091e-05,
      "loss": 0.0134,
      "step": 3675
    },
    {
      "epoch": 55.75757575757576,
      "grad_norm": 0.13620024919509888,
      "learning_rate": 7.212121212121213e-05,
      "loss": 0.0117,
      "step": 3680
    },
    {
      "epoch": 55.833333333333336,
      "grad_norm": 0.11953846365213394,
      "learning_rate": 7.208333333333334e-05,
      "loss": 0.0117,
      "step": 3685
    },
    {
      "epoch": 55.90909090909091,
      "grad_norm": 0.07127209007740021,
      "learning_rate": 7.204545454545454e-05,
      "loss": 0.0087,
      "step": 3690
    },
    {
      "epoch": 55.984848484848484,
      "grad_norm": 0.048511188477277756,
      "learning_rate": 7.200757575757576e-05,
      "loss": 0.0125,
      "step": 3695
    },
    {
      "epoch": 56.0,
      "eval_loss": 0.013337178155779839,
      "eval_mean_accuracy": 0.9531843960496164,
      "eval_mean_iou": 0.9214980769119747,
      "eval_runtime": 37.1834,
      "eval_samples_per_second": 6.32,
      "eval_steps_per_second": 0.807,
      "step": 3696
    },
    {
      "epoch": 56.06060606060606,
      "grad_norm": 0.08462590724229813,
      "learning_rate": 7.196969696969698e-05,
      "loss": 0.0126,
      "step": 3700
    },
    {
      "epoch": 56.13636363636363,
      "grad_norm": 0.05129329860210419,
      "learning_rate": 7.19318181818182e-05,
      "loss": 0.0115,
      "step": 3705
    },
    {
      "epoch": 56.21212121212121,
      "grad_norm": 0.06470350921154022,
      "learning_rate": 7.18939393939394e-05,
      "loss": 0.0099,
      "step": 3710
    },
    {
      "epoch": 56.28787878787879,
      "grad_norm": 0.07905973494052887,
      "learning_rate": 7.18560606060606e-05,
      "loss": 0.0133,
      "step": 3715
    },
    {
      "epoch": 56.36363636363637,
      "grad_norm": 0.046889904886484146,
      "learning_rate": 7.181818181818182e-05,
      "loss": 0.0125,
      "step": 3720
    },
    {
      "epoch": 56.43939393939394,
      "grad_norm": 0.06411285698413849,
      "learning_rate": 7.178030303030303e-05,
      "loss": 0.0122,
      "step": 3725
    },
    {
      "epoch": 56.515151515151516,
      "grad_norm": 0.026050303131341934,
      "learning_rate": 7.174242424242425e-05,
      "loss": 0.0098,
      "step": 3730
    },
    {
      "epoch": 56.59090909090909,
      "grad_norm": 0.10320115089416504,
      "learning_rate": 7.170454545454545e-05,
      "loss": 0.0129,
      "step": 3735
    },
    {
      "epoch": 56.666666666666664,
      "grad_norm": 0.12689624726772308,
      "learning_rate": 7.166666666666667e-05,
      "loss": 0.0132,
      "step": 3740
    },
    {
      "epoch": 56.74242424242424,
      "grad_norm": 0.0848609209060669,
      "learning_rate": 7.162878787878789e-05,
      "loss": 0.0154,
      "step": 3745
    },
    {
      "epoch": 56.81818181818182,
      "grad_norm": 0.04083557426929474,
      "learning_rate": 7.15909090909091e-05,
      "loss": 0.0126,
      "step": 3750
    },
    {
      "epoch": 56.89393939393939,
      "grad_norm": 0.04378031566739082,
      "learning_rate": 7.15530303030303e-05,
      "loss": 0.009,
      "step": 3755
    },
    {
      "epoch": 56.96969696969697,
      "grad_norm": 0.09939198195934296,
      "learning_rate": 7.151515151515152e-05,
      "loss": 0.0111,
      "step": 3760
    },
    {
      "epoch": 57.0,
      "eval_loss": 0.012748904526233673,
      "eval_mean_accuracy": 0.946765034299065,
      "eval_mean_iou": 0.9190036204224811,
      "eval_runtime": 37.5983,
      "eval_samples_per_second": 6.25,
      "eval_steps_per_second": 0.798,
      "step": 3762
    },
    {
      "epoch": 57.04545454545455,
      "grad_norm": 0.08476035296916962,
      "learning_rate": 7.147727272727273e-05,
      "loss": 0.0117,
      "step": 3765
    },
    {
      "epoch": 57.121212121212125,
      "grad_norm": 0.025504618883132935,
      "learning_rate": 7.143939393939394e-05,
      "loss": 0.0099,
      "step": 3770
    },
    {
      "epoch": 57.196969696969695,
      "grad_norm": 0.07169104367494583,
      "learning_rate": 7.140151515151516e-05,
      "loss": 0.0108,
      "step": 3775
    },
    {
      "epoch": 57.27272727272727,
      "grad_norm": 0.03472178056836128,
      "learning_rate": 7.136363636363636e-05,
      "loss": 0.0102,
      "step": 3780
    },
    {
      "epoch": 57.34848484848485,
      "grad_norm": 0.08064550906419754,
      "learning_rate": 7.132575757575758e-05,
      "loss": 0.0138,
      "step": 3785
    },
    {
      "epoch": 57.42424242424242,
      "grad_norm": 0.05632326751947403,
      "learning_rate": 7.12878787878788e-05,
      "loss": 0.0109,
      "step": 3790
    },
    {
      "epoch": 57.5,
      "grad_norm": 0.04317541792988777,
      "learning_rate": 7.125000000000001e-05,
      "loss": 0.0126,
      "step": 3795
    },
    {
      "epoch": 57.57575757575758,
      "grad_norm": 0.03845816105604172,
      "learning_rate": 7.121212121212121e-05,
      "loss": 0.012,
      "step": 3800
    },
    {
      "epoch": 57.65151515151515,
      "grad_norm": 0.037458259612321854,
      "learning_rate": 7.117424242424243e-05,
      "loss": 0.0102,
      "step": 3805
    },
    {
      "epoch": 57.72727272727273,
      "grad_norm": 0.0941748097538948,
      "learning_rate": 7.113636363636363e-05,
      "loss": 0.014,
      "step": 3810
    },
    {
      "epoch": 57.803030303030305,
      "grad_norm": 0.07625503093004227,
      "learning_rate": 7.109848484848485e-05,
      "loss": 0.0126,
      "step": 3815
    },
    {
      "epoch": 57.878787878787875,
      "grad_norm": 0.03961586952209473,
      "learning_rate": 7.106060606060607e-05,
      "loss": 0.0114,
      "step": 3820
    },
    {
      "epoch": 57.95454545454545,
      "grad_norm": 0.07360029965639114,
      "learning_rate": 7.102272727272727e-05,
      "loss": 0.0114,
      "step": 3825
    },
    {
      "epoch": 58.0,
      "eval_loss": 0.012880906462669373,
      "eval_mean_accuracy": 0.9565657469240847,
      "eval_mean_iou": 0.9221521656009101,
      "eval_runtime": 40.0553,
      "eval_samples_per_second": 5.867,
      "eval_steps_per_second": 0.749,
      "step": 3828
    },
    {
      "epoch": 58.03030303030303,
      "grad_norm": 0.0524006113409996,
      "learning_rate": 7.098484848484849e-05,
      "loss": 0.012,
      "step": 3830
    },
    {
      "epoch": 58.10606060606061,
      "grad_norm": 0.03739740699529648,
      "learning_rate": 7.09469696969697e-05,
      "loss": 0.0138,
      "step": 3835
    },
    {
      "epoch": 58.18181818181818,
      "grad_norm": 0.033809173852205276,
      "learning_rate": 7.090909090909092e-05,
      "loss": 0.0093,
      "step": 3840
    },
    {
      "epoch": 58.25757575757576,
      "grad_norm": 0.10202228277921677,
      "learning_rate": 7.087121212121212e-05,
      "loss": 0.0126,
      "step": 3845
    },
    {
      "epoch": 58.333333333333336,
      "grad_norm": 0.024866582825779915,
      "learning_rate": 7.083333333333334e-05,
      "loss": 0.0129,
      "step": 3850
    },
    {
      "epoch": 58.40909090909091,
      "grad_norm": 0.05789878964424133,
      "learning_rate": 7.079545454545455e-05,
      "loss": 0.0122,
      "step": 3855
    },
    {
      "epoch": 58.484848484848484,
      "grad_norm": 0.052291419357061386,
      "learning_rate": 7.075757575757576e-05,
      "loss": 0.0117,
      "step": 3860
    },
    {
      "epoch": 58.56060606060606,
      "grad_norm": 0.08531995862722397,
      "learning_rate": 7.071969696969697e-05,
      "loss": 0.0116,
      "step": 3865
    },
    {
      "epoch": 58.63636363636363,
      "grad_norm": 0.04689672589302063,
      "learning_rate": 7.068181818181818e-05,
      "loss": 0.0111,
      "step": 3870
    },
    {
      "epoch": 58.71212121212121,
      "grad_norm": 0.11566620320081711,
      "learning_rate": 7.06439393939394e-05,
      "loss": 0.0122,
      "step": 3875
    },
    {
      "epoch": 58.78787878787879,
      "grad_norm": 0.031542982906103134,
      "learning_rate": 7.060606060606061e-05,
      "loss": 0.0098,
      "step": 3880
    },
    {
      "epoch": 58.86363636363637,
      "grad_norm": 0.04567784070968628,
      "learning_rate": 7.056818181818183e-05,
      "loss": 0.0117,
      "step": 3885
    },
    {
      "epoch": 58.93939393939394,
      "grad_norm": 0.15457390248775482,
      "learning_rate": 7.053030303030303e-05,
      "loss": 0.0113,
      "step": 3890
    },
    {
      "epoch": 59.0,
      "eval_loss": 0.012969030998647213,
      "eval_mean_accuracy": 0.9616300452489798,
      "eval_mean_iou": 0.9223188563470325,
      "eval_runtime": 37.9106,
      "eval_samples_per_second": 6.199,
      "eval_steps_per_second": 0.791,
      "step": 3894
    },
    {
      "epoch": 59.015151515151516,
      "grad_norm": 0.09605880826711655,
      "learning_rate": 7.049242424242425e-05,
      "loss": 0.0123,
      "step": 3895
    },
    {
      "epoch": 59.09090909090909,
      "grad_norm": 0.06294688582420349,
      "learning_rate": 7.045454545454546e-05,
      "loss": 0.0125,
      "step": 3900
    },
    {
      "epoch": 59.166666666666664,
      "grad_norm": 0.09651622921228409,
      "learning_rate": 7.041666666666668e-05,
      "loss": 0.0114,
      "step": 3905
    },
    {
      "epoch": 59.24242424242424,
      "grad_norm": 0.05601264908909798,
      "learning_rate": 7.037878787878788e-05,
      "loss": 0.0125,
      "step": 3910
    },
    {
      "epoch": 59.31818181818182,
      "grad_norm": 0.08048252016305923,
      "learning_rate": 7.034090909090909e-05,
      "loss": 0.0117,
      "step": 3915
    },
    {
      "epoch": 59.39393939393939,
      "grad_norm": 0.04494091868400574,
      "learning_rate": 7.03030303030303e-05,
      "loss": 0.0112,
      "step": 3920
    },
    {
      "epoch": 59.46969696969697,
      "grad_norm": 0.036121025681495667,
      "learning_rate": 7.026515151515152e-05,
      "loss": 0.0107,
      "step": 3925
    },
    {
      "epoch": 59.54545454545455,
      "grad_norm": 0.05400189757347107,
      "learning_rate": 7.022727272727274e-05,
      "loss": 0.0121,
      "step": 3930
    },
    {
      "epoch": 59.621212121212125,
      "grad_norm": 0.13459229469299316,
      "learning_rate": 7.018939393939394e-05,
      "loss": 0.0106,
      "step": 3935
    },
    {
      "epoch": 59.696969696969695,
      "grad_norm": 0.050082024186849594,
      "learning_rate": 7.015151515151515e-05,
      "loss": 0.0125,
      "step": 3940
    },
    {
      "epoch": 59.77272727272727,
      "grad_norm": 0.02532917819917202,
      "learning_rate": 7.011363636363637e-05,
      "loss": 0.0099,
      "step": 3945
    },
    {
      "epoch": 59.84848484848485,
      "grad_norm": 0.04282253980636597,
      "learning_rate": 7.007575757575759e-05,
      "loss": 0.0131,
      "step": 3950
    },
    {
      "epoch": 59.92424242424242,
      "grad_norm": 0.03396788612008095,
      "learning_rate": 7.003787878787879e-05,
      "loss": 0.0119,
      "step": 3955
    },
    {
      "epoch": 60.0,
      "grad_norm": 0.04221852868795395,
      "learning_rate": 7e-05,
      "loss": 0.0111,
      "step": 3960
    },
    {
      "epoch": 60.0,
      "eval_loss": 0.012523246929049492,
      "eval_mean_accuracy": 0.9513905943239722,
      "eval_mean_iou": 0.9221731588282052,
      "eval_runtime": 38.7454,
      "eval_samples_per_second": 6.065,
      "eval_steps_per_second": 0.774,
      "step": 3960
    },
    {
      "epoch": 60.07575757575758,
      "grad_norm": 0.04121781885623932,
      "learning_rate": 6.996212121212121e-05,
      "loss": 0.0103,
      "step": 3965
    },
    {
      "epoch": 60.15151515151515,
      "grad_norm": 0.06508173048496246,
      "learning_rate": 6.992424242424243e-05,
      "loss": 0.0106,
      "step": 3970
    },
    {
      "epoch": 60.22727272727273,
      "grad_norm": 0.06700477004051208,
      "learning_rate": 6.988636363636364e-05,
      "loss": 0.0159,
      "step": 3975
    },
    {
      "epoch": 60.303030303030305,
      "grad_norm": 0.042721450328826904,
      "learning_rate": 6.984848484848485e-05,
      "loss": 0.0107,
      "step": 3980
    },
    {
      "epoch": 60.378787878787875,
      "grad_norm": 0.04656713083386421,
      "learning_rate": 6.981060606060606e-05,
      "loss": 0.0134,
      "step": 3985
    },
    {
      "epoch": 60.45454545454545,
      "grad_norm": 0.18768616020679474,
      "learning_rate": 6.977272727272728e-05,
      "loss": 0.0122,
      "step": 3990
    },
    {
      "epoch": 60.53030303030303,
      "grad_norm": 0.03940529003739357,
      "learning_rate": 6.97348484848485e-05,
      "loss": 0.011,
      "step": 3995
    },
    {
      "epoch": 60.60606060606061,
      "grad_norm": 0.0688980296254158,
      "learning_rate": 6.96969696969697e-05,
      "loss": 0.015,
      "step": 4000
    },
    {
      "epoch": 60.68181818181818,
      "grad_norm": 0.045764610171318054,
      "learning_rate": 6.96590909090909e-05,
      "loss": 0.0104,
      "step": 4005
    },
    {
      "epoch": 60.75757575757576,
      "grad_norm": 0.06260871142148972,
      "learning_rate": 6.962121212121212e-05,
      "loss": 0.0135,
      "step": 4010
    },
    {
      "epoch": 60.833333333333336,
      "grad_norm": 0.03793005272746086,
      "learning_rate": 6.958333333333334e-05,
      "loss": 0.0094,
      "step": 4015
    },
    {
      "epoch": 60.90909090909091,
      "grad_norm": 0.06887661665678024,
      "learning_rate": 6.954545454545455e-05,
      "loss": 0.0107,
      "step": 4020
    },
    {
      "epoch": 60.984848484848484,
      "grad_norm": 0.031375572085380554,
      "learning_rate": 6.950757575757575e-05,
      "loss": 0.0103,
      "step": 4025
    },
    {
      "epoch": 61.0,
      "eval_loss": 0.013117657974362373,
      "eval_mean_accuracy": 0.9615693474023521,
      "eval_mean_iou": 0.9214676126093998,
      "eval_runtime": 37.0758,
      "eval_samples_per_second": 6.338,
      "eval_steps_per_second": 0.809,
      "step": 4026
    },
    {
      "epoch": 61.06060606060606,
      "grad_norm": 0.09805615246295929,
      "learning_rate": 6.946969696969697e-05,
      "loss": 0.0098,
      "step": 4030
    },
    {
      "epoch": 61.13636363636363,
      "grad_norm": 0.16170643270015717,
      "learning_rate": 6.943181818181819e-05,
      "loss": 0.0156,
      "step": 4035
    },
    {
      "epoch": 61.21212121212121,
      "grad_norm": 0.0989101231098175,
      "learning_rate": 6.93939393939394e-05,
      "loss": 0.0135,
      "step": 4040
    },
    {
      "epoch": 61.28787878787879,
      "grad_norm": 0.026725921779870987,
      "learning_rate": 6.935606060606061e-05,
      "loss": 0.0116,
      "step": 4045
    },
    {
      "epoch": 61.36363636363637,
      "grad_norm": 0.11946594715118408,
      "learning_rate": 6.931818181818182e-05,
      "loss": 0.0136,
      "step": 4050
    },
    {
      "epoch": 61.43939393939394,
      "grad_norm": 0.06817812472581863,
      "learning_rate": 6.928030303030303e-05,
      "loss": 0.0111,
      "step": 4055
    },
    {
      "epoch": 61.515151515151516,
      "grad_norm": 0.07762253284454346,
      "learning_rate": 6.924242424242424e-05,
      "loss": 0.0086,
      "step": 4060
    },
    {
      "epoch": 61.59090909090909,
      "grad_norm": 0.04134995490312576,
      "learning_rate": 6.920454545454546e-05,
      "loss": 0.0117,
      "step": 4065
    },
    {
      "epoch": 61.666666666666664,
      "grad_norm": 0.07784988731145859,
      "learning_rate": 6.916666666666666e-05,
      "loss": 0.0104,
      "step": 4070
    },
    {
      "epoch": 61.74242424242424,
      "grad_norm": 0.07652796059846878,
      "learning_rate": 6.912878787878788e-05,
      "loss": 0.0103,
      "step": 4075
    },
    {
      "epoch": 61.81818181818182,
      "grad_norm": 0.07818873226642609,
      "learning_rate": 6.90909090909091e-05,
      "loss": 0.014,
      "step": 4080
    },
    {
      "epoch": 61.89393939393939,
      "grad_norm": 0.05453864857554436,
      "learning_rate": 6.905303030303031e-05,
      "loss": 0.0109,
      "step": 4085
    },
    {
      "epoch": 61.96969696969697,
      "grad_norm": 0.037294626235961914,
      "learning_rate": 6.901515151515152e-05,
      "loss": 0.01,
      "step": 4090
    },
    {
      "epoch": 62.0,
      "eval_loss": 0.01275706384330988,
      "eval_mean_accuracy": 0.9527113031771633,
      "eval_mean_iou": 0.9215169005792726,
      "eval_runtime": 38.2824,
      "eval_samples_per_second": 6.139,
      "eval_steps_per_second": 0.784,
      "step": 4092
    },
    {
      "epoch": 62.04545454545455,
      "grad_norm": 0.04737137630581856,
      "learning_rate": 6.897727272727273e-05,
      "loss": 0.012,
      "step": 4095
    },
    {
      "epoch": 62.121212121212125,
      "grad_norm": 0.03990905359387398,
      "learning_rate": 6.893939393939395e-05,
      "loss": 0.0113,
      "step": 4100
    },
    {
      "epoch": 62.196969696969695,
      "grad_norm": 0.03616999089717865,
      "learning_rate": 6.890151515151515e-05,
      "loss": 0.0107,
      "step": 4105
    },
    {
      "epoch": 62.27272727272727,
      "grad_norm": 0.037457384169101715,
      "learning_rate": 6.886363636363637e-05,
      "loss": 0.0122,
      "step": 4110
    },
    {
      "epoch": 62.34848484848485,
      "grad_norm": 0.055457111448049545,
      "learning_rate": 6.882575757575757e-05,
      "loss": 0.0123,
      "step": 4115
    },
    {
      "epoch": 62.42424242424242,
      "grad_norm": 0.03277739882469177,
      "learning_rate": 6.878787878787879e-05,
      "loss": 0.0117,
      "step": 4120
    },
    {
      "epoch": 62.5,
      "grad_norm": 0.049716051667928696,
      "learning_rate": 6.875e-05,
      "loss": 0.0114,
      "step": 4125
    },
    {
      "epoch": 62.57575757575758,
      "grad_norm": 0.05659966915845871,
      "learning_rate": 6.871212121212122e-05,
      "loss": 0.0133,
      "step": 4130
    },
    {
      "epoch": 62.65151515151515,
      "grad_norm": 0.08494408428668976,
      "learning_rate": 6.867424242424242e-05,
      "loss": 0.0088,
      "step": 4135
    },
    {
      "epoch": 62.72727272727273,
      "grad_norm": 0.06011077016592026,
      "learning_rate": 6.863636363636364e-05,
      "loss": 0.0102,
      "step": 4140
    },
    {
      "epoch": 62.803030303030305,
      "grad_norm": 0.027212150394916534,
      "learning_rate": 6.859848484848486e-05,
      "loss": 0.0101,
      "step": 4145
    },
    {
      "epoch": 62.878787878787875,
      "grad_norm": 0.0672391876578331,
      "learning_rate": 6.856060606060606e-05,
      "loss": 0.0122,
      "step": 4150
    },
    {
      "epoch": 62.95454545454545,
      "grad_norm": 0.08669383078813553,
      "learning_rate": 6.852272727272728e-05,
      "loss": 0.0144,
      "step": 4155
    },
    {
      "epoch": 63.0,
      "eval_loss": 0.012646224349737167,
      "eval_mean_accuracy": 0.9487086580832264,
      "eval_mean_iou": 0.9196621043315837,
      "eval_runtime": 37.8136,
      "eval_samples_per_second": 6.215,
      "eval_steps_per_second": 0.793,
      "step": 4158
    },
    {
      "epoch": 63.03030303030303,
      "grad_norm": 0.04667006805539131,
      "learning_rate": 6.848484848484848e-05,
      "loss": 0.0123,
      "step": 4160
    },
    {
      "epoch": 63.10606060606061,
      "grad_norm": 0.08743372559547424,
      "learning_rate": 6.84469696969697e-05,
      "loss": 0.0109,
      "step": 4165
    },
    {
      "epoch": 63.18181818181818,
      "grad_norm": 0.04360045865178108,
      "learning_rate": 6.840909090909091e-05,
      "loss": 0.0087,
      "step": 4170
    },
    {
      "epoch": 63.25757575757576,
      "grad_norm": 0.03124597668647766,
      "learning_rate": 6.837121212121213e-05,
      "loss": 0.0115,
      "step": 4175
    },
    {
      "epoch": 63.333333333333336,
      "grad_norm": 0.0662151426076889,
      "learning_rate": 6.833333333333333e-05,
      "loss": 0.0094,
      "step": 4180
    },
    {
      "epoch": 63.40909090909091,
      "grad_norm": 0.04773330315947533,
      "learning_rate": 6.829545454545455e-05,
      "loss": 0.0114,
      "step": 4185
    },
    {
      "epoch": 63.484848484848484,
      "grad_norm": 0.07932186126708984,
      "learning_rate": 6.825757575757576e-05,
      "loss": 0.0158,
      "step": 4190
    },
    {
      "epoch": 63.56060606060606,
      "grad_norm": 0.046718817204236984,
      "learning_rate": 6.821969696969698e-05,
      "loss": 0.0117,
      "step": 4195
    },
    {
      "epoch": 63.63636363636363,
      "grad_norm": 0.12640522420406342,
      "learning_rate": 6.818181818181818e-05,
      "loss": 0.011,
      "step": 4200
    },
    {
      "epoch": 63.71212121212121,
      "grad_norm": 0.04979836568236351,
      "learning_rate": 6.814393939393939e-05,
      "loss": 0.0136,
      "step": 4205
    },
    {
      "epoch": 63.78787878787879,
      "grad_norm": 0.029140133410692215,
      "learning_rate": 6.81060606060606e-05,
      "loss": 0.0136,
      "step": 4210
    },
    {
      "epoch": 63.86363636363637,
      "grad_norm": 0.05822339653968811,
      "learning_rate": 6.806818181818182e-05,
      "loss": 0.0101,
      "step": 4215
    },
    {
      "epoch": 63.93939393939394,
      "grad_norm": 0.14213977754116058,
      "learning_rate": 6.803030303030304e-05,
      "loss": 0.0124,
      "step": 4220
    },
    {
      "epoch": 64.0,
      "eval_loss": 0.012501081451773643,
      "eval_mean_accuracy": 0.952744443923253,
      "eval_mean_iou": 0.9232006954900239,
      "eval_runtime": 38.5321,
      "eval_samples_per_second": 6.099,
      "eval_steps_per_second": 0.779,
      "step": 4224
    },
    {
      "epoch": 64.01515151515152,
      "grad_norm": 0.05092529579997063,
      "learning_rate": 6.799242424242424e-05,
      "loss": 0.0123,
      "step": 4225
    },
    {
      "epoch": 64.0909090909091,
      "grad_norm": 0.03776617720723152,
      "learning_rate": 6.795454545454546e-05,
      "loss": 0.0115,
      "step": 4230
    },
    {
      "epoch": 64.16666666666667,
      "grad_norm": 0.03788723051548004,
      "learning_rate": 6.791666666666667e-05,
      "loss": 0.0091,
      "step": 4235
    },
    {
      "epoch": 64.24242424242425,
      "grad_norm": 0.03464753180742264,
      "learning_rate": 6.787878787878789e-05,
      "loss": 0.0106,
      "step": 4240
    },
    {
      "epoch": 64.31818181818181,
      "grad_norm": 0.18742735683918,
      "learning_rate": 6.784090909090909e-05,
      "loss": 0.0135,
      "step": 4245
    },
    {
      "epoch": 64.39393939393939,
      "grad_norm": 0.08644703030586243,
      "learning_rate": 6.78030303030303e-05,
      "loss": 0.0129,
      "step": 4250
    },
    {
      "epoch": 64.46969696969697,
      "grad_norm": 0.0819237232208252,
      "learning_rate": 6.776515151515151e-05,
      "loss": 0.0126,
      "step": 4255
    },
    {
      "epoch": 64.54545454545455,
      "grad_norm": 0.05149337276816368,
      "learning_rate": 6.772727272727273e-05,
      "loss": 0.0104,
      "step": 4260
    },
    {
      "epoch": 64.62121212121212,
      "grad_norm": 0.04178178310394287,
      "learning_rate": 6.768939393939395e-05,
      "loss": 0.0138,
      "step": 4265
    },
    {
      "epoch": 64.6969696969697,
      "grad_norm": 0.056223314255476,
      "learning_rate": 6.765151515151515e-05,
      "loss": 0.0101,
      "step": 4270
    },
    {
      "epoch": 64.77272727272727,
      "grad_norm": 0.028331341221928596,
      "learning_rate": 6.761363636363636e-05,
      "loss": 0.0106,
      "step": 4275
    },
    {
      "epoch": 64.84848484848484,
      "grad_norm": 0.04504998400807381,
      "learning_rate": 6.757575757575758e-05,
      "loss": 0.0103,
      "step": 4280
    },
    {
      "epoch": 64.92424242424242,
      "grad_norm": 0.06155358999967575,
      "learning_rate": 6.75378787878788e-05,
      "loss": 0.0142,
      "step": 4285
    },
    {
      "epoch": 65.0,
      "grad_norm": 0.05468761548399925,
      "learning_rate": 6.750000000000001e-05,
      "loss": 0.0114,
      "step": 4290
    },
    {
      "epoch": 65.0,
      "eval_loss": 0.013594343326985836,
      "eval_mean_accuracy": 0.9412699663909988,
      "eval_mean_iou": 0.9153205433235868,
      "eval_runtime": 38.375,
      "eval_samples_per_second": 6.124,
      "eval_steps_per_second": 0.782,
      "step": 4290
    },
    {
      "epoch": 65.07575757575758,
      "grad_norm": 0.06403299421072006,
      "learning_rate": 6.74621212121212e-05,
      "loss": 0.0138,
      "step": 4295
    },
    {
      "epoch": 65.15151515151516,
      "grad_norm": 0.028770241886377335,
      "learning_rate": 6.742424242424242e-05,
      "loss": 0.0089,
      "step": 4300
    },
    {
      "epoch": 65.22727272727273,
      "grad_norm": 0.04401589557528496,
      "learning_rate": 6.738636363636364e-05,
      "loss": 0.0131,
      "step": 4305
    },
    {
      "epoch": 65.3030303030303,
      "grad_norm": 0.03673441335558891,
      "learning_rate": 6.734848484848485e-05,
      "loss": 0.0146,
      "step": 4310
    },
    {
      "epoch": 65.37878787878788,
      "grad_norm": 0.036511268466711044,
      "learning_rate": 6.731060606060606e-05,
      "loss": 0.0105,
      "step": 4315
    },
    {
      "epoch": 65.45454545454545,
      "grad_norm": 0.06756854057312012,
      "learning_rate": 6.727272727272727e-05,
      "loss": 0.0124,
      "step": 4320
    },
    {
      "epoch": 65.53030303030303,
      "grad_norm": 0.09373059868812561,
      "learning_rate": 6.723484848484849e-05,
      "loss": 0.0113,
      "step": 4325
    },
    {
      "epoch": 65.60606060606061,
      "grad_norm": 0.02682119980454445,
      "learning_rate": 6.71969696969697e-05,
      "loss": 0.0104,
      "step": 4330
    },
    {
      "epoch": 65.68181818181819,
      "grad_norm": 0.0909670889377594,
      "learning_rate": 6.715909090909092e-05,
      "loss": 0.0117,
      "step": 4335
    },
    {
      "epoch": 65.75757575757575,
      "grad_norm": 0.049221403896808624,
      "learning_rate": 6.712121212121213e-05,
      "loss": 0.012,
      "step": 4340
    },
    {
      "epoch": 65.83333333333333,
      "grad_norm": 0.08020204305648804,
      "learning_rate": 6.708333333333333e-05,
      "loss": 0.013,
      "step": 4345
    },
    {
      "epoch": 65.9090909090909,
      "grad_norm": 0.02726997621357441,
      "learning_rate": 6.704545454545455e-05,
      "loss": 0.0075,
      "step": 4350
    },
    {
      "epoch": 65.98484848484848,
      "grad_norm": 0.0937594473361969,
      "learning_rate": 6.700757575757576e-05,
      "loss": 0.0119,
      "step": 4355
    },
    {
      "epoch": 66.0,
      "eval_loss": 0.012598187662661076,
      "eval_mean_accuracy": 0.9540747487920583,
      "eval_mean_iou": 0.9223910198811728,
      "eval_runtime": 38.4978,
      "eval_samples_per_second": 6.104,
      "eval_steps_per_second": 0.779,
      "step": 4356
    },
    {
      "epoch": 66.06060606060606,
      "grad_norm": 0.05935138463973999,
      "learning_rate": 6.696969696969696e-05,
      "loss": 0.0123,
      "step": 4360
    },
    {
      "epoch": 66.13636363636364,
      "grad_norm": 0.037199921905994415,
      "learning_rate": 6.693181818181818e-05,
      "loss": 0.0171,
      "step": 4365
    },
    {
      "epoch": 66.21212121212122,
      "grad_norm": 0.07275083661079407,
      "learning_rate": 6.68939393939394e-05,
      "loss": 0.0088,
      "step": 4370
    },
    {
      "epoch": 66.28787878787878,
      "grad_norm": 0.07925418019294739,
      "learning_rate": 6.685606060606061e-05,
      "loss": 0.0088,
      "step": 4375
    },
    {
      "epoch": 66.36363636363636,
      "grad_norm": 0.03851373866200447,
      "learning_rate": 6.681818181818183e-05,
      "loss": 0.0086,
      "step": 4380
    },
    {
      "epoch": 66.43939393939394,
      "grad_norm": 0.03146800771355629,
      "learning_rate": 6.678030303030303e-05,
      "loss": 0.0163,
      "step": 4385
    },
    {
      "epoch": 66.51515151515152,
      "grad_norm": 0.06539692729711533,
      "learning_rate": 6.674242424242425e-05,
      "loss": 0.0085,
      "step": 4390
    },
    {
      "epoch": 66.5909090909091,
      "grad_norm": 0.038253720849752426,
      "learning_rate": 6.670454545454545e-05,
      "loss": 0.0109,
      "step": 4395
    },
    {
      "epoch": 66.66666666666667,
      "grad_norm": 0.06775215268135071,
      "learning_rate": 6.666666666666667e-05,
      "loss": 0.0115,
      "step": 4400
    },
    {
      "epoch": 66.74242424242425,
      "grad_norm": 0.08585315942764282,
      "learning_rate": 6.662878787878787e-05,
      "loss": 0.0107,
      "step": 4405
    },
    {
      "epoch": 66.81818181818181,
      "grad_norm": 0.06294826418161392,
      "learning_rate": 6.659090909090909e-05,
      "loss": 0.0123,
      "step": 4410
    },
    {
      "epoch": 66.89393939393939,
      "grad_norm": 0.086001455783844,
      "learning_rate": 6.65530303030303e-05,
      "loss": 0.0116,
      "step": 4415
    },
    {
      "epoch": 66.96969696969697,
      "grad_norm": 0.047361768782138824,
      "learning_rate": 6.651515151515152e-05,
      "loss": 0.0105,
      "step": 4420
    },
    {
      "epoch": 67.0,
      "eval_loss": 0.012462368234992027,
      "eval_mean_accuracy": 0.9514472616699776,
      "eval_mean_iou": 0.9222528506608595,
      "eval_runtime": 38.807,
      "eval_samples_per_second": 6.056,
      "eval_steps_per_second": 0.773,
      "step": 4422
    },
    {
      "epoch": 67.04545454545455,
      "grad_norm": 0.024092866107821465,
      "learning_rate": 6.647727272727274e-05,
      "loss": 0.0101,
      "step": 4425
    },
    {
      "epoch": 67.12121212121212,
      "grad_norm": 0.03415961563587189,
      "learning_rate": 6.643939393939394e-05,
      "loss": 0.012,
      "step": 4430
    },
    {
      "epoch": 67.1969696969697,
      "grad_norm": 0.03541991487145424,
      "learning_rate": 6.640151515151516e-05,
      "loss": 0.0131,
      "step": 4435
    },
    {
      "epoch": 67.27272727272727,
      "grad_norm": 0.05192383751273155,
      "learning_rate": 6.636363636363638e-05,
      "loss": 0.0102,
      "step": 4440
    },
    {
      "epoch": 67.34848484848484,
      "grad_norm": 0.05383046716451645,
      "learning_rate": 6.632575757575758e-05,
      "loss": 0.0137,
      "step": 4445
    },
    {
      "epoch": 67.42424242424242,
      "grad_norm": 0.05495131015777588,
      "learning_rate": 6.628787878787878e-05,
      "loss": 0.0115,
      "step": 4450
    },
    {
      "epoch": 67.5,
      "grad_norm": 0.11727051436901093,
      "learning_rate": 6.625e-05,
      "loss": 0.0095,
      "step": 4455
    },
    {
      "epoch": 67.57575757575758,
      "grad_norm": 0.03751982003450394,
      "learning_rate": 6.621212121212121e-05,
      "loss": 0.0123,
      "step": 4460
    },
    {
      "epoch": 67.65151515151516,
      "grad_norm": 0.05747637897729874,
      "learning_rate": 6.617424242424243e-05,
      "loss": 0.0115,
      "step": 4465
    },
    {
      "epoch": 67.72727272727273,
      "grad_norm": 0.06709247082471848,
      "learning_rate": 6.613636363636365e-05,
      "loss": 0.0093,
      "step": 4470
    },
    {
      "epoch": 67.8030303030303,
      "grad_norm": 0.05704130977392197,
      "learning_rate": 6.609848484848485e-05,
      "loss": 0.01,
      "step": 4475
    },
    {
      "epoch": 67.87878787878788,
      "grad_norm": 0.034582652151584625,
      "learning_rate": 6.606060606060607e-05,
      "loss": 0.0109,
      "step": 4480
    },
    {
      "epoch": 67.95454545454545,
      "grad_norm": 0.06423749029636383,
      "learning_rate": 6.602272727272728e-05,
      "loss": 0.009,
      "step": 4485
    },
    {
      "epoch": 68.0,
      "eval_loss": 0.012783810496330261,
      "eval_mean_accuracy": 0.9468898843183459,
      "eval_mean_iou": 0.9187416737573271,
      "eval_runtime": 37.9496,
      "eval_samples_per_second": 6.192,
      "eval_steps_per_second": 0.791,
      "step": 4488
    },
    {
      "epoch": 68.03030303030303,
      "grad_norm": 0.09640944749116898,
      "learning_rate": 6.598484848484849e-05,
      "loss": 0.0122,
      "step": 4490
    },
    {
      "epoch": 68.10606060606061,
      "grad_norm": 0.02793867141008377,
      "learning_rate": 6.594696969696969e-05,
      "loss": 0.0132,
      "step": 4495
    },
    {
      "epoch": 68.18181818181819,
      "grad_norm": 0.08540845662355423,
      "learning_rate": 6.59090909090909e-05,
      "loss": 0.0135,
      "step": 4500
    },
    {
      "epoch": 68.25757575757575,
      "grad_norm": 0.09172602742910385,
      "learning_rate": 6.587121212121212e-05,
      "loss": 0.0092,
      "step": 4505
    },
    {
      "epoch": 68.33333333333333,
      "grad_norm": 0.08024314045906067,
      "learning_rate": 6.583333333333334e-05,
      "loss": 0.01,
      "step": 4510
    },
    {
      "epoch": 68.4090909090909,
      "grad_norm": 0.08395801484584808,
      "learning_rate": 6.579545454545456e-05,
      "loss": 0.0106,
      "step": 4515
    },
    {
      "epoch": 68.48484848484848,
      "grad_norm": 0.022789670154452324,
      "learning_rate": 6.575757575757576e-05,
      "loss": 0.0107,
      "step": 4520
    },
    {
      "epoch": 68.56060606060606,
      "grad_norm": 0.07401345670223236,
      "learning_rate": 6.571969696969697e-05,
      "loss": 0.0134,
      "step": 4525
    },
    {
      "epoch": 68.63636363636364,
      "grad_norm": 0.07456403970718384,
      "learning_rate": 6.568181818181819e-05,
      "loss": 0.0099,
      "step": 4530
    },
    {
      "epoch": 68.71212121212122,
      "grad_norm": 0.0486326701939106,
      "learning_rate": 6.564393939393941e-05,
      "loss": 0.0139,
      "step": 4535
    },
    {
      "epoch": 68.78787878787878,
      "grad_norm": 0.08989811688661575,
      "learning_rate": 6.560606060606061e-05,
      "loss": 0.013,
      "step": 4540
    },
    {
      "epoch": 68.86363636363636,
      "grad_norm": 0.05103142559528351,
      "learning_rate": 6.556818181818181e-05,
      "loss": 0.0101,
      "step": 4545
    },
    {
      "epoch": 68.93939393939394,
      "grad_norm": 0.05049978196620941,
      "learning_rate": 6.553030303030303e-05,
      "loss": 0.0083,
      "step": 4550
    },
    {
      "epoch": 69.0,
      "eval_loss": 0.012642125599086285,
      "eval_mean_accuracy": 0.9484531566799381,
      "eval_mean_iou": 0.9197687234985249,
      "eval_runtime": 37.6116,
      "eval_samples_per_second": 6.248,
      "eval_steps_per_second": 0.798,
      "step": 4554
    },
    {
      "epoch": 69.01515151515152,
      "grad_norm": 0.0354972742497921,
      "learning_rate": 6.549242424242425e-05,
      "loss": 0.0127,
      "step": 4555
    },
    {
      "epoch": 69.0909090909091,
      "grad_norm": 0.16375724971294403,
      "learning_rate": 6.545454545454546e-05,
      "loss": 0.0109,
      "step": 4560
    },
    {
      "epoch": 69.16666666666667,
      "grad_norm": 0.07749626040458679,
      "learning_rate": 6.541666666666667e-05,
      "loss": 0.0111,
      "step": 4565
    },
    {
      "epoch": 69.24242424242425,
      "grad_norm": 0.07019774615764618,
      "learning_rate": 6.537878787878788e-05,
      "loss": 0.0115,
      "step": 4570
    },
    {
      "epoch": 69.31818181818181,
      "grad_norm": 0.03569045662879944,
      "learning_rate": 6.53409090909091e-05,
      "loss": 0.0082,
      "step": 4575
    },
    {
      "epoch": 69.39393939393939,
      "grad_norm": 0.08635324984788895,
      "learning_rate": 6.530303030303032e-05,
      "loss": 0.011,
      "step": 4580
    },
    {
      "epoch": 69.46969696969697,
      "grad_norm": 0.08610258251428604,
      "learning_rate": 6.526515151515152e-05,
      "loss": 0.0112,
      "step": 4585
    },
    {
      "epoch": 69.54545454545455,
      "grad_norm": 0.05177152529358864,
      "learning_rate": 6.522727272727272e-05,
      "loss": 0.0088,
      "step": 4590
    },
    {
      "epoch": 69.62121212121212,
      "grad_norm": 0.06793096661567688,
      "learning_rate": 6.518939393939394e-05,
      "loss": 0.0118,
      "step": 4595
    },
    {
      "epoch": 69.6969696969697,
      "grad_norm": 0.04318660497665405,
      "learning_rate": 6.515151515151516e-05,
      "loss": 0.0115,
      "step": 4600
    },
    {
      "epoch": 69.77272727272727,
      "grad_norm": 0.05080942064523697,
      "learning_rate": 6.511363636363637e-05,
      "loss": 0.0132,
      "step": 4605
    },
    {
      "epoch": 69.84848484848484,
      "grad_norm": 0.03461077809333801,
      "learning_rate": 6.507575757575757e-05,
      "loss": 0.0119,
      "step": 4610
    },
    {
      "epoch": 69.92424242424242,
      "grad_norm": 0.03419889882206917,
      "learning_rate": 6.503787878787879e-05,
      "loss": 0.0102,
      "step": 4615
    },
    {
      "epoch": 70.0,
      "grad_norm": 0.07109799236059189,
      "learning_rate": 6.500000000000001e-05,
      "loss": 0.0131,
      "step": 4620
    },
    {
      "epoch": 70.0,
      "eval_loss": 0.012310701422393322,
      "eval_mean_accuracy": 0.9457120279661587,
      "eval_mean_iou": 0.9200704408225021,
      "eval_runtime": 36.5613,
      "eval_samples_per_second": 6.428,
      "eval_steps_per_second": 0.821,
      "step": 4620
    },
    {
      "epoch": 70.07575757575758,
      "grad_norm": 0.0567115917801857,
      "learning_rate": 6.496212121212122e-05,
      "loss": 0.0117,
      "step": 4625
    },
    {
      "epoch": 70.15151515151516,
      "grad_norm": 0.05222885683178902,
      "learning_rate": 6.492424242424243e-05,
      "loss": 0.0088,
      "step": 4630
    },
    {
      "epoch": 70.22727272727273,
      "grad_norm": 0.042311690747737885,
      "learning_rate": 6.488636363636363e-05,
      "loss": 0.0125,
      "step": 4635
    },
    {
      "epoch": 70.3030303030303,
      "grad_norm": 0.05225116387009621,
      "learning_rate": 6.484848484848485e-05,
      "loss": 0.0117,
      "step": 4640
    },
    {
      "epoch": 70.37878787878788,
      "grad_norm": 0.08837945759296417,
      "learning_rate": 6.481060606060606e-05,
      "loss": 0.0118,
      "step": 4645
    },
    {
      "epoch": 70.45454545454545,
      "grad_norm": 0.0660945400595665,
      "learning_rate": 6.477272727272728e-05,
      "loss": 0.0146,
      "step": 4650
    },
    {
      "epoch": 70.53030303030303,
      "grad_norm": 0.06746061891317368,
      "learning_rate": 6.473484848484848e-05,
      "loss": 0.0114,
      "step": 4655
    },
    {
      "epoch": 70.60606060606061,
      "grad_norm": 0.036264415830373764,
      "learning_rate": 6.46969696969697e-05,
      "loss": 0.0123,
      "step": 4660
    },
    {
      "epoch": 70.68181818181819,
      "grad_norm": 0.045903827995061874,
      "learning_rate": 6.465909090909092e-05,
      "loss": 0.0097,
      "step": 4665
    },
    {
      "epoch": 70.75757575757575,
      "grad_norm": 0.05633692443370819,
      "learning_rate": 6.462121212121213e-05,
      "loss": 0.0102,
      "step": 4670
    },
    {
      "epoch": 70.83333333333333,
      "grad_norm": 0.03714665398001671,
      "learning_rate": 6.458333333333334e-05,
      "loss": 0.0089,
      "step": 4675
    },
    {
      "epoch": 70.9090909090909,
      "grad_norm": 0.037018224596977234,
      "learning_rate": 6.454545454545455e-05,
      "loss": 0.0081,
      "step": 4680
    },
    {
      "epoch": 70.98484848484848,
      "grad_norm": 0.09219218790531158,
      "learning_rate": 6.450757575757576e-05,
      "loss": 0.0137,
      "step": 4685
    },
    {
      "epoch": 71.0,
      "eval_loss": 0.012629001401364803,
      "eval_mean_accuracy": 0.9498252486329651,
      "eval_mean_iou": 0.9207706495465638,
      "eval_runtime": 38.316,
      "eval_samples_per_second": 6.133,
      "eval_steps_per_second": 0.783,
      "step": 4686
    },
    {
      "epoch": 71.06060606060606,
      "grad_norm": 0.07760203629732132,
      "learning_rate": 6.446969696969697e-05,
      "loss": 0.0116,
      "step": 4690
    },
    {
      "epoch": 71.13636363636364,
      "grad_norm": 0.0704277828335762,
      "learning_rate": 6.443181818181819e-05,
      "loss": 0.0131,
      "step": 4695
    },
    {
      "epoch": 71.21212121212122,
      "grad_norm": 0.031509023159742355,
      "learning_rate": 6.439393939393939e-05,
      "loss": 0.0073,
      "step": 4700
    },
    {
      "epoch": 71.28787878787878,
      "grad_norm": 0.15912030637264252,
      "learning_rate": 6.435606060606061e-05,
      "loss": 0.0114,
      "step": 4705
    },
    {
      "epoch": 71.36363636363636,
      "grad_norm": 0.06734999269247055,
      "learning_rate": 6.431818181818182e-05,
      "loss": 0.0141,
      "step": 4710
    },
    {
      "epoch": 71.43939393939394,
      "grad_norm": 0.08428304642438889,
      "learning_rate": 6.428030303030304e-05,
      "loss": 0.0106,
      "step": 4715
    },
    {
      "epoch": 71.51515151515152,
      "grad_norm": 0.04047851637005806,
      "learning_rate": 6.424242424242424e-05,
      "loss": 0.0096,
      "step": 4720
    },
    {
      "epoch": 71.5909090909091,
      "grad_norm": 0.04612775519490242,
      "learning_rate": 6.420454545454546e-05,
      "loss": 0.0111,
      "step": 4725
    },
    {
      "epoch": 71.66666666666667,
      "grad_norm": 0.02849680185317993,
      "learning_rate": 6.416666666666668e-05,
      "loss": 0.0112,
      "step": 4730
    },
    {
      "epoch": 71.74242424242425,
      "grad_norm": 0.04236720874905586,
      "learning_rate": 6.412878787878788e-05,
      "loss": 0.0113,
      "step": 4735
    },
    {
      "epoch": 71.81818181818181,
      "grad_norm": 0.03216387704014778,
      "learning_rate": 6.40909090909091e-05,
      "loss": 0.0114,
      "step": 4740
    },
    {
      "epoch": 71.89393939393939,
      "grad_norm": 0.050868310034275055,
      "learning_rate": 6.40530303030303e-05,
      "loss": 0.0122,
      "step": 4745
    },
    {
      "epoch": 71.96969696969697,
      "grad_norm": 0.16013120114803314,
      "learning_rate": 6.401515151515152e-05,
      "loss": 0.0124,
      "step": 4750
    },
    {
      "epoch": 72.0,
      "eval_loss": 0.01307035330682993,
      "eval_mean_accuracy": 0.9399394312133462,
      "eval_mean_iou": 0.9151516344796438,
      "eval_runtime": 37.7152,
      "eval_samples_per_second": 6.231,
      "eval_steps_per_second": 0.795,
      "step": 4752
    },
    {
      "epoch": 72.04545454545455,
      "grad_norm": 0.1348450481891632,
      "learning_rate": 6.397727272727273e-05,
      "loss": 0.0116,
      "step": 4755
    },
    {
      "epoch": 72.12121212121212,
      "grad_norm": 0.07382166385650635,
      "learning_rate": 6.393939393939395e-05,
      "loss": 0.013,
      "step": 4760
    },
    {
      "epoch": 72.1969696969697,
      "grad_norm": 0.03599587082862854,
      "learning_rate": 6.390151515151515e-05,
      "loss": 0.0128,
      "step": 4765
    },
    {
      "epoch": 72.27272727272727,
      "grad_norm": 0.04645940661430359,
      "learning_rate": 6.386363636363637e-05,
      "loss": 0.0089,
      "step": 4770
    },
    {
      "epoch": 72.34848484848484,
      "grad_norm": 0.05478982999920845,
      "learning_rate": 6.382575757575759e-05,
      "loss": 0.0108,
      "step": 4775
    },
    {
      "epoch": 72.42424242424242,
      "grad_norm": 0.16569973528385162,
      "learning_rate": 6.37878787878788e-05,
      "loss": 0.013,
      "step": 4780
    },
    {
      "epoch": 72.5,
      "grad_norm": 0.075131356716156,
      "learning_rate": 6.375e-05,
      "loss": 0.009,
      "step": 4785
    },
    {
      "epoch": 72.57575757575758,
      "grad_norm": 0.09459608048200607,
      "learning_rate": 6.371212121212121e-05,
      "loss": 0.0111,
      "step": 4790
    },
    {
      "epoch": 72.65151515151516,
      "grad_norm": 0.025974206626415253,
      "learning_rate": 6.367424242424242e-05,
      "loss": 0.0077,
      "step": 4795
    },
    {
      "epoch": 72.72727272727273,
      "grad_norm": 0.04162737727165222,
      "learning_rate": 6.363636363636364e-05,
      "loss": 0.0079,
      "step": 4800
    },
    {
      "epoch": 72.8030303030303,
      "grad_norm": 0.036529965698719025,
      "learning_rate": 6.359848484848486e-05,
      "loss": 0.0136,
      "step": 4805
    },
    {
      "epoch": 72.87878787878788,
      "grad_norm": 0.05888577178120613,
      "learning_rate": 6.356060606060606e-05,
      "loss": 0.0169,
      "step": 4810
    },
    {
      "epoch": 72.95454545454545,
      "grad_norm": 0.09174024313688278,
      "learning_rate": 6.352272727272728e-05,
      "loss": 0.012,
      "step": 4815
    },
    {
      "epoch": 73.0,
      "eval_loss": 0.012692486867308617,
      "eval_mean_accuracy": 0.9530754639293052,
      "eval_mean_iou": 0.919156484098307,
      "eval_runtime": 38.2885,
      "eval_samples_per_second": 6.138,
      "eval_steps_per_second": 0.784,
      "step": 4818
    },
    {
      "epoch": 73.03030303030303,
      "grad_norm": 0.062425412237644196,
      "learning_rate": 6.34848484848485e-05,
      "loss": 0.0105,
      "step": 4820
    },
    {
      "epoch": 73.10606060606061,
      "grad_norm": 0.12050031870603561,
      "learning_rate": 6.344696969696971e-05,
      "loss": 0.014,
      "step": 4825
    },
    {
      "epoch": 73.18181818181819,
      "grad_norm": 0.08030742406845093,
      "learning_rate": 6.340909090909091e-05,
      "loss": 0.0111,
      "step": 4830
    },
    {
      "epoch": 73.25757575757575,
      "grad_norm": 0.11159731447696686,
      "learning_rate": 6.337121212121212e-05,
      "loss": 0.0101,
      "step": 4835
    },
    {
      "epoch": 73.33333333333333,
      "grad_norm": 0.04641450196504593,
      "learning_rate": 6.333333333333333e-05,
      "loss": 0.0137,
      "step": 4840
    },
    {
      "epoch": 73.4090909090909,
      "grad_norm": 0.11552347987890244,
      "learning_rate": 6.329545454545455e-05,
      "loss": 0.0101,
      "step": 4845
    },
    {
      "epoch": 73.48484848484848,
      "grad_norm": 0.036518074572086334,
      "learning_rate": 6.325757575757577e-05,
      "loss": 0.0122,
      "step": 4850
    },
    {
      "epoch": 73.56060606060606,
      "grad_norm": 0.13449549674987793,
      "learning_rate": 6.321969696969697e-05,
      "loss": 0.0127,
      "step": 4855
    },
    {
      "epoch": 73.63636363636364,
      "grad_norm": 0.048244498670101166,
      "learning_rate": 6.318181818181818e-05,
      "loss": 0.0137,
      "step": 4860
    },
    {
      "epoch": 73.71212121212122,
      "grad_norm": 0.04050932079553604,
      "learning_rate": 6.31439393939394e-05,
      "loss": 0.0104,
      "step": 4865
    },
    {
      "epoch": 73.78787878787878,
      "grad_norm": 0.08685606718063354,
      "learning_rate": 6.310606060606062e-05,
      "loss": 0.009,
      "step": 4870
    },
    {
      "epoch": 73.86363636363636,
      "grad_norm": 0.05132646486163139,
      "learning_rate": 6.306818181818182e-05,
      "loss": 0.0118,
      "step": 4875
    },
    {
      "epoch": 73.93939393939394,
      "grad_norm": 0.1251213550567627,
      "learning_rate": 6.303030303030302e-05,
      "loss": 0.0109,
      "step": 4880
    },
    {
      "epoch": 74.0,
      "eval_loss": 0.012270085513591766,
      "eval_mean_accuracy": 0.958222793197086,
      "eval_mean_iou": 0.9240844040588803,
      "eval_runtime": 38.0115,
      "eval_samples_per_second": 6.182,
      "eval_steps_per_second": 0.789,
      "step": 4884
    },
    {
      "epoch": 74.01515151515152,
      "grad_norm": 0.15835894644260406,
      "learning_rate": 6.299242424242424e-05,
      "loss": 0.013,
      "step": 4885
    },
    {
      "epoch": 74.0909090909091,
      "grad_norm": 0.16184474527835846,
      "learning_rate": 6.295454545454546e-05,
      "loss": 0.0125,
      "step": 4890
    },
    {
      "epoch": 74.16666666666667,
      "grad_norm": 0.037963107228279114,
      "learning_rate": 6.291666666666667e-05,
      "loss": 0.0095,
      "step": 4895
    },
    {
      "epoch": 74.24242424242425,
      "grad_norm": 0.03786008805036545,
      "learning_rate": 6.287878787878788e-05,
      "loss": 0.011,
      "step": 4900
    },
    {
      "epoch": 74.31818181818181,
      "grad_norm": 0.04996492341160774,
      "learning_rate": 6.284090909090909e-05,
      "loss": 0.0087,
      "step": 4905
    },
    {
      "epoch": 74.39393939393939,
      "grad_norm": 0.03897299990057945,
      "learning_rate": 6.280303030303031e-05,
      "loss": 0.0137,
      "step": 4910
    },
    {
      "epoch": 74.46969696969697,
      "grad_norm": 0.07473551481962204,
      "learning_rate": 6.276515151515153e-05,
      "loss": 0.0106,
      "step": 4915
    },
    {
      "epoch": 74.54545454545455,
      "grad_norm": 0.04365568608045578,
      "learning_rate": 6.272727272727273e-05,
      "loss": 0.0092,
      "step": 4920
    },
    {
      "epoch": 74.62121212121212,
      "grad_norm": 0.047594692558050156,
      "learning_rate": 6.268939393939395e-05,
      "loss": 0.0143,
      "step": 4925
    },
    {
      "epoch": 74.6969696969697,
      "grad_norm": 0.04694069176912308,
      "learning_rate": 6.265151515151515e-05,
      "loss": 0.0097,
      "step": 4930
    },
    {
      "epoch": 74.77272727272727,
      "grad_norm": 0.06841346621513367,
      "learning_rate": 6.261363636363637e-05,
      "loss": 0.0099,
      "step": 4935
    },
    {
      "epoch": 74.84848484848484,
      "grad_norm": 0.023824667558073997,
      "learning_rate": 6.257575757575758e-05,
      "loss": 0.0089,
      "step": 4940
    },
    {
      "epoch": 74.92424242424242,
      "grad_norm": 0.15073633193969727,
      "learning_rate": 6.253787878787878e-05,
      "loss": 0.0154,
      "step": 4945
    },
    {
      "epoch": 75.0,
      "grad_norm": 0.03049931861460209,
      "learning_rate": 6.25e-05,
      "loss": 0.0084,
      "step": 4950
    },
    {
      "epoch": 75.0,
      "eval_loss": 0.012616033665835857,
      "eval_mean_accuracy": 0.96535495647982,
      "eval_mean_iou": 0.9231066225215528,
      "eval_runtime": 37.9174,
      "eval_samples_per_second": 6.198,
      "eval_steps_per_second": 0.791,
      "step": 4950
    },
    {
      "epoch": 75.07575757575758,
      "grad_norm": 0.04145142808556557,
      "learning_rate": 6.246212121212122e-05,
      "loss": 0.011,
      "step": 4955
    },
    {
      "epoch": 75.15151515151516,
      "grad_norm": 0.0678715929389,
      "learning_rate": 6.242424242424243e-05,
      "loss": 0.0122,
      "step": 4960
    },
    {
      "epoch": 75.22727272727273,
      "grad_norm": 0.13034574687480927,
      "learning_rate": 6.238636363636364e-05,
      "loss": 0.0096,
      "step": 4965
    },
    {
      "epoch": 75.3030303030303,
      "grad_norm": 0.03129105269908905,
      "learning_rate": 6.234848484848485e-05,
      "loss": 0.0089,
      "step": 4970
    },
    {
      "epoch": 75.37878787878788,
      "grad_norm": 0.052110474556684494,
      "learning_rate": 6.231060606060606e-05,
      "loss": 0.0161,
      "step": 4975
    },
    {
      "epoch": 75.45454545454545,
      "grad_norm": 0.08112243562936783,
      "learning_rate": 6.227272727272727e-05,
      "loss": 0.0099,
      "step": 4980
    },
    {
      "epoch": 75.53030303030303,
      "grad_norm": 0.05403774976730347,
      "learning_rate": 6.223484848484849e-05,
      "loss": 0.0111,
      "step": 4985
    },
    {
      "epoch": 75.60606060606061,
      "grad_norm": 0.028977181762456894,
      "learning_rate": 6.219696969696969e-05,
      "loss": 0.0097,
      "step": 4990
    },
    {
      "epoch": 75.68181818181819,
      "grad_norm": 0.023542912676930428,
      "learning_rate": 6.215909090909091e-05,
      "loss": 0.0111,
      "step": 4995
    },
    {
      "epoch": 75.75757575757575,
      "grad_norm": 0.03462165594100952,
      "learning_rate": 6.212121212121213e-05,
      "loss": 0.0102,
      "step": 5000
    },
    {
      "epoch": 75.83333333333333,
      "grad_norm": 0.03769727796316147,
      "learning_rate": 6.208333333333334e-05,
      "loss": 0.0093,
      "step": 5005
    },
    {
      "epoch": 75.9090909090909,
      "grad_norm": 0.08455546945333481,
      "learning_rate": 6.204545454545455e-05,
      "loss": 0.0108,
      "step": 5010
    },
    {
      "epoch": 75.98484848484848,
      "grad_norm": 0.07797928899526596,
      "learning_rate": 6.200757575757576e-05,
      "loss": 0.0128,
      "step": 5015
    },
    {
      "epoch": 76.0,
      "eval_loss": 0.012203234247863293,
      "eval_mean_accuracy": 0.9468688748827301,
      "eval_mean_iou": 0.9207384673729923,
      "eval_runtime": 38.1285,
      "eval_samples_per_second": 6.163,
      "eval_steps_per_second": 0.787,
      "step": 5016
    },
    {
      "epoch": 76.06060606060606,
      "grad_norm": 0.033195268362760544,
      "learning_rate": 6.196969696969698e-05,
      "loss": 0.0108,
      "step": 5020
    },
    {
      "epoch": 76.13636363636364,
      "grad_norm": 0.05578048154711723,
      "learning_rate": 6.193181818181818e-05,
      "loss": 0.0111,
      "step": 5025
    },
    {
      "epoch": 76.21212121212122,
      "grad_norm": 0.020961975678801537,
      "learning_rate": 6.18939393939394e-05,
      "loss": 0.0121,
      "step": 5030
    },
    {
      "epoch": 76.28787878787878,
      "grad_norm": 0.069123275578022,
      "learning_rate": 6.18560606060606e-05,
      "loss": 0.0111,
      "step": 5035
    },
    {
      "epoch": 76.36363636363636,
      "grad_norm": 0.05566676706075668,
      "learning_rate": 6.181818181818182e-05,
      "loss": 0.0103,
      "step": 5040
    },
    {
      "epoch": 76.43939393939394,
      "grad_norm": 0.20632827281951904,
      "learning_rate": 6.178030303030303e-05,
      "loss": 0.0123,
      "step": 5045
    },
    {
      "epoch": 76.51515151515152,
      "grad_norm": 0.04644811525940895,
      "learning_rate": 6.174242424242425e-05,
      "loss": 0.0107,
      "step": 5050
    },
    {
      "epoch": 76.5909090909091,
      "grad_norm": 0.10727785527706146,
      "learning_rate": 6.170454545454545e-05,
      "loss": 0.01,
      "step": 5055
    },
    {
      "epoch": 76.66666666666667,
      "grad_norm": 0.030500786378979683,
      "learning_rate": 6.166666666666667e-05,
      "loss": 0.009,
      "step": 5060
    },
    {
      "epoch": 76.74242424242425,
      "grad_norm": 0.11938727647066116,
      "learning_rate": 6.162878787878789e-05,
      "loss": 0.0163,
      "step": 5065
    },
    {
      "epoch": 76.81818181818181,
      "grad_norm": 0.05330207943916321,
      "learning_rate": 6.15909090909091e-05,
      "loss": 0.01,
      "step": 5070
    },
    {
      "epoch": 76.89393939393939,
      "grad_norm": 0.09527183324098587,
      "learning_rate": 6.15530303030303e-05,
      "loss": 0.0079,
      "step": 5075
    },
    {
      "epoch": 76.96969696969697,
      "grad_norm": 0.06988299638032913,
      "learning_rate": 6.151515151515151e-05,
      "loss": 0.0127,
      "step": 5080
    },
    {
      "epoch": 77.0,
      "eval_loss": 0.012030522339046001,
      "eval_mean_accuracy": 0.9577746505849826,
      "eval_mean_iou": 0.9249966294551871,
      "eval_runtime": 38.6078,
      "eval_samples_per_second": 6.087,
      "eval_steps_per_second": 0.777,
      "step": 5082
    },
    {
      "epoch": 77.04545454545455,
      "grad_norm": 0.06578565388917923,
      "learning_rate": 6.147727272727273e-05,
      "loss": 0.0103,
      "step": 5085
    },
    {
      "epoch": 77.12121212121212,
      "grad_norm": 0.06173538789153099,
      "learning_rate": 6.143939393939394e-05,
      "loss": 0.0069,
      "step": 5090
    },
    {
      "epoch": 77.1969696969697,
      "grad_norm": 0.12510448694229126,
      "learning_rate": 6.140151515151516e-05,
      "loss": 0.0119,
      "step": 5095
    },
    {
      "epoch": 77.27272727272727,
      "grad_norm": 0.06616706401109695,
      "learning_rate": 6.136363636363636e-05,
      "loss": 0.0098,
      "step": 5100
    },
    {
      "epoch": 77.34848484848484,
      "grad_norm": 0.08143316209316254,
      "learning_rate": 6.132575757575758e-05,
      "loss": 0.0145,
      "step": 5105
    },
    {
      "epoch": 77.42424242424242,
      "grad_norm": 0.029967322945594788,
      "learning_rate": 6.12878787878788e-05,
      "loss": 0.0119,
      "step": 5110
    },
    {
      "epoch": 77.5,
      "grad_norm": 0.028705701231956482,
      "learning_rate": 6.125000000000001e-05,
      "loss": 0.0135,
      "step": 5115
    },
    {
      "epoch": 77.57575757575758,
      "grad_norm": 0.13109959661960602,
      "learning_rate": 6.121212121212121e-05,
      "loss": 0.0111,
      "step": 5120
    },
    {
      "epoch": 77.65151515151516,
      "grad_norm": 0.06633663922548294,
      "learning_rate": 6.117424242424242e-05,
      "loss": 0.011,
      "step": 5125
    },
    {
      "epoch": 77.72727272727273,
      "grad_norm": 0.057570528239011765,
      "learning_rate": 6.113636363636363e-05,
      "loss": 0.0113,
      "step": 5130
    },
    {
      "epoch": 77.8030303030303,
      "grad_norm": 0.05112088844180107,
      "learning_rate": 6.109848484848485e-05,
      "loss": 0.0098,
      "step": 5135
    },
    {
      "epoch": 77.87878787878788,
      "grad_norm": 0.08773958683013916,
      "learning_rate": 6.106060606060607e-05,
      "loss": 0.012,
      "step": 5140
    },
    {
      "epoch": 77.95454545454545,
      "grad_norm": 0.08041108399629593,
      "learning_rate": 6.102272727272728e-05,
      "loss": 0.0129,
      "step": 5145
    },
    {
      "epoch": 78.0,
      "eval_loss": 0.012219646014273167,
      "eval_mean_accuracy": 0.9450409543191239,
      "eval_mean_iou": 0.9191327953708684,
      "eval_runtime": 38.5616,
      "eval_samples_per_second": 6.094,
      "eval_steps_per_second": 0.778,
      "step": 5148
    },
    {
      "epoch": 78.03030303030303,
      "grad_norm": 0.0643332228064537,
      "learning_rate": 6.098484848484849e-05,
      "loss": 0.0109,
      "step": 5150
    },
    {
      "epoch": 78.10606060606061,
      "grad_norm": 0.04110164940357208,
      "learning_rate": 6.09469696969697e-05,
      "loss": 0.0074,
      "step": 5155
    },
    {
      "epoch": 78.18181818181819,
      "grad_norm": 0.08478015661239624,
      "learning_rate": 6.090909090909091e-05,
      "loss": 0.0101,
      "step": 5160
    },
    {
      "epoch": 78.25757575757575,
      "grad_norm": 0.07190626114606857,
      "learning_rate": 6.087121212121213e-05,
      "loss": 0.0116,
      "step": 5165
    },
    {
      "epoch": 78.33333333333333,
      "grad_norm": 0.04038595035672188,
      "learning_rate": 6.083333333333333e-05,
      "loss": 0.0122,
      "step": 5170
    },
    {
      "epoch": 78.4090909090909,
      "grad_norm": 0.047902707010507584,
      "learning_rate": 6.079545454545454e-05,
      "loss": 0.0123,
      "step": 5175
    },
    {
      "epoch": 78.48484848484848,
      "grad_norm": 0.039987962692976,
      "learning_rate": 6.075757575757576e-05,
      "loss": 0.0111,
      "step": 5180
    },
    {
      "epoch": 78.56060606060606,
      "grad_norm": 0.02486814185976982,
      "learning_rate": 6.071969696969697e-05,
      "loss": 0.0114,
      "step": 5185
    },
    {
      "epoch": 78.63636363636364,
      "grad_norm": 0.037390146404504776,
      "learning_rate": 6.0681818181818185e-05,
      "loss": 0.0106,
      "step": 5190
    },
    {
      "epoch": 78.71212121212122,
      "grad_norm": 0.041209183633327484,
      "learning_rate": 6.0643939393939395e-05,
      "loss": 0.0112,
      "step": 5195
    },
    {
      "epoch": 78.78787878787878,
      "grad_norm": 0.13878729939460754,
      "learning_rate": 6.060606060606061e-05,
      "loss": 0.0121,
      "step": 5200
    },
    {
      "epoch": 78.86363636363636,
      "grad_norm": 0.029451536014676094,
      "learning_rate": 6.056818181818182e-05,
      "loss": 0.0103,
      "step": 5205
    },
    {
      "epoch": 78.93939393939394,
      "grad_norm": 0.03991081938147545,
      "learning_rate": 6.053030303030304e-05,
      "loss": 0.0122,
      "step": 5210
    },
    {
      "epoch": 79.0,
      "eval_loss": 0.011890086345374584,
      "eval_mean_accuracy": 0.9537089717441192,
      "eval_mean_iou": 0.924275847609146,
      "eval_runtime": 37.8854,
      "eval_samples_per_second": 6.203,
      "eval_steps_per_second": 0.792,
      "step": 5214
    },
    {
      "epoch": 79.01515151515152,
      "grad_norm": 0.05769460275769234,
      "learning_rate": 6.049242424242425e-05,
      "loss": 0.0101,
      "step": 5215
    },
    {
      "epoch": 79.0909090909091,
      "grad_norm": 0.058120809495449066,
      "learning_rate": 6.045454545454545e-05,
      "loss": 0.0114,
      "step": 5220
    },
    {
      "epoch": 79.16666666666667,
      "grad_norm": 0.04622367396950722,
      "learning_rate": 6.041666666666667e-05,
      "loss": 0.0077,
      "step": 5225
    },
    {
      "epoch": 79.24242424242425,
      "grad_norm": 0.1976555734872818,
      "learning_rate": 6.037878787878788e-05,
      "loss": 0.0095,
      "step": 5230
    },
    {
      "epoch": 79.31818181818181,
      "grad_norm": 0.08812398463487625,
      "learning_rate": 6.034090909090909e-05,
      "loss": 0.0119,
      "step": 5235
    },
    {
      "epoch": 79.39393939393939,
      "grad_norm": 0.06684183329343796,
      "learning_rate": 6.03030303030303e-05,
      "loss": 0.0106,
      "step": 5240
    },
    {
      "epoch": 79.46969696969697,
      "grad_norm": 0.04615764692425728,
      "learning_rate": 6.026515151515152e-05,
      "loss": 0.0113,
      "step": 5245
    },
    {
      "epoch": 79.54545454545455,
      "grad_norm": 0.03554265573620796,
      "learning_rate": 6.022727272727273e-05,
      "loss": 0.0095,
      "step": 5250
    },
    {
      "epoch": 79.62121212121212,
      "grad_norm": 0.03789900243282318,
      "learning_rate": 6.0189393939393946e-05,
      "loss": 0.012,
      "step": 5255
    },
    {
      "epoch": 79.6969696969697,
      "grad_norm": 0.02807208150625229,
      "learning_rate": 6.0151515151515156e-05,
      "loss": 0.0087,
      "step": 5260
    },
    {
      "epoch": 79.77272727272727,
      "grad_norm": 0.04348227009177208,
      "learning_rate": 6.011363636363637e-05,
      "loss": 0.0098,
      "step": 5265
    },
    {
      "epoch": 79.84848484848484,
      "grad_norm": 0.04057144373655319,
      "learning_rate": 6.0075757575757575e-05,
      "loss": 0.0132,
      "step": 5270
    },
    {
      "epoch": 79.92424242424242,
      "grad_norm": 0.058795034885406494,
      "learning_rate": 6.0037878787878785e-05,
      "loss": 0.0109,
      "step": 5275
    },
    {
      "epoch": 80.0,
      "grad_norm": 0.5864346623420715,
      "learning_rate": 6e-05,
      "loss": 0.015,
      "step": 5280
    },
    {
      "epoch": 80.0,
      "eval_loss": 0.012187034823000431,
      "eval_mean_accuracy": 0.9612084743835716,
      "eval_mean_iou": 0.9249165031943463,
      "eval_runtime": 37.7907,
      "eval_samples_per_second": 6.218,
      "eval_steps_per_second": 0.794,
      "step": 5280
    },
    {
      "epoch": 80.07575757575758,
      "grad_norm": 0.13725991547107697,
      "learning_rate": 5.996212121212121e-05,
      "loss": 0.0121,
      "step": 5285
    },
    {
      "epoch": 80.15151515151516,
      "grad_norm": 0.04899454116821289,
      "learning_rate": 5.992424242424243e-05,
      "loss": 0.009,
      "step": 5290
    },
    {
      "epoch": 80.22727272727273,
      "grad_norm": 0.046539995819330215,
      "learning_rate": 5.988636363636364e-05,
      "loss": 0.0111,
      "step": 5295
    },
    {
      "epoch": 80.3030303030303,
      "grad_norm": 0.08182474970817566,
      "learning_rate": 5.9848484848484854e-05,
      "loss": 0.0104,
      "step": 5300
    },
    {
      "epoch": 80.37878787878788,
      "grad_norm": 0.03240705281496048,
      "learning_rate": 5.9810606060606064e-05,
      "loss": 0.0201,
      "step": 5305
    },
    {
      "epoch": 80.45454545454545,
      "grad_norm": 0.059028830379247665,
      "learning_rate": 5.977272727272728e-05,
      "loss": 0.0141,
      "step": 5310
    },
    {
      "epoch": 80.53030303030303,
      "grad_norm": 0.16419073939323425,
      "learning_rate": 5.9734848484848483e-05,
      "loss": 0.0118,
      "step": 5315
    },
    {
      "epoch": 80.60606060606061,
      "grad_norm": 0.051342036575078964,
      "learning_rate": 5.969696969696969e-05,
      "loss": 0.0127,
      "step": 5320
    },
    {
      "epoch": 80.68181818181819,
      "grad_norm": 0.11851360648870468,
      "learning_rate": 5.965909090909091e-05,
      "loss": 0.0133,
      "step": 5325
    },
    {
      "epoch": 80.75757575757575,
      "grad_norm": 0.05440409854054451,
      "learning_rate": 5.962121212121212e-05,
      "loss": 0.0147,
      "step": 5330
    },
    {
      "epoch": 80.83333333333333,
      "grad_norm": 0.20603637397289276,
      "learning_rate": 5.9583333333333336e-05,
      "loss": 0.0245,
      "step": 5335
    },
    {
      "epoch": 80.9090909090909,
      "grad_norm": 0.11738941818475723,
      "learning_rate": 5.9545454545454546e-05,
      "loss": 0.009,
      "step": 5340
    },
    {
      "epoch": 80.98484848484848,
      "grad_norm": 0.05742305517196655,
      "learning_rate": 5.950757575757576e-05,
      "loss": 0.0126,
      "step": 5345
    },
    {
      "epoch": 81.0,
      "eval_loss": 0.02780001237988472,
      "eval_mean_accuracy": 0.9397194637747933,
      "eval_mean_iou": 0.9104387627282181,
      "eval_runtime": 37.9669,
      "eval_samples_per_second": 6.19,
      "eval_steps_per_second": 0.79,
      "step": 5346
    },
    {
      "epoch": 81.06060606060606,
      "grad_norm": 0.05530901625752449,
      "learning_rate": 5.946969696969697e-05,
      "loss": 0.0353,
      "step": 5350
    },
    {
      "epoch": 81.13636363636364,
      "grad_norm": 0.10689505189657211,
      "learning_rate": 5.943181818181819e-05,
      "loss": 0.0117,
      "step": 5355
    },
    {
      "epoch": 81.21212121212122,
      "grad_norm": 0.04367786645889282,
      "learning_rate": 5.93939393939394e-05,
      "loss": 0.0122,
      "step": 5360
    },
    {
      "epoch": 81.28787878787878,
      "grad_norm": 0.026988252997398376,
      "learning_rate": 5.93560606060606e-05,
      "loss": 0.0125,
      "step": 5365
    },
    {
      "epoch": 81.36363636363636,
      "grad_norm": 0.05469467490911484,
      "learning_rate": 5.931818181818182e-05,
      "loss": 0.0137,
      "step": 5370
    },
    {
      "epoch": 81.43939393939394,
      "grad_norm": 0.034001775085926056,
      "learning_rate": 5.928030303030303e-05,
      "loss": 0.0134,
      "step": 5375
    },
    {
      "epoch": 81.51515151515152,
      "grad_norm": 0.03239731118083,
      "learning_rate": 5.9242424242424244e-05,
      "loss": 0.0116,
      "step": 5380
    },
    {
      "epoch": 81.5909090909091,
      "grad_norm": 0.08132766932249069,
      "learning_rate": 5.9204545454545454e-05,
      "loss": 0.0113,
      "step": 5385
    },
    {
      "epoch": 81.66666666666667,
      "grad_norm": 0.039843056350946426,
      "learning_rate": 5.916666666666667e-05,
      "loss": 0.0108,
      "step": 5390
    },
    {
      "epoch": 81.74242424242425,
      "grad_norm": 0.0217888243496418,
      "learning_rate": 5.912878787878788e-05,
      "loss": 0.0093,
      "step": 5395
    },
    {
      "epoch": 81.81818181818181,
      "grad_norm": 0.08429674059152603,
      "learning_rate": 5.90909090909091e-05,
      "loss": 0.0143,
      "step": 5400
    },
    {
      "epoch": 81.89393939393939,
      "grad_norm": 0.04619276896119118,
      "learning_rate": 5.905303030303031e-05,
      "loss": 0.0133,
      "step": 5405
    },
    {
      "epoch": 81.96969696969697,
      "grad_norm": 0.04300420731306076,
      "learning_rate": 5.901515151515152e-05,
      "loss": 0.0128,
      "step": 5410
    },
    {
      "epoch": 82.0,
      "eval_loss": 0.012430435977876186,
      "eval_mean_accuracy": 0.9508356091216649,
      "eval_mean_iou": 0.9219090437764786,
      "eval_runtime": 37.6501,
      "eval_samples_per_second": 6.242,
      "eval_steps_per_second": 0.797,
      "step": 5412
    },
    {
      "epoch": 82.04545454545455,
      "grad_norm": 0.04875261336565018,
      "learning_rate": 5.8977272727272726e-05,
      "loss": 0.0072,
      "step": 5415
    },
    {
      "epoch": 82.12121212121212,
      "grad_norm": 0.03968634828925133,
      "learning_rate": 5.8939393939393936e-05,
      "loss": 0.0158,
      "step": 5420
    },
    {
      "epoch": 82.1969696969697,
      "grad_norm": 0.120216503739357,
      "learning_rate": 5.890151515151515e-05,
      "loss": 0.0104,
      "step": 5425
    },
    {
      "epoch": 82.27272727272727,
      "grad_norm": 0.030231354758143425,
      "learning_rate": 5.886363636363636e-05,
      "loss": 0.0132,
      "step": 5430
    },
    {
      "epoch": 82.34848484848484,
      "grad_norm": 0.03835824877023697,
      "learning_rate": 5.882575757575758e-05,
      "loss": 0.0109,
      "step": 5435
    },
    {
      "epoch": 82.42424242424242,
      "grad_norm": 0.05523248016834259,
      "learning_rate": 5.878787878787879e-05,
      "loss": 0.0128,
      "step": 5440
    },
    {
      "epoch": 82.5,
      "grad_norm": 0.03873053938150406,
      "learning_rate": 5.8750000000000005e-05,
      "loss": 0.0078,
      "step": 5445
    },
    {
      "epoch": 82.57575757575758,
      "grad_norm": 0.08267053216695786,
      "learning_rate": 5.871212121212122e-05,
      "loss": 0.0106,
      "step": 5450
    },
    {
      "epoch": 82.65151515151516,
      "grad_norm": 0.15179097652435303,
      "learning_rate": 5.867424242424243e-05,
      "loss": 0.0126,
      "step": 5455
    },
    {
      "epoch": 82.72727272727273,
      "grad_norm": 0.04992326721549034,
      "learning_rate": 5.8636363636363634e-05,
      "loss": 0.0096,
      "step": 5460
    },
    {
      "epoch": 82.8030303030303,
      "grad_norm": 0.039940230548381805,
      "learning_rate": 5.8598484848484844e-05,
      "loss": 0.0104,
      "step": 5465
    },
    {
      "epoch": 82.87878787878788,
      "grad_norm": 0.07003412395715714,
      "learning_rate": 5.856060606060606e-05,
      "loss": 0.0115,
      "step": 5470
    },
    {
      "epoch": 82.95454545454545,
      "grad_norm": 0.04271642118692398,
      "learning_rate": 5.852272727272727e-05,
      "loss": 0.0093,
      "step": 5475
    },
    {
      "epoch": 83.0,
      "eval_loss": 0.011978263035416603,
      "eval_mean_accuracy": 0.9517377883065784,
      "eval_mean_iou": 0.9239993462525935,
      "eval_runtime": 38.0552,
      "eval_samples_per_second": 6.175,
      "eval_steps_per_second": 0.788,
      "step": 5478
    },
    {
      "epoch": 83.03030303030303,
      "grad_norm": 0.020281262695789337,
      "learning_rate": 5.848484848484849e-05,
      "loss": 0.0116,
      "step": 5480
    },
    {
      "epoch": 83.10606060606061,
      "grad_norm": 0.043769288808107376,
      "learning_rate": 5.84469696969697e-05,
      "loss": 0.0114,
      "step": 5485
    },
    {
      "epoch": 83.18181818181819,
      "grad_norm": 0.03450915217399597,
      "learning_rate": 5.840909090909091e-05,
      "loss": 0.01,
      "step": 5490
    },
    {
      "epoch": 83.25757575757575,
      "grad_norm": 0.04385576769709587,
      "learning_rate": 5.837121212121213e-05,
      "loss": 0.012,
      "step": 5495
    },
    {
      "epoch": 83.33333333333333,
      "grad_norm": 0.08134505897760391,
      "learning_rate": 5.833333333333334e-05,
      "loss": 0.0136,
      "step": 5500
    },
    {
      "epoch": 83.4090909090909,
      "grad_norm": 0.04804539307951927,
      "learning_rate": 5.8295454545454556e-05,
      "loss": 0.0115,
      "step": 5505
    },
    {
      "epoch": 83.48484848484848,
      "grad_norm": 0.05074438825249672,
      "learning_rate": 5.825757575757575e-05,
      "loss": 0.0083,
      "step": 5510
    },
    {
      "epoch": 83.56060606060606,
      "grad_norm": 0.05816187337040901,
      "learning_rate": 5.821969696969697e-05,
      "loss": 0.0121,
      "step": 5515
    },
    {
      "epoch": 83.63636363636364,
      "grad_norm": 0.11151707172393799,
      "learning_rate": 5.818181818181818e-05,
      "loss": 0.0124,
      "step": 5520
    },
    {
      "epoch": 83.71212121212122,
      "grad_norm": 0.04578902944922447,
      "learning_rate": 5.8143939393939395e-05,
      "loss": 0.0122,
      "step": 5525
    },
    {
      "epoch": 83.78787878787878,
      "grad_norm": 0.039695948362350464,
      "learning_rate": 5.810606060606061e-05,
      "loss": 0.0119,
      "step": 5530
    },
    {
      "epoch": 83.86363636363636,
      "grad_norm": 0.03422868996858597,
      "learning_rate": 5.806818181818182e-05,
      "loss": 0.0092,
      "step": 5535
    },
    {
      "epoch": 83.93939393939394,
      "grad_norm": 0.03236297890543938,
      "learning_rate": 5.803030303030304e-05,
      "loss": 0.009,
      "step": 5540
    },
    {
      "epoch": 84.0,
      "eval_loss": 0.012253516353666782,
      "eval_mean_accuracy": 0.9460659927651071,
      "eval_mean_iou": 0.919551529784646,
      "eval_runtime": 37.5924,
      "eval_samples_per_second": 6.251,
      "eval_steps_per_second": 0.798,
      "step": 5544
    },
    {
      "epoch": 84.01515151515152,
      "grad_norm": 0.04444459453225136,
      "learning_rate": 5.799242424242425e-05,
      "loss": 0.0099,
      "step": 5545
    },
    {
      "epoch": 84.0909090909091,
      "grad_norm": 0.05794501304626465,
      "learning_rate": 5.7954545454545464e-05,
      "loss": 0.0112,
      "step": 5550
    },
    {
      "epoch": 84.16666666666667,
      "grad_norm": 0.03296046331524849,
      "learning_rate": 5.7916666666666674e-05,
      "loss": 0.0167,
      "step": 5555
    },
    {
      "epoch": 84.24242424242425,
      "grad_norm": 0.07193636149168015,
      "learning_rate": 5.787878787878788e-05,
      "loss": 0.0119,
      "step": 5560
    },
    {
      "epoch": 84.31818181818181,
      "grad_norm": 0.0791182592511177,
      "learning_rate": 5.784090909090909e-05,
      "loss": 0.0142,
      "step": 5565
    },
    {
      "epoch": 84.39393939393939,
      "grad_norm": 0.1411008983850479,
      "learning_rate": 5.78030303030303e-05,
      "loss": 0.0118,
      "step": 5570
    },
    {
      "epoch": 84.46969696969697,
      "grad_norm": 0.043072160333395004,
      "learning_rate": 5.776515151515152e-05,
      "loss": 0.0099,
      "step": 5575
    },
    {
      "epoch": 84.54545454545455,
      "grad_norm": 0.04133640602231026,
      "learning_rate": 5.772727272727273e-05,
      "loss": 0.0098,
      "step": 5580
    },
    {
      "epoch": 84.62121212121212,
      "grad_norm": 0.04977918043732643,
      "learning_rate": 5.7689393939393946e-05,
      "loss": 0.0169,
      "step": 5585
    },
    {
      "epoch": 84.6969696969697,
      "grad_norm": 0.07605769485235214,
      "learning_rate": 5.7651515151515156e-05,
      "loss": 0.0126,
      "step": 5590
    },
    {
      "epoch": 84.77272727272727,
      "grad_norm": 0.09223145246505737,
      "learning_rate": 5.761363636363637e-05,
      "loss": 0.0127,
      "step": 5595
    },
    {
      "epoch": 84.84848484848484,
      "grad_norm": 0.036935970187187195,
      "learning_rate": 5.757575757575758e-05,
      "loss": 0.0099,
      "step": 5600
    },
    {
      "epoch": 84.92424242424242,
      "grad_norm": 0.033807992935180664,
      "learning_rate": 5.75378787878788e-05,
      "loss": 0.0088,
      "step": 5605
    },
    {
      "epoch": 85.0,
      "grad_norm": 0.08899591118097305,
      "learning_rate": 5.7499999999999995e-05,
      "loss": 0.0113,
      "step": 5610
    },
    {
      "epoch": 85.0,
      "eval_loss": 0.01209044735878706,
      "eval_mean_accuracy": 0.9547555623312437,
      "eval_mean_iou": 0.9243781413529661,
      "eval_runtime": 37.3111,
      "eval_samples_per_second": 6.298,
      "eval_steps_per_second": 0.804,
      "step": 5610
    },
    {
      "epoch": 85.07575757575758,
      "grad_norm": 0.04387102648615837,
      "learning_rate": 5.746212121212121e-05,
      "loss": 0.008,
      "step": 5615
    },
    {
      "epoch": 85.15151515151516,
      "grad_norm": 0.08484287559986115,
      "learning_rate": 5.742424242424243e-05,
      "loss": 0.0099,
      "step": 5620
    },
    {
      "epoch": 85.22727272727273,
      "grad_norm": 0.04197641834616661,
      "learning_rate": 5.738636363636364e-05,
      "loss": 0.0103,
      "step": 5625
    },
    {
      "epoch": 85.3030303030303,
      "grad_norm": 0.018458792939782143,
      "learning_rate": 5.7348484848484854e-05,
      "loss": 0.0125,
      "step": 5630
    },
    {
      "epoch": 85.37878787878788,
      "grad_norm": 0.0641879290342331,
      "learning_rate": 5.7310606060606064e-05,
      "loss": 0.0115,
      "step": 5635
    },
    {
      "epoch": 85.45454545454545,
      "grad_norm": 0.13399891555309296,
      "learning_rate": 5.727272727272728e-05,
      "loss": 0.0142,
      "step": 5640
    },
    {
      "epoch": 85.53030303030303,
      "grad_norm": 0.04540137201547623,
      "learning_rate": 5.723484848484849e-05,
      "loss": 0.01,
      "step": 5645
    },
    {
      "epoch": 85.60606060606061,
      "grad_norm": 0.04876042529940605,
      "learning_rate": 5.719696969696971e-05,
      "loss": 0.0101,
      "step": 5650
    },
    {
      "epoch": 85.68181818181819,
      "grad_norm": 0.10302470624446869,
      "learning_rate": 5.715909090909091e-05,
      "loss": 0.0139,
      "step": 5655
    },
    {
      "epoch": 85.75757575757575,
      "grad_norm": 0.05748822167515755,
      "learning_rate": 5.712121212121212e-05,
      "loss": 0.0106,
      "step": 5660
    },
    {
      "epoch": 85.83333333333333,
      "grad_norm": 0.10959959030151367,
      "learning_rate": 5.7083333333333336e-05,
      "loss": 0.0113,
      "step": 5665
    },
    {
      "epoch": 85.9090909090909,
      "grad_norm": 0.03975432738661766,
      "learning_rate": 5.7045454545454546e-05,
      "loss": 0.0126,
      "step": 5670
    },
    {
      "epoch": 85.98484848484848,
      "grad_norm": 0.050623178482055664,
      "learning_rate": 5.700757575757576e-05,
      "loss": 0.0083,
      "step": 5675
    },
    {
      "epoch": 86.0,
      "eval_loss": 0.011844090186059475,
      "eval_mean_accuracy": 0.9592235606148579,
      "eval_mean_iou": 0.9254347805314138,
      "eval_runtime": 37.7426,
      "eval_samples_per_second": 6.226,
      "eval_steps_per_second": 0.795,
      "step": 5676
    },
    {
      "epoch": 86.06060606060606,
      "grad_norm": 0.045534852892160416,
      "learning_rate": 5.696969696969697e-05,
      "loss": 0.0138,
      "step": 5680
    },
    {
      "epoch": 86.13636363636364,
      "grad_norm": 0.14888247847557068,
      "learning_rate": 5.693181818181819e-05,
      "loss": 0.011,
      "step": 5685
    },
    {
      "epoch": 86.21212121212122,
      "grad_norm": 0.10461845993995667,
      "learning_rate": 5.68939393939394e-05,
      "loss": 0.0079,
      "step": 5690
    },
    {
      "epoch": 86.28787878787878,
      "grad_norm": 0.07529419660568237,
      "learning_rate": 5.6856060606060615e-05,
      "loss": 0.0111,
      "step": 5695
    },
    {
      "epoch": 86.36363636363636,
      "grad_norm": 0.024968508630990982,
      "learning_rate": 5.6818181818181825e-05,
      "loss": 0.0085,
      "step": 5700
    },
    {
      "epoch": 86.43939393939394,
      "grad_norm": 0.04841415584087372,
      "learning_rate": 5.678030303030303e-05,
      "loss": 0.0146,
      "step": 5705
    },
    {
      "epoch": 86.51515151515152,
      "grad_norm": 0.08570636808872223,
      "learning_rate": 5.6742424242424244e-05,
      "loss": 0.0135,
      "step": 5710
    },
    {
      "epoch": 86.5909090909091,
      "grad_norm": 0.13056956231594086,
      "learning_rate": 5.6704545454545454e-05,
      "loss": 0.0139,
      "step": 5715
    },
    {
      "epoch": 86.66666666666667,
      "grad_norm": 0.02174949273467064,
      "learning_rate": 5.666666666666667e-05,
      "loss": 0.0087,
      "step": 5720
    },
    {
      "epoch": 86.74242424242425,
      "grad_norm": 0.07762351632118225,
      "learning_rate": 5.662878787878788e-05,
      "loss": 0.0105,
      "step": 5725
    },
    {
      "epoch": 86.81818181818181,
      "grad_norm": 0.07200455665588379,
      "learning_rate": 5.65909090909091e-05,
      "loss": 0.0117,
      "step": 5730
    },
    {
      "epoch": 86.89393939393939,
      "grad_norm": 0.04761195555329323,
      "learning_rate": 5.655303030303031e-05,
      "loss": 0.0113,
      "step": 5735
    },
    {
      "epoch": 86.96969696969697,
      "grad_norm": 0.0848870649933815,
      "learning_rate": 5.651515151515152e-05,
      "loss": 0.0118,
      "step": 5740
    },
    {
      "epoch": 87.0,
      "eval_loss": 0.01241371314972639,
      "eval_mean_accuracy": 0.9580446942978035,
      "eval_mean_iou": 0.9250777100455987,
      "eval_runtime": 38.1426,
      "eval_samples_per_second": 6.161,
      "eval_steps_per_second": 0.787,
      "step": 5742
    },
    {
      "epoch": 87.04545454545455,
      "grad_norm": 0.0707048550248146,
      "learning_rate": 5.647727272727273e-05,
      "loss": 0.0115,
      "step": 5745
    },
    {
      "epoch": 87.12121212121212,
      "grad_norm": 0.025665629655122757,
      "learning_rate": 5.643939393939395e-05,
      "loss": 0.0089,
      "step": 5750
    },
    {
      "epoch": 87.1969696969697,
      "grad_norm": 0.06208712235093117,
      "learning_rate": 5.640151515151515e-05,
      "loss": 0.0125,
      "step": 5755
    },
    {
      "epoch": 87.27272727272727,
      "grad_norm": 0.056949734687805176,
      "learning_rate": 5.636363636363636e-05,
      "loss": 0.0123,
      "step": 5760
    },
    {
      "epoch": 87.34848484848484,
      "grad_norm": 0.05626838281750679,
      "learning_rate": 5.632575757575758e-05,
      "loss": 0.0111,
      "step": 5765
    },
    {
      "epoch": 87.42424242424242,
      "grad_norm": 0.12429925054311752,
      "learning_rate": 5.628787878787879e-05,
      "loss": 0.0099,
      "step": 5770
    },
    {
      "epoch": 87.5,
      "grad_norm": 0.030479269102215767,
      "learning_rate": 5.6250000000000005e-05,
      "loss": 0.0084,
      "step": 5775
    },
    {
      "epoch": 87.57575757575758,
      "grad_norm": 0.061555929481983185,
      "learning_rate": 5.6212121212121215e-05,
      "loss": 0.0143,
      "step": 5780
    },
    {
      "epoch": 87.65151515151516,
      "grad_norm": 0.03702424839138985,
      "learning_rate": 5.617424242424243e-05,
      "loss": 0.0099,
      "step": 5785
    },
    {
      "epoch": 87.72727272727273,
      "grad_norm": 0.08771682530641556,
      "learning_rate": 5.613636363636364e-05,
      "loss": 0.0098,
      "step": 5790
    },
    {
      "epoch": 87.8030303030303,
      "grad_norm": 0.043102361261844635,
      "learning_rate": 5.609848484848486e-05,
      "loss": 0.0089,
      "step": 5795
    },
    {
      "epoch": 87.87878787878788,
      "grad_norm": 0.1446639597415924,
      "learning_rate": 5.606060606060606e-05,
      "loss": 0.012,
      "step": 5800
    },
    {
      "epoch": 87.95454545454545,
      "grad_norm": 0.048752665519714355,
      "learning_rate": 5.602272727272727e-05,
      "loss": 0.0106,
      "step": 5805
    },
    {
      "epoch": 88.0,
      "eval_loss": 0.01182025671005249,
      "eval_mean_accuracy": 0.9629315905463143,
      "eval_mean_iou": 0.9252627578731359,
      "eval_runtime": 35.9419,
      "eval_samples_per_second": 6.538,
      "eval_steps_per_second": 0.835,
      "step": 5808
    },
    {
      "epoch": 88.03030303030303,
      "grad_norm": 0.08088354766368866,
      "learning_rate": 5.598484848484849e-05,
      "loss": 0.0098,
      "step": 5810
    },
    {
      "epoch": 88.10606060606061,
      "grad_norm": 0.060566410422325134,
      "learning_rate": 5.59469696969697e-05,
      "loss": 0.0099,
      "step": 5815
    },
    {
      "epoch": 88.18181818181819,
      "grad_norm": 0.07347115874290466,
      "learning_rate": 5.5909090909090913e-05,
      "loss": 0.0088,
      "step": 5820
    },
    {
      "epoch": 88.25757575757575,
      "grad_norm": 0.036012452095746994,
      "learning_rate": 5.587121212121212e-05,
      "loss": 0.011,
      "step": 5825
    },
    {
      "epoch": 88.33333333333333,
      "grad_norm": 0.09866835922002792,
      "learning_rate": 5.583333333333334e-05,
      "loss": 0.01,
      "step": 5830
    },
    {
      "epoch": 88.4090909090909,
      "grad_norm": 0.04072349891066551,
      "learning_rate": 5.579545454545455e-05,
      "loss": 0.0102,
      "step": 5835
    },
    {
      "epoch": 88.48484848484848,
      "grad_norm": 0.14677642285823822,
      "learning_rate": 5.5757575757575766e-05,
      "loss": 0.0128,
      "step": 5840
    },
    {
      "epoch": 88.56060606060606,
      "grad_norm": 0.07439061999320984,
      "learning_rate": 5.5719696969696976e-05,
      "loss": 0.0115,
      "step": 5845
    },
    {
      "epoch": 88.63636363636364,
      "grad_norm": 0.06539683789014816,
      "learning_rate": 5.568181818181818e-05,
      "loss": 0.013,
      "step": 5850
    },
    {
      "epoch": 88.71212121212122,
      "grad_norm": 0.0733487531542778,
      "learning_rate": 5.5643939393939395e-05,
      "loss": 0.0137,
      "step": 5855
    },
    {
      "epoch": 88.78787878787878,
      "grad_norm": 0.033583130687475204,
      "learning_rate": 5.5606060606060605e-05,
      "loss": 0.0096,
      "step": 5860
    },
    {
      "epoch": 88.86363636363636,
      "grad_norm": 0.050879135727882385,
      "learning_rate": 5.556818181818182e-05,
      "loss": 0.0097,
      "step": 5865
    },
    {
      "epoch": 88.93939393939394,
      "grad_norm": 0.05321991443634033,
      "learning_rate": 5.553030303030303e-05,
      "loss": 0.009,
      "step": 5870
    },
    {
      "epoch": 89.0,
      "eval_loss": 0.011984692886471748,
      "eval_mean_accuracy": 0.9524827493999641,
      "eval_mean_iou": 0.9238197350643819,
      "eval_runtime": 36.3379,
      "eval_samples_per_second": 6.467,
      "eval_steps_per_second": 0.826,
      "step": 5874
    },
    {
      "epoch": 89.01515151515152,
      "grad_norm": 0.026051169261336327,
      "learning_rate": 5.549242424242425e-05,
      "loss": 0.0078,
      "step": 5875
    },
    {
      "epoch": 89.0909090909091,
      "grad_norm": 0.06830645352602005,
      "learning_rate": 5.545454545454546e-05,
      "loss": 0.0105,
      "step": 5880
    },
    {
      "epoch": 89.16666666666667,
      "grad_norm": 0.044922687113285065,
      "learning_rate": 5.5416666666666674e-05,
      "loss": 0.0082,
      "step": 5885
    },
    {
      "epoch": 89.24242424242425,
      "grad_norm": 0.04160988703370094,
      "learning_rate": 5.5378787878787884e-05,
      "loss": 0.0124,
      "step": 5890
    },
    {
      "epoch": 89.31818181818181,
      "grad_norm": 0.056926555931568146,
      "learning_rate": 5.53409090909091e-05,
      "loss": 0.0073,
      "step": 5895
    },
    {
      "epoch": 89.39393939393939,
      "grad_norm": 0.04986729472875595,
      "learning_rate": 5.5303030303030304e-05,
      "loss": 0.0095,
      "step": 5900
    },
    {
      "epoch": 89.46969696969697,
      "grad_norm": 0.2659572958946228,
      "learning_rate": 5.526515151515151e-05,
      "loss": 0.0116,
      "step": 5905
    },
    {
      "epoch": 89.54545454545455,
      "grad_norm": 0.06467900425195694,
      "learning_rate": 5.522727272727273e-05,
      "loss": 0.0152,
      "step": 5910
    },
    {
      "epoch": 89.62121212121212,
      "grad_norm": 0.14420247077941895,
      "learning_rate": 5.518939393939394e-05,
      "loss": 0.01,
      "step": 5915
    },
    {
      "epoch": 89.6969696969697,
      "grad_norm": 0.041974928230047226,
      "learning_rate": 5.5151515151515156e-05,
      "loss": 0.0097,
      "step": 5920
    },
    {
      "epoch": 89.77272727272727,
      "grad_norm": 0.06596788763999939,
      "learning_rate": 5.5113636363636366e-05,
      "loss": 0.0123,
      "step": 5925
    },
    {
      "epoch": 89.84848484848484,
      "grad_norm": 0.034767668694257736,
      "learning_rate": 5.507575757575758e-05,
      "loss": 0.0079,
      "step": 5930
    },
    {
      "epoch": 89.92424242424242,
      "grad_norm": 0.08259923756122589,
      "learning_rate": 5.503787878787879e-05,
      "loss": 0.0124,
      "step": 5935
    },
    {
      "epoch": 90.0,
      "grad_norm": 0.11512812227010727,
      "learning_rate": 5.500000000000001e-05,
      "loss": 0.012,
      "step": 5940
    },
    {
      "epoch": 90.0,
      "eval_loss": 0.012059438042342663,
      "eval_mean_accuracy": 0.9536937812853354,
      "eval_mean_iou": 0.923442995208657,
      "eval_runtime": 37.9567,
      "eval_samples_per_second": 6.191,
      "eval_steps_per_second": 0.79,
      "step": 5940
    },
    {
      "epoch": 90.07575757575758,
      "grad_norm": 0.04509193077683449,
      "learning_rate": 5.496212121212121e-05,
      "loss": 0.0101,
      "step": 5945
    },
    {
      "epoch": 90.15151515151516,
      "grad_norm": 0.054981522262096405,
      "learning_rate": 5.492424242424242e-05,
      "loss": 0.0102,
      "step": 5950
    },
    {
      "epoch": 90.22727272727273,
      "grad_norm": 0.025574510917067528,
      "learning_rate": 5.488636363636364e-05,
      "loss": 0.0086,
      "step": 5955
    },
    {
      "epoch": 90.3030303030303,
      "grad_norm": 0.09717600792646408,
      "learning_rate": 5.484848484848485e-05,
      "loss": 0.0122,
      "step": 5960
    },
    {
      "epoch": 90.37878787878788,
      "grad_norm": 0.07982099801301956,
      "learning_rate": 5.4810606060606064e-05,
      "loss": 0.012,
      "step": 5965
    },
    {
      "epoch": 90.45454545454545,
      "grad_norm": 0.08756443858146667,
      "learning_rate": 5.4772727272727274e-05,
      "loss": 0.0123,
      "step": 5970
    },
    {
      "epoch": 90.53030303030303,
      "grad_norm": 0.08022715896368027,
      "learning_rate": 5.473484848484849e-05,
      "loss": 0.012,
      "step": 5975
    },
    {
      "epoch": 90.60606060606061,
      "grad_norm": 0.04485933855175972,
      "learning_rate": 5.46969696969697e-05,
      "loss": 0.0068,
      "step": 5980
    },
    {
      "epoch": 90.68181818181819,
      "grad_norm": 0.08497916907072067,
      "learning_rate": 5.465909090909092e-05,
      "loss": 0.0101,
      "step": 5985
    },
    {
      "epoch": 90.75757575757575,
      "grad_norm": 0.07064521312713623,
      "learning_rate": 5.462121212121213e-05,
      "loss": 0.0144,
      "step": 5990
    },
    {
      "epoch": 90.83333333333333,
      "grad_norm": 0.0647803470492363,
      "learning_rate": 5.458333333333333e-05,
      "loss": 0.0091,
      "step": 5995
    },
    {
      "epoch": 90.9090909090909,
      "grad_norm": 0.05198395252227783,
      "learning_rate": 5.4545454545454546e-05,
      "loss": 0.0146,
      "step": 6000
    },
    {
      "epoch": 90.98484848484848,
      "grad_norm": 0.062054768204689026,
      "learning_rate": 5.4507575757575756e-05,
      "loss": 0.0076,
      "step": 6005
    },
    {
      "epoch": 91.0,
      "eval_loss": 0.011856780387461185,
      "eval_mean_accuracy": 0.9644703696038676,
      "eval_mean_iou": 0.9252690615503788,
      "eval_runtime": 38.2724,
      "eval_samples_per_second": 6.14,
      "eval_steps_per_second": 0.784,
      "step": 6006
    },
    {
      "epoch": 91.06060606060606,
      "grad_norm": 0.1371425986289978,
      "learning_rate": 5.446969696969697e-05,
      "loss": 0.0125,
      "step": 6010
    },
    {
      "epoch": 91.13636363636364,
      "grad_norm": 0.03798961266875267,
      "learning_rate": 5.443181818181818e-05,
      "loss": 0.009,
      "step": 6015
    },
    {
      "epoch": 91.21212121212122,
      "grad_norm": 0.04205932468175888,
      "learning_rate": 5.43939393939394e-05,
      "loss": 0.0098,
      "step": 6020
    },
    {
      "epoch": 91.28787878787878,
      "grad_norm": 0.10898420959711075,
      "learning_rate": 5.435606060606061e-05,
      "loss": 0.0116,
      "step": 6025
    },
    {
      "epoch": 91.36363636363636,
      "grad_norm": 0.12546664476394653,
      "learning_rate": 5.4318181818181825e-05,
      "loss": 0.0137,
      "step": 6030
    },
    {
      "epoch": 91.43939393939394,
      "grad_norm": 0.03197912499308586,
      "learning_rate": 5.4280303030303035e-05,
      "loss": 0.0113,
      "step": 6035
    },
    {
      "epoch": 91.51515151515152,
      "grad_norm": 0.06644611060619354,
      "learning_rate": 5.424242424242425e-05,
      "loss": 0.0106,
      "step": 6040
    },
    {
      "epoch": 91.5909090909091,
      "grad_norm": 0.023100942373275757,
      "learning_rate": 5.4204545454545454e-05,
      "loss": 0.0114,
      "step": 6045
    },
    {
      "epoch": 91.66666666666667,
      "grad_norm": 0.0547044612467289,
      "learning_rate": 5.4166666666666664e-05,
      "loss": 0.0082,
      "step": 6050
    },
    {
      "epoch": 91.74242424242425,
      "grad_norm": 0.07674729079008102,
      "learning_rate": 5.412878787878788e-05,
      "loss": 0.0082,
      "step": 6055
    },
    {
      "epoch": 91.81818181818181,
      "grad_norm": 0.07751553505659103,
      "learning_rate": 5.409090909090909e-05,
      "loss": 0.0121,
      "step": 6060
    },
    {
      "epoch": 91.89393939393939,
      "grad_norm": 0.01899191364645958,
      "learning_rate": 5.405303030303031e-05,
      "loss": 0.0093,
      "step": 6065
    },
    {
      "epoch": 91.96969696969697,
      "grad_norm": 0.04015909880399704,
      "learning_rate": 5.401515151515152e-05,
      "loss": 0.0104,
      "step": 6070
    },
    {
      "epoch": 92.0,
      "eval_loss": 0.011939671821892262,
      "eval_mean_accuracy": 0.9549669784683411,
      "eval_mean_iou": 0.9254030800443188,
      "eval_runtime": 38.9489,
      "eval_samples_per_second": 6.034,
      "eval_steps_per_second": 0.77,
      "step": 6072
    },
    {
      "epoch": 92.04545454545455,
      "grad_norm": 0.037428490817546844,
      "learning_rate": 5.397727272727273e-05,
      "loss": 0.0146,
      "step": 6075
    },
    {
      "epoch": 92.12121212121212,
      "grad_norm": 0.09828735142946243,
      "learning_rate": 5.393939393939394e-05,
      "loss": 0.0116,
      "step": 6080
    },
    {
      "epoch": 92.1969696969697,
      "grad_norm": 0.03378356248140335,
      "learning_rate": 5.390151515151516e-05,
      "loss": 0.0119,
      "step": 6085
    },
    {
      "epoch": 92.27272727272727,
      "grad_norm": 0.13702324032783508,
      "learning_rate": 5.386363636363637e-05,
      "loss": 0.0105,
      "step": 6090
    },
    {
      "epoch": 92.34848484848484,
      "grad_norm": 0.09515448659658432,
      "learning_rate": 5.382575757575757e-05,
      "loss": 0.012,
      "step": 6095
    },
    {
      "epoch": 92.42424242424242,
      "grad_norm": 0.07175347208976746,
      "learning_rate": 5.378787878787879e-05,
      "loss": 0.0095,
      "step": 6100
    },
    {
      "epoch": 92.5,
      "grad_norm": 0.03587895259261131,
      "learning_rate": 5.375e-05,
      "loss": 0.0106,
      "step": 6105
    },
    {
      "epoch": 92.57575757575758,
      "grad_norm": 0.052851103246212006,
      "learning_rate": 5.3712121212121215e-05,
      "loss": 0.0119,
      "step": 6110
    },
    {
      "epoch": 92.65151515151516,
      "grad_norm": 0.035065874457359314,
      "learning_rate": 5.3674242424242425e-05,
      "loss": 0.0104,
      "step": 6115
    },
    {
      "epoch": 92.72727272727273,
      "grad_norm": 0.01736905425786972,
      "learning_rate": 5.363636363636364e-05,
      "loss": 0.0091,
      "step": 6120
    },
    {
      "epoch": 92.8030303030303,
      "grad_norm": 0.025128522887825966,
      "learning_rate": 5.359848484848485e-05,
      "loss": 0.0097,
      "step": 6125
    },
    {
      "epoch": 92.87878787878788,
      "grad_norm": 0.13567227125167847,
      "learning_rate": 5.356060606060607e-05,
      "loss": 0.0109,
      "step": 6130
    },
    {
      "epoch": 92.95454545454545,
      "grad_norm": 0.034488048404455185,
      "learning_rate": 5.352272727272728e-05,
      "loss": 0.0093,
      "step": 6135
    },
    {
      "epoch": 93.0,
      "eval_loss": 0.012174108065664768,
      "eval_mean_accuracy": 0.9661168875440519,
      "eval_mean_iou": 0.9239065404117609,
      "eval_runtime": 37.9949,
      "eval_samples_per_second": 6.185,
      "eval_steps_per_second": 0.79,
      "step": 6138
    },
    {
      "epoch": 93.03030303030303,
      "grad_norm": 0.048192478716373444,
      "learning_rate": 5.348484848484848e-05,
      "loss": 0.0108,
      "step": 6140
    },
    {
      "epoch": 93.10606060606061,
      "grad_norm": 0.03638065233826637,
      "learning_rate": 5.34469696969697e-05,
      "loss": 0.0112,
      "step": 6145
    },
    {
      "epoch": 93.18181818181819,
      "grad_norm": 0.2524271309375763,
      "learning_rate": 5.340909090909091e-05,
      "loss": 0.0151,
      "step": 6150
    },
    {
      "epoch": 93.25757575757575,
      "grad_norm": 0.0416405014693737,
      "learning_rate": 5.3371212121212123e-05,
      "loss": 0.0088,
      "step": 6155
    },
    {
      "epoch": 93.33333333333333,
      "grad_norm": 0.047953248023986816,
      "learning_rate": 5.333333333333333e-05,
      "loss": 0.012,
      "step": 6160
    },
    {
      "epoch": 93.4090909090909,
      "grad_norm": 0.04203718155622482,
      "learning_rate": 5.329545454545455e-05,
      "loss": 0.0084,
      "step": 6165
    },
    {
      "epoch": 93.48484848484848,
      "grad_norm": 0.018247565254569054,
      "learning_rate": 5.325757575757576e-05,
      "loss": 0.011,
      "step": 6170
    },
    {
      "epoch": 93.56060606060606,
      "grad_norm": 0.06402482092380524,
      "learning_rate": 5.3219696969696976e-05,
      "loss": 0.0117,
      "step": 6175
    },
    {
      "epoch": 93.63636363636364,
      "grad_norm": 0.024794064462184906,
      "learning_rate": 5.3181818181818186e-05,
      "loss": 0.0101,
      "step": 6180
    },
    {
      "epoch": 93.71212121212122,
      "grad_norm": 0.045491911470890045,
      "learning_rate": 5.31439393939394e-05,
      "loss": 0.0084,
      "step": 6185
    },
    {
      "epoch": 93.78787878787878,
      "grad_norm": 0.05663684383034706,
      "learning_rate": 5.3106060606060605e-05,
      "loss": 0.0112,
      "step": 6190
    },
    {
      "epoch": 93.86363636363636,
      "grad_norm": 0.03720458969473839,
      "learning_rate": 5.3068181818181815e-05,
      "loss": 0.0111,
      "step": 6195
    },
    {
      "epoch": 93.93939393939394,
      "grad_norm": 0.029325420036911964,
      "learning_rate": 5.303030303030303e-05,
      "loss": 0.009,
      "step": 6200
    },
    {
      "epoch": 94.0,
      "eval_loss": 0.012219621799886227,
      "eval_mean_accuracy": 0.9464391354055479,
      "eval_mean_iou": 0.9206799479468354,
      "eval_runtime": 35.8769,
      "eval_samples_per_second": 6.55,
      "eval_steps_per_second": 0.836,
      "step": 6204
    },
    {
      "epoch": 94.01515151515152,
      "grad_norm": 0.06176697090268135,
      "learning_rate": 5.299242424242424e-05,
      "loss": 0.0132,
      "step": 6205
    },
    {
      "epoch": 94.0909090909091,
      "grad_norm": 0.09122408181428909,
      "learning_rate": 5.295454545454546e-05,
      "loss": 0.0123,
      "step": 6210
    },
    {
      "epoch": 94.16666666666667,
      "grad_norm": 0.03718971461057663,
      "learning_rate": 5.291666666666667e-05,
      "loss": 0.0106,
      "step": 6215
    },
    {
      "epoch": 94.24242424242425,
      "grad_norm": 0.05132707953453064,
      "learning_rate": 5.2878787878787884e-05,
      "loss": 0.0099,
      "step": 6220
    },
    {
      "epoch": 94.31818181818181,
      "grad_norm": 0.05317656323313713,
      "learning_rate": 5.2840909090909094e-05,
      "loss": 0.0091,
      "step": 6225
    },
    {
      "epoch": 94.39393939393939,
      "grad_norm": 0.03662780299782753,
      "learning_rate": 5.280303030303031e-05,
      "loss": 0.0084,
      "step": 6230
    },
    {
      "epoch": 94.46969696969697,
      "grad_norm": 0.0820523202419281,
      "learning_rate": 5.276515151515152e-05,
      "loss": 0.0134,
      "step": 6235
    },
    {
      "epoch": 94.54545454545455,
      "grad_norm": 0.04373392462730408,
      "learning_rate": 5.272727272727272e-05,
      "loss": 0.0114,
      "step": 6240
    },
    {
      "epoch": 94.62121212121212,
      "grad_norm": 0.03673020750284195,
      "learning_rate": 5.268939393939394e-05,
      "loss": 0.0132,
      "step": 6245
    },
    {
      "epoch": 94.6969696969697,
      "grad_norm": 0.051661454141139984,
      "learning_rate": 5.265151515151515e-05,
      "loss": 0.0089,
      "step": 6250
    },
    {
      "epoch": 94.77272727272727,
      "grad_norm": 0.029781755059957504,
      "learning_rate": 5.2613636363636366e-05,
      "loss": 0.0094,
      "step": 6255
    },
    {
      "epoch": 94.84848484848484,
      "grad_norm": 0.046023178845644,
      "learning_rate": 5.2575757575757576e-05,
      "loss": 0.0094,
      "step": 6260
    },
    {
      "epoch": 94.92424242424242,
      "grad_norm": 0.03694437816739082,
      "learning_rate": 5.253787878787879e-05,
      "loss": 0.0097,
      "step": 6265
    },
    {
      "epoch": 95.0,
      "grad_norm": 0.08381907641887665,
      "learning_rate": 5.25e-05,
      "loss": 0.012,
      "step": 6270
    },
    {
      "epoch": 95.0,
      "eval_loss": 0.011530919000506401,
      "eval_mean_accuracy": 0.9538900075844992,
      "eval_mean_iou": 0.925397833563634,
      "eval_runtime": 37.8283,
      "eval_samples_per_second": 6.212,
      "eval_steps_per_second": 0.793,
      "step": 6270
    },
    {
      "epoch": 95.07575757575758,
      "grad_norm": 0.02503017522394657,
      "learning_rate": 5.246212121212122e-05,
      "loss": 0.009,
      "step": 6275
    },
    {
      "epoch": 95.15151515151516,
      "grad_norm": 0.05131182819604874,
      "learning_rate": 5.242424242424243e-05,
      "loss": 0.013,
      "step": 6280
    },
    {
      "epoch": 95.22727272727273,
      "grad_norm": 0.12758226692676544,
      "learning_rate": 5.238636363636363e-05,
      "loss": 0.0117,
      "step": 6285
    },
    {
      "epoch": 95.3030303030303,
      "grad_norm": 0.024913907051086426,
      "learning_rate": 5.234848484848485e-05,
      "loss": 0.0098,
      "step": 6290
    },
    {
      "epoch": 95.37878787878788,
      "grad_norm": 0.05781831964850426,
      "learning_rate": 5.231060606060606e-05,
      "loss": 0.008,
      "step": 6295
    },
    {
      "epoch": 95.45454545454545,
      "grad_norm": 0.17411963641643524,
      "learning_rate": 5.2272727272727274e-05,
      "loss": 0.0124,
      "step": 6300
    },
    {
      "epoch": 95.53030303030303,
      "grad_norm": 0.10018199682235718,
      "learning_rate": 5.2234848484848484e-05,
      "loss": 0.0102,
      "step": 6305
    },
    {
      "epoch": 95.60606060606061,
      "grad_norm": 0.0442533940076828,
      "learning_rate": 5.21969696969697e-05,
      "loss": 0.0103,
      "step": 6310
    },
    {
      "epoch": 95.68181818181819,
      "grad_norm": 0.05044003948569298,
      "learning_rate": 5.215909090909091e-05,
      "loss": 0.0081,
      "step": 6315
    },
    {
      "epoch": 95.75757575757575,
      "grad_norm": 0.13622035086154938,
      "learning_rate": 5.212121212121213e-05,
      "loss": 0.0134,
      "step": 6320
    },
    {
      "epoch": 95.83333333333333,
      "grad_norm": 0.041710611432790756,
      "learning_rate": 5.208333333333334e-05,
      "loss": 0.0138,
      "step": 6325
    },
    {
      "epoch": 95.9090909090909,
      "grad_norm": 0.02802814356982708,
      "learning_rate": 5.204545454545455e-05,
      "loss": 0.008,
      "step": 6330
    },
    {
      "epoch": 95.98484848484848,
      "grad_norm": 0.05353168770670891,
      "learning_rate": 5.2007575757575756e-05,
      "loss": 0.0102,
      "step": 6335
    },
    {
      "epoch": 96.0,
      "eval_loss": 0.012063601054251194,
      "eval_mean_accuracy": 0.9460991450595205,
      "eval_mean_iou": 0.9199884072232244,
      "eval_runtime": 2680.6746,
      "eval_samples_per_second": 0.088,
      "eval_steps_per_second": 0.011,
      "step": 6336
    },
    {
      "epoch": 96.06060606060606,
      "grad_norm": 0.06989242881536484,
      "learning_rate": 5.1969696969696966e-05,
      "loss": 0.0123,
      "step": 6340
    },
    {
      "epoch": 96.13636363636364,
      "grad_norm": 0.03883448615670204,
      "learning_rate": 5.193181818181818e-05,
      "loss": 0.0101,
      "step": 6345
    },
    {
      "epoch": 96.21212121212122,
      "grad_norm": 0.03779088333249092,
      "learning_rate": 5.189393939393939e-05,
      "loss": 0.0093,
      "step": 6350
    },
    {
      "epoch": 96.28787878787878,
      "grad_norm": 0.057634782046079636,
      "learning_rate": 5.185606060606061e-05,
      "loss": 0.0089,
      "step": 6355
    },
    {
      "epoch": 96.36363636363636,
      "grad_norm": 0.038876257836818695,
      "learning_rate": 5.181818181818182e-05,
      "loss": 0.0109,
      "step": 6360
    },
    {
      "epoch": 96.43939393939394,
      "grad_norm": 0.055217355489730835,
      "learning_rate": 5.1780303030303035e-05,
      "loss": 0.0117,
      "step": 6365
    },
    {
      "epoch": 96.51515151515152,
      "grad_norm": 0.03494484722614288,
      "learning_rate": 5.1742424242424245e-05,
      "loss": 0.0098,
      "step": 6370
    },
    {
      "epoch": 96.5909090909091,
      "grad_norm": 0.039199914783239365,
      "learning_rate": 5.170454545454546e-05,
      "loss": 0.0096,
      "step": 6375
    },
    {
      "epoch": 96.66666666666667,
      "grad_norm": 0.0465189591050148,
      "learning_rate": 5.166666666666667e-05,
      "loss": 0.0097,
      "step": 6380
    },
    {
      "epoch": 96.74242424242425,
      "grad_norm": 0.051250480115413666,
      "learning_rate": 5.1628787878787874e-05,
      "loss": 0.0113,
      "step": 6385
    },
    {
      "epoch": 96.81818181818181,
      "grad_norm": 0.07720223814249039,
      "learning_rate": 5.159090909090909e-05,
      "loss": 0.0148,
      "step": 6390
    },
    {
      "epoch": 96.89393939393939,
      "grad_norm": 0.10584262758493423,
      "learning_rate": 5.15530303030303e-05,
      "loss": 0.0088,
      "step": 6395
    },
    {
      "epoch": 96.96969696969697,
      "grad_norm": 0.04779966175556183,
      "learning_rate": 5.151515151515152e-05,
      "loss": 0.0096,
      "step": 6400
    },
    {
      "epoch": 97.0,
      "eval_loss": 0.012011293321847916,
      "eval_mean_accuracy": 0.9463830864239501,
      "eval_mean_iou": 0.9213579350836856,
      "eval_runtime": 21.1596,
      "eval_samples_per_second": 11.106,
      "eval_steps_per_second": 1.418,
      "step": 6402
    },
    {
      "epoch": 97.04545454545455,
      "grad_norm": 0.06158211827278137,
      "learning_rate": 5.147727272727273e-05,
      "loss": 0.0103,
      "step": 6405
    },
    {
      "epoch": 97.12121212121212,
      "grad_norm": 0.10089283436536789,
      "learning_rate": 5.143939393939394e-05,
      "loss": 0.0122,
      "step": 6410
    },
    {
      "epoch": 97.1969696969697,
      "grad_norm": 0.07031193375587463,
      "learning_rate": 5.140151515151515e-05,
      "loss": 0.0106,
      "step": 6415
    },
    {
      "epoch": 97.27272727272727,
      "grad_norm": 0.07967378944158554,
      "learning_rate": 5.136363636363637e-05,
      "loss": 0.01,
      "step": 6420
    },
    {
      "epoch": 97.34848484848484,
      "grad_norm": 0.06332171708345413,
      "learning_rate": 5.132575757575758e-05,
      "loss": 0.0128,
      "step": 6425
    },
    {
      "epoch": 97.42424242424242,
      "grad_norm": 0.03093346394598484,
      "learning_rate": 5.1287878787878796e-05,
      "loss": 0.0096,
      "step": 6430
    },
    {
      "epoch": 97.5,
      "grad_norm": 0.04033166542649269,
      "learning_rate": 5.125e-05,
      "loss": 0.0092,
      "step": 6435
    },
    {
      "epoch": 97.57575757575758,
      "grad_norm": 0.03249778226017952,
      "learning_rate": 5.121212121212121e-05,
      "loss": 0.0084,
      "step": 6440
    },
    {
      "epoch": 97.65151515151516,
      "grad_norm": 0.08839982002973557,
      "learning_rate": 5.1174242424242425e-05,
      "loss": 0.0116,
      "step": 6445
    },
    {
      "epoch": 97.72727272727273,
      "grad_norm": 0.08634117245674133,
      "learning_rate": 5.1136363636363635e-05,
      "loss": 0.0092,
      "step": 6450
    },
    {
      "epoch": 97.8030303030303,
      "grad_norm": 0.028230078518390656,
      "learning_rate": 5.109848484848485e-05,
      "loss": 0.0108,
      "step": 6455
    },
    {
      "epoch": 97.87878787878788,
      "grad_norm": 0.03684651851654053,
      "learning_rate": 5.106060606060606e-05,
      "loss": 0.0097,
      "step": 6460
    },
    {
      "epoch": 97.95454545454545,
      "grad_norm": 0.046680841594934464,
      "learning_rate": 5.102272727272728e-05,
      "loss": 0.0119,
      "step": 6465
    },
    {
      "epoch": 98.0,
      "eval_loss": 0.011796515434980392,
      "eval_mean_accuracy": 0.9562821532169625,
      "eval_mean_iou": 0.9256269530967044,
      "eval_runtime": 34.5244,
      "eval_samples_per_second": 6.807,
      "eval_steps_per_second": 0.869,
      "step": 6468
    },
    {
      "epoch": 98.03030303030303,
      "grad_norm": 0.07016526162624359,
      "learning_rate": 5.098484848484849e-05,
      "loss": 0.0085,
      "step": 6470
    },
    {
      "epoch": 98.10606060606061,
      "grad_norm": 0.061788592487573624,
      "learning_rate": 5.0946969696969704e-05,
      "loss": 0.0114,
      "step": 6475
    },
    {
      "epoch": 98.18181818181819,
      "grad_norm": 0.03205444663763046,
      "learning_rate": 5.090909090909091e-05,
      "loss": 0.0097,
      "step": 6480
    },
    {
      "epoch": 98.25757575757575,
      "grad_norm": 0.06380458921194077,
      "learning_rate": 5.087121212121212e-05,
      "loss": 0.009,
      "step": 6485
    },
    {
      "epoch": 98.33333333333333,
      "grad_norm": 0.046501271426677704,
      "learning_rate": 5.0833333333333333e-05,
      "loss": 0.0113,
      "step": 6490
    },
    {
      "epoch": 98.4090909090909,
      "grad_norm": 0.030922193080186844,
      "learning_rate": 5.079545454545454e-05,
      "loss": 0.0084,
      "step": 6495
    },
    {
      "epoch": 98.48484848484848,
      "grad_norm": 0.05521837994456291,
      "learning_rate": 5.075757575757576e-05,
      "loss": 0.0087,
      "step": 6500
    },
    {
      "epoch": 98.56060606060606,
      "grad_norm": 0.037061888724565506,
      "learning_rate": 5.071969696969697e-05,
      "loss": 0.0107,
      "step": 6505
    },
    {
      "epoch": 98.63636363636364,
      "grad_norm": 0.058939456939697266,
      "learning_rate": 5.0681818181818186e-05,
      "loss": 0.0092,
      "step": 6510
    },
    {
      "epoch": 98.71212121212122,
      "grad_norm": 0.051558274775743484,
      "learning_rate": 5.0643939393939396e-05,
      "loss": 0.0091,
      "step": 6515
    },
    {
      "epoch": 98.78787878787878,
      "grad_norm": 0.0387292206287384,
      "learning_rate": 5.060606060606061e-05,
      "loss": 0.0106,
      "step": 6520
    },
    {
      "epoch": 98.86363636363636,
      "grad_norm": 0.08564168214797974,
      "learning_rate": 5.056818181818183e-05,
      "loss": 0.0115,
      "step": 6525
    },
    {
      "epoch": 98.93939393939394,
      "grad_norm": 0.08744985610246658,
      "learning_rate": 5.0530303030303025e-05,
      "loss": 0.0117,
      "step": 6530
    },
    {
      "epoch": 99.0,
      "eval_loss": 0.011519311927258968,
      "eval_mean_accuracy": 0.9591494509811498,
      "eval_mean_iou": 0.9267561468339368,
      "eval_runtime": 36.0904,
      "eval_samples_per_second": 6.511,
      "eval_steps_per_second": 0.831,
      "step": 6534
    },
    {
      "epoch": 99.01515151515152,
      "grad_norm": 0.049160879105329514,
      "learning_rate": 5.049242424242424e-05,
      "loss": 0.0125,
      "step": 6535
    },
    {
      "epoch": 99.0909090909091,
      "grad_norm": 0.035761404782533646,
      "learning_rate": 5.045454545454545e-05,
      "loss": 0.0085,
      "step": 6540
    },
    {
      "epoch": 99.16666666666667,
      "grad_norm": 0.02251742221415043,
      "learning_rate": 5.041666666666667e-05,
      "loss": 0.0084,
      "step": 6545
    },
    {
      "epoch": 99.24242424242425,
      "grad_norm": 0.02420506253838539,
      "learning_rate": 5.037878787878788e-05,
      "loss": 0.0104,
      "step": 6550
    },
    {
      "epoch": 99.31818181818181,
      "grad_norm": 0.03436785936355591,
      "learning_rate": 5.0340909090909094e-05,
      "loss": 0.0068,
      "step": 6555
    },
    {
      "epoch": 99.39393939393939,
      "grad_norm": 0.09515592455863953,
      "learning_rate": 5.030303030303031e-05,
      "loss": 0.0092,
      "step": 6560
    },
    {
      "epoch": 99.46969696969697,
      "grad_norm": 0.024656545370817184,
      "learning_rate": 5.026515151515152e-05,
      "loss": 0.0102,
      "step": 6565
    },
    {
      "epoch": 99.54545454545455,
      "grad_norm": 0.05138121545314789,
      "learning_rate": 5.022727272727274e-05,
      "loss": 0.0106,
      "step": 6570
    },
    {
      "epoch": 99.62121212121212,
      "grad_norm": 0.031164860352873802,
      "learning_rate": 5.018939393939395e-05,
      "loss": 0.0108,
      "step": 6575
    },
    {
      "epoch": 99.6969696969697,
      "grad_norm": 0.027614641934633255,
      "learning_rate": 5.015151515151515e-05,
      "loss": 0.0135,
      "step": 6580
    },
    {
      "epoch": 99.77272727272727,
      "grad_norm": 0.05616153031587601,
      "learning_rate": 5.011363636363636e-05,
      "loss": 0.0103,
      "step": 6585
    },
    {
      "epoch": 99.84848484848484,
      "grad_norm": 0.08332528918981552,
      "learning_rate": 5.0075757575757576e-05,
      "loss": 0.0113,
      "step": 6590
    },
    {
      "epoch": 99.92424242424242,
      "grad_norm": 0.09366441518068314,
      "learning_rate": 5.0037878787878786e-05,
      "loss": 0.0128,
      "step": 6595
    },
    {
      "epoch": 100.0,
      "grad_norm": 0.030269917100667953,
      "learning_rate": 5e-05,
      "loss": 0.0094,
      "step": 6600
    },
    {
      "epoch": 100.0,
      "eval_loss": 0.011539186351001263,
      "eval_mean_accuracy": 0.9513434432636041,
      "eval_mean_iou": 0.9254758653549858,
      "eval_runtime": 35.6655,
      "eval_samples_per_second": 6.589,
      "eval_steps_per_second": 0.841,
      "step": 6600
    },
    {
      "epoch": 100.07575757575758,
      "grad_norm": 0.018214555457234383,
      "learning_rate": 4.996212121212122e-05,
      "loss": 0.0089,
      "step": 6605
    },
    {
      "epoch": 100.15151515151516,
      "grad_norm": 0.06073681637644768,
      "learning_rate": 4.992424242424243e-05,
      "loss": 0.0069,
      "step": 6610
    },
    {
      "epoch": 100.22727272727273,
      "grad_norm": 0.0321633443236351,
      "learning_rate": 4.988636363636364e-05,
      "loss": 0.0147,
      "step": 6615
    },
    {
      "epoch": 100.3030303030303,
      "grad_norm": 0.06882026791572571,
      "learning_rate": 4.984848484848485e-05,
      "loss": 0.0134,
      "step": 6620
    },
    {
      "epoch": 100.37878787878788,
      "grad_norm": 0.06814602017402649,
      "learning_rate": 4.9810606060606065e-05,
      "loss": 0.0091,
      "step": 6625
    },
    {
      "epoch": 100.45454545454545,
      "grad_norm": 0.06110388785600662,
      "learning_rate": 4.9772727272727275e-05,
      "loss": 0.0113,
      "step": 6630
    },
    {
      "epoch": 100.53030303030303,
      "grad_norm": 0.037348754703998566,
      "learning_rate": 4.973484848484849e-05,
      "loss": 0.0099,
      "step": 6635
    },
    {
      "epoch": 100.60606060606061,
      "grad_norm": 0.06564950942993164,
      "learning_rate": 4.9696969696969694e-05,
      "loss": 0.0126,
      "step": 6640
    },
    {
      "epoch": 100.68181818181819,
      "grad_norm": 0.08975940197706223,
      "learning_rate": 4.965909090909091e-05,
      "loss": 0.0119,
      "step": 6645
    },
    {
      "epoch": 100.75757575757575,
      "grad_norm": 0.12491611391305923,
      "learning_rate": 4.962121212121213e-05,
      "loss": 0.0086,
      "step": 6650
    },
    {
      "epoch": 100.83333333333333,
      "grad_norm": 0.030476097017526627,
      "learning_rate": 4.958333333333334e-05,
      "loss": 0.0092,
      "step": 6655
    },
    {
      "epoch": 100.9090909090909,
      "grad_norm": 0.042186494916677475,
      "learning_rate": 4.9545454545454553e-05,
      "loss": 0.01,
      "step": 6660
    },
    {
      "epoch": 100.98484848484848,
      "grad_norm": 0.026987498626112938,
      "learning_rate": 4.9507575757575756e-05,
      "loss": 0.0105,
      "step": 6665
    },
    {
      "epoch": 101.0,
      "eval_loss": 0.012006869539618492,
      "eval_mean_accuracy": 0.9649867522905781,
      "eval_mean_iou": 0.9257836716875647,
      "eval_runtime": 39.5661,
      "eval_samples_per_second": 5.939,
      "eval_steps_per_second": 0.758,
      "step": 6666
    },
    {
      "epoch": 101.06060606060606,
      "grad_norm": 0.09767469018697739,
      "learning_rate": 4.946969696969697e-05,
      "loss": 0.0101,
      "step": 6670
    },
    {
      "epoch": 101.13636363636364,
      "grad_norm": 0.06865951418876648,
      "learning_rate": 4.943181818181818e-05,
      "loss": 0.012,
      "step": 6675
    },
    {
      "epoch": 101.21212121212122,
      "grad_norm": 0.045363765209913254,
      "learning_rate": 4.93939393939394e-05,
      "loss": 0.0119,
      "step": 6680
    },
    {
      "epoch": 101.28787878787878,
      "grad_norm": 0.05240585282444954,
      "learning_rate": 4.935606060606061e-05,
      "loss": 0.0091,
      "step": 6685
    },
    {
      "epoch": 101.36363636363636,
      "grad_norm": 0.038888078182935715,
      "learning_rate": 4.931818181818182e-05,
      "loss": 0.009,
      "step": 6690
    },
    {
      "epoch": 101.43939393939394,
      "grad_norm": 0.053481727838516235,
      "learning_rate": 4.9280303030303035e-05,
      "loss": 0.0122,
      "step": 6695
    },
    {
      "epoch": 101.51515151515152,
      "grad_norm": 0.042755041271448135,
      "learning_rate": 4.9242424242424245e-05,
      "loss": 0.0108,
      "step": 6700
    },
    {
      "epoch": 101.5909090909091,
      "grad_norm": 0.05318368226289749,
      "learning_rate": 4.920454545454546e-05,
      "loss": 0.0094,
      "step": 6705
    },
    {
      "epoch": 101.66666666666667,
      "grad_norm": 0.045214615762233734,
      "learning_rate": 4.9166666666666665e-05,
      "loss": 0.0104,
      "step": 6710
    },
    {
      "epoch": 101.74242424242425,
      "grad_norm": 0.036563362926244736,
      "learning_rate": 4.912878787878788e-05,
      "loss": 0.0097,
      "step": 6715
    },
    {
      "epoch": 101.81818181818181,
      "grad_norm": 0.06649522483348846,
      "learning_rate": 4.909090909090909e-05,
      "loss": 0.0104,
      "step": 6720
    },
    {
      "epoch": 101.89393939393939,
      "grad_norm": 0.05458517000079155,
      "learning_rate": 4.905303030303031e-05,
      "loss": 0.0131,
      "step": 6725
    },
    {
      "epoch": 101.96969696969697,
      "grad_norm": 0.05488090589642525,
      "learning_rate": 4.901515151515152e-05,
      "loss": 0.009,
      "step": 6730
    },
    {
      "epoch": 102.0,
      "eval_loss": 0.011632543988525867,
      "eval_mean_accuracy": 0.9514548145598285,
      "eval_mean_iou": 0.9247713040389166,
      "eval_runtime": 35.9286,
      "eval_samples_per_second": 6.541,
      "eval_steps_per_second": 0.835,
      "step": 6732
    },
    {
      "epoch": 102.04545454545455,
      "grad_norm": 0.061124879866838455,
      "learning_rate": 4.897727272727273e-05,
      "loss": 0.0102,
      "step": 6735
    },
    {
      "epoch": 102.12121212121212,
      "grad_norm": 0.020882749930024147,
      "learning_rate": 4.8939393939393944e-05,
      "loss": 0.008,
      "step": 6740
    },
    {
      "epoch": 102.1969696969697,
      "grad_norm": 0.050887253135442734,
      "learning_rate": 4.890151515151515e-05,
      "loss": 0.0101,
      "step": 6745
    },
    {
      "epoch": 102.27272727272727,
      "grad_norm": 0.7013552188873291,
      "learning_rate": 4.886363636363637e-05,
      "loss": 0.0143,
      "step": 6750
    },
    {
      "epoch": 102.34848484848484,
      "grad_norm": 0.04738401621580124,
      "learning_rate": 4.882575757575758e-05,
      "loss": 0.0101,
      "step": 6755
    },
    {
      "epoch": 102.42424242424242,
      "grad_norm": 0.11413994431495667,
      "learning_rate": 4.878787878787879e-05,
      "loss": 0.012,
      "step": 6760
    },
    {
      "epoch": 102.5,
      "grad_norm": 0.07265022397041321,
      "learning_rate": 4.875e-05,
      "loss": 0.0105,
      "step": 6765
    },
    {
      "epoch": 102.57575757575758,
      "grad_norm": 0.09485647082328796,
      "learning_rate": 4.8712121212121216e-05,
      "loss": 0.0116,
      "step": 6770
    },
    {
      "epoch": 102.65151515151516,
      "grad_norm": 0.5005685091018677,
      "learning_rate": 4.8674242424242425e-05,
      "loss": 0.0133,
      "step": 6775
    },
    {
      "epoch": 102.72727272727273,
      "grad_norm": 0.04144004359841347,
      "learning_rate": 4.863636363636364e-05,
      "loss": 0.0134,
      "step": 6780
    },
    {
      "epoch": 102.8030303030303,
      "grad_norm": 0.025291526690125465,
      "learning_rate": 4.859848484848485e-05,
      "loss": 0.007,
      "step": 6785
    },
    {
      "epoch": 102.87878787878788,
      "grad_norm": 0.038150861859321594,
      "learning_rate": 4.856060606060606e-05,
      "loss": 0.01,
      "step": 6790
    },
    {
      "epoch": 102.95454545454545,
      "grad_norm": 0.062906414270401,
      "learning_rate": 4.852272727272728e-05,
      "loss": 0.0105,
      "step": 6795
    },
    {
      "epoch": 103.0,
      "eval_loss": 0.012074150145053864,
      "eval_mean_accuracy": 0.9645511261800032,
      "eval_mean_iou": 0.9255165895405667,
      "eval_runtime": 37.7878,
      "eval_samples_per_second": 6.219,
      "eval_steps_per_second": 0.794,
      "step": 6798
    },
    {
      "epoch": 103.03030303030303,
      "grad_norm": 0.03307465836405754,
      "learning_rate": 4.848484848484849e-05,
      "loss": 0.0097,
      "step": 6800
    },
    {
      "epoch": 103.10606060606061,
      "grad_norm": 0.045298099517822266,
      "learning_rate": 4.8446969696969704e-05,
      "loss": 0.0125,
      "step": 6805
    },
    {
      "epoch": 103.18181818181819,
      "grad_norm": 0.14913134276866913,
      "learning_rate": 4.840909090909091e-05,
      "loss": 0.0099,
      "step": 6810
    },
    {
      "epoch": 103.25757575757575,
      "grad_norm": 0.10715030878782272,
      "learning_rate": 4.8371212121212124e-05,
      "loss": 0.0102,
      "step": 6815
    },
    {
      "epoch": 103.33333333333333,
      "grad_norm": 0.10568010807037354,
      "learning_rate": 4.8333333333333334e-05,
      "loss": 0.0125,
      "step": 6820
    },
    {
      "epoch": 103.4090909090909,
      "grad_norm": 0.027569571509957314,
      "learning_rate": 4.829545454545455e-05,
      "loss": 0.0095,
      "step": 6825
    },
    {
      "epoch": 103.48484848484848,
      "grad_norm": 0.03040696680545807,
      "learning_rate": 4.825757575757576e-05,
      "loss": 0.0085,
      "step": 6830
    },
    {
      "epoch": 103.56060606060606,
      "grad_norm": 0.08251451700925827,
      "learning_rate": 4.821969696969697e-05,
      "loss": 0.0094,
      "step": 6835
    },
    {
      "epoch": 103.63636363636364,
      "grad_norm": 0.06557298451662064,
      "learning_rate": 4.8181818181818186e-05,
      "loss": 0.0086,
      "step": 6840
    },
    {
      "epoch": 103.71212121212122,
      "grad_norm": 0.03794822841882706,
      "learning_rate": 4.8143939393939396e-05,
      "loss": 0.0105,
      "step": 6845
    },
    {
      "epoch": 103.78787878787878,
      "grad_norm": 0.05270567536354065,
      "learning_rate": 4.810606060606061e-05,
      "loss": 0.0116,
      "step": 6850
    },
    {
      "epoch": 103.86363636363636,
      "grad_norm": 0.07539138197898865,
      "learning_rate": 4.8068181818181816e-05,
      "loss": 0.0115,
      "step": 6855
    },
    {
      "epoch": 103.93939393939394,
      "grad_norm": 0.036371152848005295,
      "learning_rate": 4.803030303030303e-05,
      "loss": 0.0104,
      "step": 6860
    },
    {
      "epoch": 104.0,
      "eval_loss": 0.011841142550110817,
      "eval_mean_accuracy": 0.9629224591969185,
      "eval_mean_iou": 0.9249570152789081,
      "eval_runtime": 42.4197,
      "eval_samples_per_second": 5.54,
      "eval_steps_per_second": 0.707,
      "step": 6864
    },
    {
      "epoch": 104.01515151515152,
      "grad_norm": 0.08802874386310577,
      "learning_rate": 4.799242424242424e-05,
      "loss": 0.0156,
      "step": 6865
    },
    {
      "epoch": 104.0909090909091,
      "grad_norm": 0.0347924642264843,
      "learning_rate": 4.795454545454546e-05,
      "loss": 0.0081,
      "step": 6870
    },
    {
      "epoch": 104.16666666666667,
      "grad_norm": 0.10149374604225159,
      "learning_rate": 4.791666666666667e-05,
      "loss": 0.0121,
      "step": 6875
    },
    {
      "epoch": 104.24242424242425,
      "grad_norm": 0.07136045396327972,
      "learning_rate": 4.787878787878788e-05,
      "loss": 0.0107,
      "step": 6880
    },
    {
      "epoch": 104.31818181818181,
      "grad_norm": 0.01910111866891384,
      "learning_rate": 4.7840909090909094e-05,
      "loss": 0.0097,
      "step": 6885
    },
    {
      "epoch": 104.39393939393939,
      "grad_norm": 0.08469144999980927,
      "learning_rate": 4.7803030303030304e-05,
      "loss": 0.0141,
      "step": 6890
    },
    {
      "epoch": 104.46969696969697,
      "grad_norm": 0.05533278360962868,
      "learning_rate": 4.776515151515152e-05,
      "loss": 0.011,
      "step": 6895
    },
    {
      "epoch": 104.54545454545455,
      "grad_norm": 0.03108971379697323,
      "learning_rate": 4.772727272727273e-05,
      "loss": 0.0077,
      "step": 6900
    },
    {
      "epoch": 104.62121212121212,
      "grad_norm": 0.10320001095533371,
      "learning_rate": 4.768939393939394e-05,
      "loss": 0.0125,
      "step": 6905
    },
    {
      "epoch": 104.6969696969697,
      "grad_norm": 0.032905325293540955,
      "learning_rate": 4.765151515151515e-05,
      "loss": 0.0102,
      "step": 6910
    },
    {
      "epoch": 104.77272727272727,
      "grad_norm": 0.07492777705192566,
      "learning_rate": 4.7613636363636367e-05,
      "loss": 0.0113,
      "step": 6915
    },
    {
      "epoch": 104.84848484848484,
      "grad_norm": 0.04699547588825226,
      "learning_rate": 4.7575757575757576e-05,
      "loss": 0.0086,
      "step": 6920
    },
    {
      "epoch": 104.92424242424242,
      "grad_norm": 0.026096070185303688,
      "learning_rate": 4.753787878787879e-05,
      "loss": 0.01,
      "step": 6925
    },
    {
      "epoch": 105.0,
      "grad_norm": 0.08026830852031708,
      "learning_rate": 4.75e-05,
      "loss": 0.0108,
      "step": 6930
    },
    {
      "epoch": 105.0,
      "eval_loss": 0.012310993857681751,
      "eval_mean_accuracy": 0.9419552428682056,
      "eval_mean_iou": 0.9179153691170366,
      "eval_runtime": 39.0389,
      "eval_samples_per_second": 6.02,
      "eval_steps_per_second": 0.768,
      "step": 6930
    },
    {
      "epoch": 105.07575757575758,
      "grad_norm": 0.05963383615016937,
      "learning_rate": 4.746212121212121e-05,
      "loss": 0.0109,
      "step": 6935
    },
    {
      "epoch": 105.15151515151516,
      "grad_norm": 0.04898751154541969,
      "learning_rate": 4.742424242424243e-05,
      "loss": 0.0093,
      "step": 6940
    },
    {
      "epoch": 105.22727272727273,
      "grad_norm": 0.0720474123954773,
      "learning_rate": 4.738636363636364e-05,
      "loss": 0.0084,
      "step": 6945
    },
    {
      "epoch": 105.3030303030303,
      "grad_norm": 0.1247444599866867,
      "learning_rate": 4.7348484848484855e-05,
      "loss": 0.0113,
      "step": 6950
    },
    {
      "epoch": 105.37878787878788,
      "grad_norm": 0.0300770103931427,
      "learning_rate": 4.731060606060606e-05,
      "loss": 0.0085,
      "step": 6955
    },
    {
      "epoch": 105.45454545454545,
      "grad_norm": 0.0781722143292427,
      "learning_rate": 4.7272727272727275e-05,
      "loss": 0.012,
      "step": 6960
    },
    {
      "epoch": 105.53030303030303,
      "grad_norm": 0.11073745787143707,
      "learning_rate": 4.7234848484848485e-05,
      "loss": 0.0093,
      "step": 6965
    },
    {
      "epoch": 105.60606060606061,
      "grad_norm": 0.10823550820350647,
      "learning_rate": 4.71969696969697e-05,
      "loss": 0.0103,
      "step": 6970
    },
    {
      "epoch": 105.68181818181819,
      "grad_norm": 0.06218842417001724,
      "learning_rate": 4.715909090909091e-05,
      "loss": 0.0115,
      "step": 6975
    },
    {
      "epoch": 105.75757575757575,
      "grad_norm": 0.03239862993359566,
      "learning_rate": 4.712121212121212e-05,
      "loss": 0.0101,
      "step": 6980
    },
    {
      "epoch": 105.83333333333333,
      "grad_norm": 0.08355525881052017,
      "learning_rate": 4.708333333333334e-05,
      "loss": 0.0119,
      "step": 6985
    },
    {
      "epoch": 105.9090909090909,
      "grad_norm": 0.03316887840628624,
      "learning_rate": 4.704545454545455e-05,
      "loss": 0.0075,
      "step": 6990
    },
    {
      "epoch": 105.98484848484848,
      "grad_norm": 0.04796439781785011,
      "learning_rate": 4.7007575757575763e-05,
      "loss": 0.0132,
      "step": 6995
    },
    {
      "epoch": 106.0,
      "eval_loss": 0.01150845643132925,
      "eval_mean_accuracy": 0.9609921850237192,
      "eval_mean_iou": 0.927750064957332,
      "eval_runtime": 46.0303,
      "eval_samples_per_second": 5.105,
      "eval_steps_per_second": 0.652,
      "step": 6996
    },
    {
      "epoch": 106.06060606060606,
      "grad_norm": 0.044679198414087296,
      "learning_rate": 4.696969696969697e-05,
      "loss": 0.0094,
      "step": 7000
    },
    {
      "epoch": 106.13636363636364,
      "grad_norm": 0.12504459917545319,
      "learning_rate": 4.693181818181818e-05,
      "loss": 0.0107,
      "step": 7005
    },
    {
      "epoch": 106.21212121212122,
      "grad_norm": 0.028727315366268158,
      "learning_rate": 4.689393939393939e-05,
      "loss": 0.0089,
      "step": 7010
    },
    {
      "epoch": 106.28787878787878,
      "grad_norm": 0.16648826003074646,
      "learning_rate": 4.685606060606061e-05,
      "loss": 0.0113,
      "step": 7015
    },
    {
      "epoch": 106.36363636363636,
      "grad_norm": 0.16193129122257233,
      "learning_rate": 4.681818181818182e-05,
      "loss": 0.0131,
      "step": 7020
    },
    {
      "epoch": 106.43939393939394,
      "grad_norm": 0.04774007946252823,
      "learning_rate": 4.678030303030303e-05,
      "loss": 0.0134,
      "step": 7025
    },
    {
      "epoch": 106.51515151515152,
      "grad_norm": 0.048429813235998154,
      "learning_rate": 4.6742424242424245e-05,
      "loss": 0.0091,
      "step": 7030
    },
    {
      "epoch": 106.5909090909091,
      "grad_norm": 0.02864069677889347,
      "learning_rate": 4.6704545454545455e-05,
      "loss": 0.0086,
      "step": 7035
    },
    {
      "epoch": 106.66666666666667,
      "grad_norm": 0.0685524120926857,
      "learning_rate": 4.666666666666667e-05,
      "loss": 0.0085,
      "step": 7040
    },
    {
      "epoch": 106.74242424242425,
      "grad_norm": 0.03905734419822693,
      "learning_rate": 4.662878787878788e-05,
      "loss": 0.0109,
      "step": 7045
    },
    {
      "epoch": 106.81818181818181,
      "grad_norm": 0.06628044694662094,
      "learning_rate": 4.659090909090909e-05,
      "loss": 0.0111,
      "step": 7050
    },
    {
      "epoch": 106.89393939393939,
      "grad_norm": 0.04707733914256096,
      "learning_rate": 4.65530303030303e-05,
      "loss": 0.0104,
      "step": 7055
    },
    {
      "epoch": 106.96969696969697,
      "grad_norm": 0.07674609124660492,
      "learning_rate": 4.651515151515152e-05,
      "loss": 0.012,
      "step": 7060
    },
    {
      "epoch": 107.0,
      "eval_loss": 0.01146995835006237,
      "eval_mean_accuracy": 0.9572793171330217,
      "eval_mean_iou": 0.9268664136299333,
      "eval_runtime": 46.7308,
      "eval_samples_per_second": 5.029,
      "eval_steps_per_second": 0.642,
      "step": 7062
    },
    {
      "epoch": 107.04545454545455,
      "grad_norm": 0.02818676084280014,
      "learning_rate": 4.647727272727273e-05,
      "loss": 0.0098,
      "step": 7065
    },
    {
      "epoch": 107.12121212121212,
      "grad_norm": 0.06265316158533096,
      "learning_rate": 4.6439393939393944e-05,
      "loss": 0.0099,
      "step": 7070
    },
    {
      "epoch": 107.1969696969697,
      "grad_norm": 0.028576232492923737,
      "learning_rate": 4.6401515151515154e-05,
      "loss": 0.0073,
      "step": 7075
    },
    {
      "epoch": 107.27272727272727,
      "grad_norm": 0.06371459364891052,
      "learning_rate": 4.636363636363636e-05,
      "loss": 0.01,
      "step": 7080
    },
    {
      "epoch": 107.34848484848484,
      "grad_norm": 0.01904589682817459,
      "learning_rate": 4.632575757575758e-05,
      "loss": 0.0097,
      "step": 7085
    },
    {
      "epoch": 107.42424242424242,
      "grad_norm": 0.03668300062417984,
      "learning_rate": 4.628787878787879e-05,
      "loss": 0.0124,
      "step": 7090
    },
    {
      "epoch": 107.5,
      "grad_norm": 0.0280336644500494,
      "learning_rate": 4.6250000000000006e-05,
      "loss": 0.0095,
      "step": 7095
    },
    {
      "epoch": 107.57575757575758,
      "grad_norm": 0.060470689088106155,
      "learning_rate": 4.621212121212121e-05,
      "loss": 0.0068,
      "step": 7100
    },
    {
      "epoch": 107.65151515151516,
      "grad_norm": 0.02990957535803318,
      "learning_rate": 4.6174242424242426e-05,
      "loss": 0.0116,
      "step": 7105
    },
    {
      "epoch": 107.72727272727273,
      "grad_norm": 0.039525844156742096,
      "learning_rate": 4.6136363636363635e-05,
      "loss": 0.012,
      "step": 7110
    },
    {
      "epoch": 107.8030303030303,
      "grad_norm": 0.09177317470312119,
      "learning_rate": 4.609848484848485e-05,
      "loss": 0.0114,
      "step": 7115
    },
    {
      "epoch": 107.87878787878788,
      "grad_norm": 0.07570403069257736,
      "learning_rate": 4.606060606060607e-05,
      "loss": 0.0115,
      "step": 7120
    },
    {
      "epoch": 107.95454545454545,
      "grad_norm": 0.06257466971874237,
      "learning_rate": 4.602272727272727e-05,
      "loss": 0.0092,
      "step": 7125
    },
    {
      "epoch": 108.0,
      "eval_loss": 0.011306455358862877,
      "eval_mean_accuracy": 0.960343894656455,
      "eval_mean_iou": 0.9283153889668601,
      "eval_runtime": 37.1809,
      "eval_samples_per_second": 6.32,
      "eval_steps_per_second": 0.807,
      "step": 7128
    },
    {
      "epoch": 108.03030303030303,
      "grad_norm": 0.023701298981904984,
      "learning_rate": 4.598484848484849e-05,
      "loss": 0.0099,
      "step": 7130
    },
    {
      "epoch": 108.10606060606061,
      "grad_norm": 0.044543031603097916,
      "learning_rate": 4.59469696969697e-05,
      "loss": 0.009,
      "step": 7135
    },
    {
      "epoch": 108.18181818181819,
      "grad_norm": 0.040117647498846054,
      "learning_rate": 4.5909090909090914e-05,
      "loss": 0.0103,
      "step": 7140
    },
    {
      "epoch": 108.25757575757575,
      "grad_norm": 0.048889704048633575,
      "learning_rate": 4.5871212121212124e-05,
      "loss": 0.01,
      "step": 7145
    },
    {
      "epoch": 108.33333333333333,
      "grad_norm": 0.04210358485579491,
      "learning_rate": 4.5833333333333334e-05,
      "loss": 0.0069,
      "step": 7150
    },
    {
      "epoch": 108.4090909090909,
      "grad_norm": 0.05562072619795799,
      "learning_rate": 4.5795454545454544e-05,
      "loss": 0.0078,
      "step": 7155
    },
    {
      "epoch": 108.48484848484848,
      "grad_norm": 0.0422515794634819,
      "learning_rate": 4.575757575757576e-05,
      "loss": 0.0144,
      "step": 7160
    },
    {
      "epoch": 108.56060606060606,
      "grad_norm": 0.04490511119365692,
      "learning_rate": 4.571969696969698e-05,
      "loss": 0.0115,
      "step": 7165
    },
    {
      "epoch": 108.63636363636364,
      "grad_norm": 0.10601650178432465,
      "learning_rate": 4.5681818181818186e-05,
      "loss": 0.01,
      "step": 7170
    },
    {
      "epoch": 108.71212121212122,
      "grad_norm": 0.03474460542201996,
      "learning_rate": 4.5643939393939396e-05,
      "loss": 0.01,
      "step": 7175
    },
    {
      "epoch": 108.78787878787878,
      "grad_norm": 0.04237590357661247,
      "learning_rate": 4.5606060606060606e-05,
      "loss": 0.0095,
      "step": 7180
    },
    {
      "epoch": 108.86363636363636,
      "grad_norm": 0.11687144637107849,
      "learning_rate": 4.556818181818182e-05,
      "loss": 0.0119,
      "step": 7185
    },
    {
      "epoch": 108.93939393939394,
      "grad_norm": 0.0356426015496254,
      "learning_rate": 4.553030303030303e-05,
      "loss": 0.0104,
      "step": 7190
    },
    {
      "epoch": 109.0,
      "eval_loss": 0.01154931541532278,
      "eval_mean_accuracy": 0.9522986577590707,
      "eval_mean_iou": 0.9254458944341059,
      "eval_runtime": 33.4123,
      "eval_samples_per_second": 7.033,
      "eval_steps_per_second": 0.898,
      "step": 7194
    },
    {
      "epoch": 109.01515151515152,
      "grad_norm": 0.04079962521791458,
      "learning_rate": 4.549242424242424e-05,
      "loss": 0.0086,
      "step": 7195
    },
    {
      "epoch": 109.0909090909091,
      "grad_norm": 0.09283585101366043,
      "learning_rate": 4.545454545454546e-05,
      "loss": 0.0092,
      "step": 7200
    },
    {
      "epoch": 109.16666666666667,
      "grad_norm": 0.03813755512237549,
      "learning_rate": 4.541666666666667e-05,
      "loss": 0.0095,
      "step": 7205
    },
    {
      "epoch": 109.24242424242425,
      "grad_norm": 0.0308834295719862,
      "learning_rate": 4.5378787878787885e-05,
      "loss": 0.0099,
      "step": 7210
    },
    {
      "epoch": 109.31818181818181,
      "grad_norm": 0.06589963287115097,
      "learning_rate": 4.5340909090909095e-05,
      "loss": 0.0123,
      "step": 7215
    },
    {
      "epoch": 109.39393939393939,
      "grad_norm": 0.0344465970993042,
      "learning_rate": 4.5303030303030304e-05,
      "loss": 0.0102,
      "step": 7220
    },
    {
      "epoch": 109.46969696969697,
      "grad_norm": 0.02113848552107811,
      "learning_rate": 4.5265151515151514e-05,
      "loss": 0.0103,
      "step": 7225
    },
    {
      "epoch": 109.54545454545455,
      "grad_norm": 0.031061075627803802,
      "learning_rate": 4.522727272727273e-05,
      "loss": 0.0114,
      "step": 7230
    },
    {
      "epoch": 109.62121212121212,
      "grad_norm": 0.05576871708035469,
      "learning_rate": 4.518939393939394e-05,
      "loss": 0.0111,
      "step": 7235
    },
    {
      "epoch": 109.6969696969697,
      "grad_norm": 0.052739594131708145,
      "learning_rate": 4.515151515151516e-05,
      "loss": 0.01,
      "step": 7240
    },
    {
      "epoch": 109.77272727272727,
      "grad_norm": 0.06298878788948059,
      "learning_rate": 4.511363636363637e-05,
      "loss": 0.0088,
      "step": 7245
    },
    {
      "epoch": 109.84848484848484,
      "grad_norm": 0.06598102301359177,
      "learning_rate": 4.5075757575757577e-05,
      "loss": 0.0109,
      "step": 7250
    },
    {
      "epoch": 109.92424242424242,
      "grad_norm": 0.09480015933513641,
      "learning_rate": 4.503787878787879e-05,
      "loss": 0.0101,
      "step": 7255
    },
    {
      "epoch": 110.0,
      "grad_norm": 0.1140013039112091,
      "learning_rate": 4.5e-05,
      "loss": 0.0074,
      "step": 7260
    },
    {
      "epoch": 110.0,
      "eval_loss": 0.011566760949790478,
      "eval_mean_accuracy": 0.9568448051137644,
      "eval_mean_iou": 0.9270292974592215,
      "eval_runtime": 34.0052,
      "eval_samples_per_second": 6.911,
      "eval_steps_per_second": 0.882,
      "step": 7260
    },
    {
      "epoch": 110.07575757575758,
      "grad_norm": 0.05321098491549492,
      "learning_rate": 4.496212121212122e-05,
      "loss": 0.0108,
      "step": 7265
    },
    {
      "epoch": 110.15151515151516,
      "grad_norm": 0.03535084426403046,
      "learning_rate": 4.492424242424242e-05,
      "loss": 0.0104,
      "step": 7270
    },
    {
      "epoch": 110.22727272727273,
      "grad_norm": 0.03835120052099228,
      "learning_rate": 4.488636363636364e-05,
      "loss": 0.0102,
      "step": 7275
    },
    {
      "epoch": 110.3030303030303,
      "grad_norm": 0.06595432013273239,
      "learning_rate": 4.484848484848485e-05,
      "loss": 0.0104,
      "step": 7280
    },
    {
      "epoch": 110.37878787878788,
      "grad_norm": 0.09697867184877396,
      "learning_rate": 4.4810606060606065e-05,
      "loss": 0.0101,
      "step": 7285
    },
    {
      "epoch": 110.45454545454545,
      "grad_norm": 0.05611006170511246,
      "learning_rate": 4.4772727272727275e-05,
      "loss": 0.0089,
      "step": 7290
    },
    {
      "epoch": 110.53030303030303,
      "grad_norm": 0.02066979557275772,
      "learning_rate": 4.4734848484848485e-05,
      "loss": 0.0123,
      "step": 7295
    },
    {
      "epoch": 110.60606060606061,
      "grad_norm": 0.04658261686563492,
      "learning_rate": 4.46969696969697e-05,
      "loss": 0.009,
      "step": 7300
    },
    {
      "epoch": 110.68181818181819,
      "grad_norm": 0.04481909051537514,
      "learning_rate": 4.465909090909091e-05,
      "loss": 0.0126,
      "step": 7305
    },
    {
      "epoch": 110.75757575757575,
      "grad_norm": 0.05807652324438095,
      "learning_rate": 4.462121212121213e-05,
      "loss": 0.0105,
      "step": 7310
    },
    {
      "epoch": 110.83333333333333,
      "grad_norm": 0.043146975338459015,
      "learning_rate": 4.458333333333334e-05,
      "loss": 0.0088,
      "step": 7315
    },
    {
      "epoch": 110.9090909090909,
      "grad_norm": 0.08434805274009705,
      "learning_rate": 4.454545454545455e-05,
      "loss": 0.0072,
      "step": 7320
    },
    {
      "epoch": 110.98484848484848,
      "grad_norm": 0.05115601420402527,
      "learning_rate": 4.450757575757576e-05,
      "loss": 0.0101,
      "step": 7325
    },
    {
      "epoch": 111.0,
      "eval_loss": 0.011752334423363209,
      "eval_mean_accuracy": 0.9533579943404923,
      "eval_mean_iou": 0.9246253424614187,
      "eval_runtime": 35.6138,
      "eval_samples_per_second": 6.599,
      "eval_steps_per_second": 0.842,
      "step": 7326
    },
    {
      "epoch": 111.06060606060606,
      "grad_norm": 0.08878179639577866,
      "learning_rate": 4.4469696969696973e-05,
      "loss": 0.0095,
      "step": 7330
    },
    {
      "epoch": 111.13636363636364,
      "grad_norm": 0.03255094215273857,
      "learning_rate": 4.443181818181818e-05,
      "loss": 0.0099,
      "step": 7335
    },
    {
      "epoch": 111.21212121212122,
      "grad_norm": 0.09583541005849838,
      "learning_rate": 4.43939393939394e-05,
      "loss": 0.0096,
      "step": 7340
    },
    {
      "epoch": 111.28787878787878,
      "grad_norm": 0.028399396687746048,
      "learning_rate": 4.435606060606061e-05,
      "loss": 0.0109,
      "step": 7345
    },
    {
      "epoch": 111.36363636363636,
      "grad_norm": 0.04720287024974823,
      "learning_rate": 4.431818181818182e-05,
      "loss": 0.0093,
      "step": 7350
    },
    {
      "epoch": 111.43939393939394,
      "grad_norm": 0.042487289756536484,
      "learning_rate": 4.4280303030303036e-05,
      "loss": 0.009,
      "step": 7355
    },
    {
      "epoch": 111.51515151515152,
      "grad_norm": 0.09927821904420853,
      "learning_rate": 4.4242424242424246e-05,
      "loss": 0.0083,
      "step": 7360
    },
    {
      "epoch": 111.5909090909091,
      "grad_norm": 0.04834410548210144,
      "learning_rate": 4.4204545454545455e-05,
      "loss": 0.0084,
      "step": 7365
    },
    {
      "epoch": 111.66666666666667,
      "grad_norm": 0.10682867467403412,
      "learning_rate": 4.4166666666666665e-05,
      "loss": 0.01,
      "step": 7370
    },
    {
      "epoch": 111.74242424242425,
      "grad_norm": 0.05707812309265137,
      "learning_rate": 4.412878787878788e-05,
      "loss": 0.0105,
      "step": 7375
    },
    {
      "epoch": 111.81818181818181,
      "grad_norm": 0.052570004016160965,
      "learning_rate": 4.409090909090909e-05,
      "loss": 0.0108,
      "step": 7380
    },
    {
      "epoch": 111.89393939393939,
      "grad_norm": 0.04427159205079079,
      "learning_rate": 4.405303030303031e-05,
      "loss": 0.0097,
      "step": 7385
    },
    {
      "epoch": 111.96969696969697,
      "grad_norm": 0.06325545907020569,
      "learning_rate": 4.401515151515152e-05,
      "loss": 0.0116,
      "step": 7390
    },
    {
      "epoch": 112.0,
      "eval_loss": 0.01182156428694725,
      "eval_mean_accuracy": 0.9554599875742883,
      "eval_mean_iou": 0.9260621679872346,
      "eval_runtime": 34.4488,
      "eval_samples_per_second": 6.822,
      "eval_steps_per_second": 0.871,
      "step": 7392
    },
    {
      "epoch": 112.04545454545455,
      "grad_norm": 0.08302820473909378,
      "learning_rate": 4.397727272727273e-05,
      "loss": 0.0122,
      "step": 7395
    },
    {
      "epoch": 112.12121212121212,
      "grad_norm": 0.05292226001620293,
      "learning_rate": 4.3939393939393944e-05,
      "loss": 0.0099,
      "step": 7400
    },
    {
      "epoch": 112.1969696969697,
      "grad_norm": 0.05361492186784744,
      "learning_rate": 4.3901515151515154e-05,
      "loss": 0.0106,
      "step": 7405
    },
    {
      "epoch": 112.27272727272727,
      "grad_norm": 0.05836218595504761,
      "learning_rate": 4.386363636363637e-05,
      "loss": 0.01,
      "step": 7410
    },
    {
      "epoch": 112.34848484848484,
      "grad_norm": 0.04380941390991211,
      "learning_rate": 4.382575757575757e-05,
      "loss": 0.0088,
      "step": 7415
    },
    {
      "epoch": 112.42424242424242,
      "grad_norm": 0.032519884407520294,
      "learning_rate": 4.378787878787879e-05,
      "loss": 0.0094,
      "step": 7420
    },
    {
      "epoch": 112.5,
      "grad_norm": 0.022537540644407272,
      "learning_rate": 4.375e-05,
      "loss": 0.0114,
      "step": 7425
    },
    {
      "epoch": 112.57575757575758,
      "grad_norm": 0.042146217077970505,
      "learning_rate": 4.3712121212121216e-05,
      "loss": 0.0086,
      "step": 7430
    },
    {
      "epoch": 112.65151515151516,
      "grad_norm": 0.06870914250612259,
      "learning_rate": 4.3674242424242426e-05,
      "loss": 0.0088,
      "step": 7435
    },
    {
      "epoch": 112.72727272727273,
      "grad_norm": 0.03706345334649086,
      "learning_rate": 4.3636363636363636e-05,
      "loss": 0.0114,
      "step": 7440
    },
    {
      "epoch": 112.8030303030303,
      "grad_norm": 0.028785251080989838,
      "learning_rate": 4.359848484848485e-05,
      "loss": 0.011,
      "step": 7445
    },
    {
      "epoch": 112.87878787878788,
      "grad_norm": 0.05112234875559807,
      "learning_rate": 4.356060606060606e-05,
      "loss": 0.0099,
      "step": 7450
    },
    {
      "epoch": 112.95454545454545,
      "grad_norm": 0.04163798689842224,
      "learning_rate": 4.352272727272728e-05,
      "loss": 0.0125,
      "step": 7455
    },
    {
      "epoch": 113.0,
      "eval_loss": 0.01185600645840168,
      "eval_mean_accuracy": 0.9665135754918103,
      "eval_mean_iou": 0.9256760284862405,
      "eval_runtime": 38.7174,
      "eval_samples_per_second": 6.07,
      "eval_steps_per_second": 0.775,
      "step": 7458
    },
    {
      "epoch": 113.03030303030303,
      "grad_norm": 0.1261305809020996,
      "learning_rate": 4.348484848484849e-05,
      "loss": 0.0099,
      "step": 7460
    },
    {
      "epoch": 113.10606060606061,
      "grad_norm": 0.11331699788570404,
      "learning_rate": 4.34469696969697e-05,
      "loss": 0.0092,
      "step": 7465
    },
    {
      "epoch": 113.18181818181819,
      "grad_norm": 0.13711562752723694,
      "learning_rate": 4.340909090909091e-05,
      "loss": 0.0123,
      "step": 7470
    },
    {
      "epoch": 113.25757575757575,
      "grad_norm": 0.034490540623664856,
      "learning_rate": 4.3371212121212124e-05,
      "loss": 0.0095,
      "step": 7475
    },
    {
      "epoch": 113.33333333333333,
      "grad_norm": 0.045582231134176254,
      "learning_rate": 4.3333333333333334e-05,
      "loss": 0.0116,
      "step": 7480
    },
    {
      "epoch": 113.4090909090909,
      "grad_norm": 0.0788147896528244,
      "learning_rate": 4.329545454545455e-05,
      "loss": 0.0087,
      "step": 7485
    },
    {
      "epoch": 113.48484848484848,
      "grad_norm": 0.03433598205447197,
      "learning_rate": 4.325757575757576e-05,
      "loss": 0.0085,
      "step": 7490
    },
    {
      "epoch": 113.56060606060606,
      "grad_norm": 0.04410016909241676,
      "learning_rate": 4.321969696969697e-05,
      "loss": 0.0104,
      "step": 7495
    },
    {
      "epoch": 113.63636363636364,
      "grad_norm": 0.10544668138027191,
      "learning_rate": 4.318181818181819e-05,
      "loss": 0.0094,
      "step": 7500
    },
    {
      "epoch": 113.71212121212122,
      "grad_norm": 0.0801689550280571,
      "learning_rate": 4.3143939393939396e-05,
      "loss": 0.0085,
      "step": 7505
    },
    {
      "epoch": 113.78787878787878,
      "grad_norm": 0.056659914553165436,
      "learning_rate": 4.3106060606060606e-05,
      "loss": 0.0088,
      "step": 7510
    },
    {
      "epoch": 113.86363636363636,
      "grad_norm": 0.06800572574138641,
      "learning_rate": 4.3068181818181816e-05,
      "loss": 0.0105,
      "step": 7515
    },
    {
      "epoch": 113.93939393939394,
      "grad_norm": 0.05186724662780762,
      "learning_rate": 4.303030303030303e-05,
      "loss": 0.0102,
      "step": 7520
    },
    {
      "epoch": 114.0,
      "eval_loss": 0.011427159421145916,
      "eval_mean_accuracy": 0.9566195181491924,
      "eval_mean_iou": 0.9273138227164308,
      "eval_runtime": 35.4013,
      "eval_samples_per_second": 6.638,
      "eval_steps_per_second": 0.847,
      "step": 7524
    },
    {
      "epoch": 114.01515151515152,
      "grad_norm": 0.04824354127049446,
      "learning_rate": 4.299242424242424e-05,
      "loss": 0.0116,
      "step": 7525
    },
    {
      "epoch": 114.0909090909091,
      "grad_norm": 0.03439062088727951,
      "learning_rate": 4.295454545454546e-05,
      "loss": 0.0104,
      "step": 7530
    },
    {
      "epoch": 114.16666666666667,
      "grad_norm": 0.04964764043688774,
      "learning_rate": 4.291666666666667e-05,
      "loss": 0.011,
      "step": 7535
    },
    {
      "epoch": 114.24242424242425,
      "grad_norm": 0.04707629233598709,
      "learning_rate": 4.287878787878788e-05,
      "loss": 0.0103,
      "step": 7540
    },
    {
      "epoch": 114.31818181818181,
      "grad_norm": 0.03227897733449936,
      "learning_rate": 4.2840909090909095e-05,
      "loss": 0.0105,
      "step": 7545
    },
    {
      "epoch": 114.39393939393939,
      "grad_norm": 0.03118087537586689,
      "learning_rate": 4.2803030303030305e-05,
      "loss": 0.0086,
      "step": 7550
    },
    {
      "epoch": 114.46969696969697,
      "grad_norm": 0.06566847115755081,
      "learning_rate": 4.276515151515152e-05,
      "loss": 0.0101,
      "step": 7555
    },
    {
      "epoch": 114.54545454545455,
      "grad_norm": 0.0406372994184494,
      "learning_rate": 4.2727272727272724e-05,
      "loss": 0.0086,
      "step": 7560
    },
    {
      "epoch": 114.62121212121212,
      "grad_norm": 0.07635261118412018,
      "learning_rate": 4.268939393939394e-05,
      "loss": 0.0119,
      "step": 7565
    },
    {
      "epoch": 114.6969696969697,
      "grad_norm": 0.03414742276072502,
      "learning_rate": 4.265151515151515e-05,
      "loss": 0.0087,
      "step": 7570
    },
    {
      "epoch": 114.77272727272727,
      "grad_norm": 0.02686985395848751,
      "learning_rate": 4.261363636363637e-05,
      "loss": 0.0111,
      "step": 7575
    },
    {
      "epoch": 114.84848484848484,
      "grad_norm": 0.04253414645791054,
      "learning_rate": 4.257575757575758e-05,
      "loss": 0.009,
      "step": 7580
    },
    {
      "epoch": 114.92424242424242,
      "grad_norm": 0.10535052418708801,
      "learning_rate": 4.2537878787878787e-05,
      "loss": 0.0091,
      "step": 7585
    },
    {
      "epoch": 115.0,
      "grad_norm": 0.03335623815655708,
      "learning_rate": 4.25e-05,
      "loss": 0.0074,
      "step": 7590
    },
    {
      "epoch": 115.0,
      "eval_loss": 0.011401617899537086,
      "eval_mean_accuracy": 0.9578658952071478,
      "eval_mean_iou": 0.9274510383746776,
      "eval_runtime": 36.6727,
      "eval_samples_per_second": 6.408,
      "eval_steps_per_second": 0.818,
      "step": 7590
    },
    {
      "epoch": 115.07575757575758,
      "grad_norm": 0.09919910132884979,
      "learning_rate": 4.246212121212121e-05,
      "loss": 0.0108,
      "step": 7595
    },
    {
      "epoch": 115.15151515151516,
      "grad_norm": 0.05125391483306885,
      "learning_rate": 4.242424242424243e-05,
      "loss": 0.0074,
      "step": 7600
    },
    {
      "epoch": 115.22727272727273,
      "grad_norm": 0.04781901091337204,
      "learning_rate": 4.238636363636364e-05,
      "loss": 0.0082,
      "step": 7605
    },
    {
      "epoch": 115.3030303030303,
      "grad_norm": 0.05214628204703331,
      "learning_rate": 4.234848484848485e-05,
      "loss": 0.0107,
      "step": 7610
    },
    {
      "epoch": 115.37878787878788,
      "grad_norm": 0.09162145107984543,
      "learning_rate": 4.231060606060606e-05,
      "loss": 0.0088,
      "step": 7615
    },
    {
      "epoch": 115.45454545454545,
      "grad_norm": 0.06280729174613953,
      "learning_rate": 4.2272727272727275e-05,
      "loss": 0.0112,
      "step": 7620
    },
    {
      "epoch": 115.53030303030303,
      "grad_norm": 0.05043043941259384,
      "learning_rate": 4.2234848484848485e-05,
      "loss": 0.0125,
      "step": 7625
    },
    {
      "epoch": 115.60606060606061,
      "grad_norm": 0.04006355628371239,
      "learning_rate": 4.21969696969697e-05,
      "loss": 0.0101,
      "step": 7630
    },
    {
      "epoch": 115.68181818181819,
      "grad_norm": 0.05503642186522484,
      "learning_rate": 4.215909090909091e-05,
      "loss": 0.0111,
      "step": 7635
    },
    {
      "epoch": 115.75757575757575,
      "grad_norm": 0.07606371492147446,
      "learning_rate": 4.212121212121212e-05,
      "loss": 0.0112,
      "step": 7640
    },
    {
      "epoch": 115.83333333333333,
      "grad_norm": 0.044223252683877945,
      "learning_rate": 4.208333333333334e-05,
      "loss": 0.0098,
      "step": 7645
    },
    {
      "epoch": 115.9090909090909,
      "grad_norm": 0.03691675886511803,
      "learning_rate": 4.204545454545455e-05,
      "loss": 0.0085,
      "step": 7650
    },
    {
      "epoch": 115.98484848484848,
      "grad_norm": 0.029907019808888435,
      "learning_rate": 4.2007575757575764e-05,
      "loss": 0.0119,
      "step": 7655
    },
    {
      "epoch": 116.0,
      "eval_loss": 0.01146587822586298,
      "eval_mean_accuracy": 0.9563253583974513,
      "eval_mean_iou": 0.9267950741616144,
      "eval_runtime": 35.8545,
      "eval_samples_per_second": 6.554,
      "eval_steps_per_second": 0.837,
      "step": 7656
    },
    {
      "epoch": 116.06060606060606,
      "grad_norm": 0.151488795876503,
      "learning_rate": 4.196969696969697e-05,
      "loss": 0.0131,
      "step": 7660
    },
    {
      "epoch": 116.13636363636364,
      "grad_norm": 0.07661306113004684,
      "learning_rate": 4.193181818181818e-05,
      "loss": 0.0101,
      "step": 7665
    },
    {
      "epoch": 116.21212121212122,
      "grad_norm": 0.05721113085746765,
      "learning_rate": 4.189393939393939e-05,
      "loss": 0.0082,
      "step": 7670
    },
    {
      "epoch": 116.28787878787878,
      "grad_norm": 0.02952100895345211,
      "learning_rate": 4.185606060606061e-05,
      "loss": 0.0105,
      "step": 7675
    },
    {
      "epoch": 116.36363636363636,
      "grad_norm": 0.022809501737356186,
      "learning_rate": 4.181818181818182e-05,
      "loss": 0.0072,
      "step": 7680
    },
    {
      "epoch": 116.43939393939394,
      "grad_norm": 0.04448573291301727,
      "learning_rate": 4.178030303030303e-05,
      "loss": 0.0101,
      "step": 7685
    },
    {
      "epoch": 116.51515151515152,
      "grad_norm": 0.03427058458328247,
      "learning_rate": 4.1742424242424246e-05,
      "loss": 0.0108,
      "step": 7690
    },
    {
      "epoch": 116.5909090909091,
      "grad_norm": 0.059891797602176666,
      "learning_rate": 4.1704545454545456e-05,
      "loss": 0.0101,
      "step": 7695
    },
    {
      "epoch": 116.66666666666667,
      "grad_norm": 0.038300905376672745,
      "learning_rate": 4.166666666666667e-05,
      "loss": 0.0097,
      "step": 7700
    },
    {
      "epoch": 116.74242424242425,
      "grad_norm": 0.04695119708776474,
      "learning_rate": 4.1628787878787875e-05,
      "loss": 0.0077,
      "step": 7705
    },
    {
      "epoch": 116.81818181818181,
      "grad_norm": 0.03690522536635399,
      "learning_rate": 4.159090909090909e-05,
      "loss": 0.0102,
      "step": 7710
    },
    {
      "epoch": 116.89393939393939,
      "grad_norm": 0.0801897794008255,
      "learning_rate": 4.155303030303031e-05,
      "loss": 0.0103,
      "step": 7715
    },
    {
      "epoch": 116.96969696969697,
      "grad_norm": 0.03733587637543678,
      "learning_rate": 4.151515151515152e-05,
      "loss": 0.009,
      "step": 7720
    },
    {
      "epoch": 117.0,
      "eval_loss": 0.011392646469175816,
      "eval_mean_accuracy": 0.9604893687661292,
      "eval_mean_iou": 0.9279392522981739,
      "eval_runtime": 34.7478,
      "eval_samples_per_second": 6.763,
      "eval_steps_per_second": 0.863,
      "step": 7722
    },
    {
      "epoch": 117.04545454545455,
      "grad_norm": 0.03041350096464157,
      "learning_rate": 4.1477272727272734e-05,
      "loss": 0.0107,
      "step": 7725
    },
    {
      "epoch": 117.12121212121212,
      "grad_norm": 0.06420326232910156,
      "learning_rate": 4.143939393939394e-05,
      "loss": 0.0073,
      "step": 7730
    },
    {
      "epoch": 117.1969696969697,
      "grad_norm": 0.05731750279664993,
      "learning_rate": 4.1401515151515154e-05,
      "loss": 0.0108,
      "step": 7735
    },
    {
      "epoch": 117.27272727272727,
      "grad_norm": 0.02091885916888714,
      "learning_rate": 4.1363636363636364e-05,
      "loss": 0.012,
      "step": 7740
    },
    {
      "epoch": 117.34848484848484,
      "grad_norm": 0.06024421751499176,
      "learning_rate": 4.132575757575758e-05,
      "loss": 0.01,
      "step": 7745
    },
    {
      "epoch": 117.42424242424242,
      "grad_norm": 0.0672437995672226,
      "learning_rate": 4.128787878787879e-05,
      "loss": 0.0124,
      "step": 7750
    },
    {
      "epoch": 117.5,
      "grad_norm": 0.054587822407484055,
      "learning_rate": 4.125e-05,
      "loss": 0.0103,
      "step": 7755
    },
    {
      "epoch": 117.57575757575758,
      "grad_norm": 0.024738920852541924,
      "learning_rate": 4.1212121212121216e-05,
      "loss": 0.0098,
      "step": 7760
    },
    {
      "epoch": 117.65151515151516,
      "grad_norm": 0.035995300859212875,
      "learning_rate": 4.1174242424242426e-05,
      "loss": 0.0085,
      "step": 7765
    },
    {
      "epoch": 117.72727272727273,
      "grad_norm": 0.04177010804414749,
      "learning_rate": 4.113636363636364e-05,
      "loss": 0.0091,
      "step": 7770
    },
    {
      "epoch": 117.8030303030303,
      "grad_norm": 0.07871409505605698,
      "learning_rate": 4.109848484848485e-05,
      "loss": 0.0109,
      "step": 7775
    },
    {
      "epoch": 117.87878787878788,
      "grad_norm": 0.030489210039377213,
      "learning_rate": 4.106060606060606e-05,
      "loss": 0.0091,
      "step": 7780
    },
    {
      "epoch": 117.95454545454545,
      "grad_norm": 0.05413529649376869,
      "learning_rate": 4.102272727272727e-05,
      "loss": 0.0112,
      "step": 7785
    },
    {
      "epoch": 118.0,
      "eval_loss": 0.011528912000358105,
      "eval_mean_accuracy": 0.9589576173672975,
      "eval_mean_iou": 0.9273994451487426,
      "eval_runtime": 35.501,
      "eval_samples_per_second": 6.62,
      "eval_steps_per_second": 0.845,
      "step": 7788
    },
    {
      "epoch": 118.03030303030303,
      "grad_norm": 0.04203132539987564,
      "learning_rate": 4.098484848484849e-05,
      "loss": 0.0093,
      "step": 7790
    },
    {
      "epoch": 118.10606060606061,
      "grad_norm": 0.03576911613345146,
      "learning_rate": 4.09469696969697e-05,
      "loss": 0.0111,
      "step": 7795
    },
    {
      "epoch": 118.18181818181819,
      "grad_norm": 0.05897991731762886,
      "learning_rate": 4.0909090909090915e-05,
      "loss": 0.0097,
      "step": 7800
    },
    {
      "epoch": 118.25757575757575,
      "grad_norm": 0.031033039093017578,
      "learning_rate": 4.0871212121212125e-05,
      "loss": 0.0088,
      "step": 7805
    },
    {
      "epoch": 118.33333333333333,
      "grad_norm": 0.2997583746910095,
      "learning_rate": 4.0833333333333334e-05,
      "loss": 0.0127,
      "step": 7810
    },
    {
      "epoch": 118.4090909090909,
      "grad_norm": 0.10112172365188599,
      "learning_rate": 4.079545454545455e-05,
      "loss": 0.0097,
      "step": 7815
    },
    {
      "epoch": 118.48484848484848,
      "grad_norm": 0.05503738299012184,
      "learning_rate": 4.075757575757576e-05,
      "loss": 0.0102,
      "step": 7820
    },
    {
      "epoch": 118.56060606060606,
      "grad_norm": 0.12008272856473923,
      "learning_rate": 4.071969696969698e-05,
      "loss": 0.0094,
      "step": 7825
    },
    {
      "epoch": 118.63636363636364,
      "grad_norm": 0.05120488256216049,
      "learning_rate": 4.068181818181818e-05,
      "loss": 0.0093,
      "step": 7830
    },
    {
      "epoch": 118.71212121212122,
      "grad_norm": 0.046350326389074326,
      "learning_rate": 4.06439393939394e-05,
      "loss": 0.0096,
      "step": 7835
    },
    {
      "epoch": 118.78787878787878,
      "grad_norm": 0.0499519482254982,
      "learning_rate": 4.0606060606060606e-05,
      "loss": 0.0082,
      "step": 7840
    },
    {
      "epoch": 118.86363636363636,
      "grad_norm": 0.04410199820995331,
      "learning_rate": 4.056818181818182e-05,
      "loss": 0.0144,
      "step": 7845
    },
    {
      "epoch": 118.93939393939394,
      "grad_norm": 0.05844845250248909,
      "learning_rate": 4.053030303030303e-05,
      "loss": 0.0077,
      "step": 7850
    },
    {
      "epoch": 119.0,
      "eval_loss": 0.011765029281377792,
      "eval_mean_accuracy": 0.9461845709670959,
      "eval_mean_iou": 0.9222772344092475,
      "eval_runtime": 33.9275,
      "eval_samples_per_second": 6.927,
      "eval_steps_per_second": 0.884,
      "step": 7854
    },
    {
      "epoch": 119.01515151515152,
      "grad_norm": 0.0506209060549736,
      "learning_rate": 4.049242424242424e-05,
      "loss": 0.0121,
      "step": 7855
    },
    {
      "epoch": 119.0909090909091,
      "grad_norm": 0.08737337589263916,
      "learning_rate": 4.045454545454546e-05,
      "loss": 0.0091,
      "step": 7860
    },
    {
      "epoch": 119.16666666666667,
      "grad_norm": 0.17255261540412903,
      "learning_rate": 4.041666666666667e-05,
      "loss": 0.0127,
      "step": 7865
    },
    {
      "epoch": 119.24242424242425,
      "grad_norm": 0.048920463770627975,
      "learning_rate": 4.0378787878787885e-05,
      "loss": 0.0114,
      "step": 7870
    },
    {
      "epoch": 119.31818181818181,
      "grad_norm": 0.10267198085784912,
      "learning_rate": 4.034090909090909e-05,
      "loss": 0.0081,
      "step": 7875
    },
    {
      "epoch": 119.39393939393939,
      "grad_norm": 0.04326149821281433,
      "learning_rate": 4.0303030303030305e-05,
      "loss": 0.0078,
      "step": 7880
    },
    {
      "epoch": 119.46969696969697,
      "grad_norm": 0.07896868884563446,
      "learning_rate": 4.0265151515151515e-05,
      "loss": 0.0116,
      "step": 7885
    },
    {
      "epoch": 119.54545454545455,
      "grad_norm": 0.03161108121275902,
      "learning_rate": 4.022727272727273e-05,
      "loss": 0.01,
      "step": 7890
    },
    {
      "epoch": 119.62121212121212,
      "grad_norm": 0.08486254513263702,
      "learning_rate": 4.018939393939394e-05,
      "loss": 0.0095,
      "step": 7895
    },
    {
      "epoch": 119.6969696969697,
      "grad_norm": 0.033353812992572784,
      "learning_rate": 4.015151515151515e-05,
      "loss": 0.0101,
      "step": 7900
    },
    {
      "epoch": 119.77272727272727,
      "grad_norm": 0.05104556679725647,
      "learning_rate": 4.011363636363637e-05,
      "loss": 0.0126,
      "step": 7905
    },
    {
      "epoch": 119.84848484848484,
      "grad_norm": 0.02961340919137001,
      "learning_rate": 4.007575757575758e-05,
      "loss": 0.0104,
      "step": 7910
    },
    {
      "epoch": 119.92424242424242,
      "grad_norm": 0.030298568308353424,
      "learning_rate": 4.0037878787878794e-05,
      "loss": 0.0099,
      "step": 7915
    },
    {
      "epoch": 120.0,
      "grad_norm": 0.028010228648781776,
      "learning_rate": 4e-05,
      "loss": 0.008,
      "step": 7920
    },
    {
      "epoch": 120.0,
      "eval_loss": 0.011559019796550274,
      "eval_mean_accuracy": 0.959856375943602,
      "eval_mean_iou": 0.9271022091547435,
      "eval_runtime": 35.7472,
      "eval_samples_per_second": 6.574,
      "eval_steps_per_second": 0.839,
      "step": 7920
    },
    {
      "epoch": 120.07575757575758,
      "grad_norm": 0.07802994549274445,
      "learning_rate": 3.996212121212121e-05,
      "loss": 0.0099,
      "step": 7925
    },
    {
      "epoch": 120.15151515151516,
      "grad_norm": 0.0541093535721302,
      "learning_rate": 3.992424242424242e-05,
      "loss": 0.0106,
      "step": 7930
    },
    {
      "epoch": 120.22727272727273,
      "grad_norm": 0.06496841460466385,
      "learning_rate": 3.988636363636364e-05,
      "loss": 0.0097,
      "step": 7935
    },
    {
      "epoch": 120.3030303030303,
      "grad_norm": 0.07196808606386185,
      "learning_rate": 3.984848484848485e-05,
      "loss": 0.0098,
      "step": 7940
    },
    {
      "epoch": 120.37878787878788,
      "grad_norm": 0.0354468859732151,
      "learning_rate": 3.9810606060606066e-05,
      "loss": 0.0088,
      "step": 7945
    },
    {
      "epoch": 120.45454545454545,
      "grad_norm": 0.048268966376781464,
      "learning_rate": 3.9772727272727275e-05,
      "loss": 0.0133,
      "step": 7950
    },
    {
      "epoch": 120.53030303030303,
      "grad_norm": 0.03151046857237816,
      "learning_rate": 3.9734848484848485e-05,
      "loss": 0.0076,
      "step": 7955
    },
    {
      "epoch": 120.60606060606061,
      "grad_norm": 0.041045594960451126,
      "learning_rate": 3.96969696969697e-05,
      "loss": 0.0086,
      "step": 7960
    },
    {
      "epoch": 120.68181818181819,
      "grad_norm": 0.04824323207139969,
      "learning_rate": 3.965909090909091e-05,
      "loss": 0.0114,
      "step": 7965
    },
    {
      "epoch": 120.75757575757575,
      "grad_norm": 0.032620273530483246,
      "learning_rate": 3.962121212121213e-05,
      "loss": 0.0098,
      "step": 7970
    },
    {
      "epoch": 120.83333333333333,
      "grad_norm": 0.120726577937603,
      "learning_rate": 3.958333333333333e-05,
      "loss": 0.0104,
      "step": 7975
    },
    {
      "epoch": 120.9090909090909,
      "grad_norm": 0.02508428879082203,
      "learning_rate": 3.954545454545455e-05,
      "loss": 0.0079,
      "step": 7980
    },
    {
      "epoch": 120.98484848484848,
      "grad_norm": 0.06270240247249603,
      "learning_rate": 3.950757575757576e-05,
      "loss": 0.0092,
      "step": 7985
    },
    {
      "epoch": 121.0,
      "eval_loss": 0.011343863792717457,
      "eval_mean_accuracy": 0.9568757163066048,
      "eval_mean_iou": 0.9276314073012363,
      "eval_runtime": 35.1631,
      "eval_samples_per_second": 6.683,
      "eval_steps_per_second": 0.853,
      "step": 7986
    },
    {
      "epoch": 121.06060606060606,
      "grad_norm": 0.015478801913559437,
      "learning_rate": 3.9469696969696974e-05,
      "loss": 0.0096,
      "step": 7990
    },
    {
      "epoch": 121.13636363636364,
      "grad_norm": 0.0871434211730957,
      "learning_rate": 3.9431818181818184e-05,
      "loss": 0.0088,
      "step": 7995
    },
    {
      "epoch": 121.21212121212122,
      "grad_norm": 0.07840181142091751,
      "learning_rate": 3.939393939393939e-05,
      "loss": 0.0116,
      "step": 8000
    },
    {
      "epoch": 121.28787878787878,
      "grad_norm": 0.05582057312130928,
      "learning_rate": 3.935606060606061e-05,
      "loss": 0.0093,
      "step": 8005
    },
    {
      "epoch": 121.36363636363636,
      "grad_norm": 0.08227861672639847,
      "learning_rate": 3.931818181818182e-05,
      "loss": 0.0123,
      "step": 8010
    },
    {
      "epoch": 121.43939393939394,
      "grad_norm": 0.028080297634005547,
      "learning_rate": 3.9280303030303036e-05,
      "loss": 0.0098,
      "step": 8015
    },
    {
      "epoch": 121.51515151515152,
      "grad_norm": 0.04775175824761391,
      "learning_rate": 3.924242424242424e-05,
      "loss": 0.0095,
      "step": 8020
    },
    {
      "epoch": 121.5909090909091,
      "grad_norm": 0.11681926250457764,
      "learning_rate": 3.9204545454545456e-05,
      "loss": 0.009,
      "step": 8025
    },
    {
      "epoch": 121.66666666666667,
      "grad_norm": 0.028493843972682953,
      "learning_rate": 3.9166666666666665e-05,
      "loss": 0.0084,
      "step": 8030
    },
    {
      "epoch": 121.74242424242425,
      "grad_norm": 0.038892555981874466,
      "learning_rate": 3.912878787878788e-05,
      "loss": 0.0114,
      "step": 8035
    },
    {
      "epoch": 121.81818181818181,
      "grad_norm": 0.03477760776877403,
      "learning_rate": 3.909090909090909e-05,
      "loss": 0.01,
      "step": 8040
    },
    {
      "epoch": 121.89393939393939,
      "grad_norm": 0.05694646015763283,
      "learning_rate": 3.90530303030303e-05,
      "loss": 0.0096,
      "step": 8045
    },
    {
      "epoch": 121.96969696969697,
      "grad_norm": 0.03533066436648369,
      "learning_rate": 3.901515151515152e-05,
      "loss": 0.0098,
      "step": 8050
    },
    {
      "epoch": 122.0,
      "eval_loss": 0.011725744232535362,
      "eval_mean_accuracy": 0.9579326513346752,
      "eval_mean_iou": 0.9257731930136147,
      "eval_runtime": 40.6766,
      "eval_samples_per_second": 5.777,
      "eval_steps_per_second": 0.738,
      "step": 8052
    },
    {
      "epoch": 122.04545454545455,
      "grad_norm": 0.04543990269303322,
      "learning_rate": 3.897727272727273e-05,
      "loss": 0.011,
      "step": 8055
    },
    {
      "epoch": 122.12121212121212,
      "grad_norm": 0.04960137978196144,
      "learning_rate": 3.8939393939393944e-05,
      "loss": 0.0079,
      "step": 8060
    },
    {
      "epoch": 122.1969696969697,
      "grad_norm": 0.048826202750205994,
      "learning_rate": 3.8901515151515154e-05,
      "loss": 0.0103,
      "step": 8065
    },
    {
      "epoch": 122.27272727272727,
      "grad_norm": 0.034918077290058136,
      "learning_rate": 3.8863636363636364e-05,
      "loss": 0.0082,
      "step": 8070
    },
    {
      "epoch": 122.34848484848484,
      "grad_norm": 0.17813225090503693,
      "learning_rate": 3.8825757575757574e-05,
      "loss": 0.0088,
      "step": 8075
    },
    {
      "epoch": 122.42424242424242,
      "grad_norm": 0.022743700072169304,
      "learning_rate": 3.878787878787879e-05,
      "loss": 0.0087,
      "step": 8080
    },
    {
      "epoch": 122.5,
      "grad_norm": 0.04618868976831436,
      "learning_rate": 3.875e-05,
      "loss": 0.0122,
      "step": 8085
    },
    {
      "epoch": 122.57575757575758,
      "grad_norm": 0.04268888384103775,
      "learning_rate": 3.8712121212121217e-05,
      "loss": 0.0068,
      "step": 8090
    },
    {
      "epoch": 122.65151515151516,
      "grad_norm": 0.05451355129480362,
      "learning_rate": 3.8674242424242426e-05,
      "loss": 0.0104,
      "step": 8095
    },
    {
      "epoch": 122.72727272727273,
      "grad_norm": 0.0346740260720253,
      "learning_rate": 3.8636363636363636e-05,
      "loss": 0.0112,
      "step": 8100
    },
    {
      "epoch": 122.8030303030303,
      "grad_norm": 0.04337818920612335,
      "learning_rate": 3.859848484848485e-05,
      "loss": 0.0091,
      "step": 8105
    },
    {
      "epoch": 122.87878787878788,
      "grad_norm": 0.07118333131074905,
      "learning_rate": 3.856060606060606e-05,
      "loss": 0.0141,
      "step": 8110
    },
    {
      "epoch": 122.95454545454545,
      "grad_norm": 0.03622877597808838,
      "learning_rate": 3.852272727272728e-05,
      "loss": 0.0114,
      "step": 8115
    },
    {
      "epoch": 123.0,
      "eval_loss": 0.01203148439526558,
      "eval_mean_accuracy": 0.9437784102449503,
      "eval_mean_iou": 0.9198706727291786,
      "eval_runtime": 34.3971,
      "eval_samples_per_second": 6.832,
      "eval_steps_per_second": 0.872,
      "step": 8118
    },
    {
      "epoch": 123.03030303030303,
      "grad_norm": 0.043108221143484116,
      "learning_rate": 3.848484848484848e-05,
      "loss": 0.0078,
      "step": 8120
    },
    {
      "epoch": 123.10606060606061,
      "grad_norm": 0.026124270632863045,
      "learning_rate": 3.84469696969697e-05,
      "loss": 0.0111,
      "step": 8125
    },
    {
      "epoch": 123.18181818181819,
      "grad_norm": 0.03937513381242752,
      "learning_rate": 3.840909090909091e-05,
      "loss": 0.0104,
      "step": 8130
    },
    {
      "epoch": 123.25757575757575,
      "grad_norm": 0.06427106261253357,
      "learning_rate": 3.8371212121212125e-05,
      "loss": 0.0071,
      "step": 8135
    },
    {
      "epoch": 123.33333333333333,
      "grad_norm": 0.024193864315748215,
      "learning_rate": 3.8333333333333334e-05,
      "loss": 0.0085,
      "step": 8140
    },
    {
      "epoch": 123.4090909090909,
      "grad_norm": 0.04871522635221481,
      "learning_rate": 3.8295454545454544e-05,
      "loss": 0.0104,
      "step": 8145
    },
    {
      "epoch": 123.48484848484848,
      "grad_norm": 0.08775445073843002,
      "learning_rate": 3.825757575757576e-05,
      "loss": 0.0098,
      "step": 8150
    },
    {
      "epoch": 123.56060606060606,
      "grad_norm": 0.03694203868508339,
      "learning_rate": 3.821969696969697e-05,
      "loss": 0.0086,
      "step": 8155
    },
    {
      "epoch": 123.63636363636364,
      "grad_norm": 0.09518029540777206,
      "learning_rate": 3.818181818181819e-05,
      "loss": 0.0107,
      "step": 8160
    },
    {
      "epoch": 123.71212121212122,
      "grad_norm": 0.036771632730960846,
      "learning_rate": 3.81439393939394e-05,
      "loss": 0.0089,
      "step": 8165
    },
    {
      "epoch": 123.78787878787878,
      "grad_norm": 0.04062000289559364,
      "learning_rate": 3.810606060606061e-05,
      "loss": 0.0114,
      "step": 8170
    },
    {
      "epoch": 123.86363636363636,
      "grad_norm": 0.05793893709778786,
      "learning_rate": 3.8068181818181816e-05,
      "loss": 0.0092,
      "step": 8175
    },
    {
      "epoch": 123.93939393939394,
      "grad_norm": 0.034293681383132935,
      "learning_rate": 3.803030303030303e-05,
      "loss": 0.0089,
      "step": 8180
    },
    {
      "epoch": 124.0,
      "eval_loss": 0.011426865123212337,
      "eval_mean_accuracy": 0.9602345395457985,
      "eval_mean_iou": 0.9277061308384176,
      "eval_runtime": 36.7968,
      "eval_samples_per_second": 6.386,
      "eval_steps_per_second": 0.815,
      "step": 8184
    },
    {
      "epoch": 124.01515151515152,
      "grad_norm": 0.0780995786190033,
      "learning_rate": 3.799242424242424e-05,
      "loss": 0.0423,
      "step": 8185
    },
    {
      "epoch": 124.0909090909091,
      "grad_norm": 0.03279482573270798,
      "learning_rate": 3.795454545454545e-05,
      "loss": 0.0084,
      "step": 8190
    },
    {
      "epoch": 124.16666666666667,
      "grad_norm": 0.04126240313053131,
      "learning_rate": 3.791666666666667e-05,
      "loss": 0.0079,
      "step": 8195
    },
    {
      "epoch": 124.24242424242425,
      "grad_norm": 0.17293687164783478,
      "learning_rate": 3.787878787878788e-05,
      "loss": 0.0119,
      "step": 8200
    },
    {
      "epoch": 124.31818181818181,
      "grad_norm": 0.04419407248497009,
      "learning_rate": 3.7840909090909095e-05,
      "loss": 0.0097,
      "step": 8205
    },
    {
      "epoch": 124.39393939393939,
      "grad_norm": 0.038720179349184036,
      "learning_rate": 3.7803030303030305e-05,
      "loss": 0.0124,
      "step": 8210
    },
    {
      "epoch": 124.46969696969697,
      "grad_norm": 0.07850942760705948,
      "learning_rate": 3.7765151515151515e-05,
      "loss": 0.0121,
      "step": 8215
    },
    {
      "epoch": 124.54545454545455,
      "grad_norm": 0.033612463623285294,
      "learning_rate": 3.7727272727272725e-05,
      "loss": 0.0092,
      "step": 8220
    },
    {
      "epoch": 124.62121212121212,
      "grad_norm": 0.09047693759202957,
      "learning_rate": 3.768939393939394e-05,
      "loss": 0.009,
      "step": 8225
    },
    {
      "epoch": 124.6969696969697,
      "grad_norm": 0.029846273362636566,
      "learning_rate": 3.765151515151516e-05,
      "loss": 0.0098,
      "step": 8230
    },
    {
      "epoch": 124.77272727272727,
      "grad_norm": 0.07646148651838303,
      "learning_rate": 3.761363636363637e-05,
      "loss": 0.0103,
      "step": 8235
    },
    {
      "epoch": 124.84848484848484,
      "grad_norm": 0.026828819885849953,
      "learning_rate": 3.757575757575758e-05,
      "loss": 0.0083,
      "step": 8240
    },
    {
      "epoch": 124.92424242424242,
      "grad_norm": 0.023842118680477142,
      "learning_rate": 3.753787878787879e-05,
      "loss": 0.009,
      "step": 8245
    },
    {
      "epoch": 125.0,
      "grad_norm": 0.23472216725349426,
      "learning_rate": 3.7500000000000003e-05,
      "loss": 0.0114,
      "step": 8250
    },
    {
      "epoch": 125.0,
      "eval_loss": 0.011499553918838501,
      "eval_mean_accuracy": 0.954917351041079,
      "eval_mean_iou": 0.9263578198730379,
      "eval_runtime": 38.4364,
      "eval_samples_per_second": 6.114,
      "eval_steps_per_second": 0.781,
      "step": 8250
    },
    {
      "epoch": 125.07575757575758,
      "grad_norm": 0.18983297049999237,
      "learning_rate": 3.746212121212121e-05,
      "loss": 0.0104,
      "step": 8255
    },
    {
      "epoch": 125.15151515151516,
      "grad_norm": 0.06970106810331345,
      "learning_rate": 3.742424242424243e-05,
      "loss": 0.0092,
      "step": 8260
    },
    {
      "epoch": 125.22727272727273,
      "grad_norm": 0.12817084789276123,
      "learning_rate": 3.738636363636363e-05,
      "loss": 0.0113,
      "step": 8265
    },
    {
      "epoch": 125.3030303030303,
      "grad_norm": 0.05871003121137619,
      "learning_rate": 3.734848484848485e-05,
      "loss": 0.0129,
      "step": 8270
    },
    {
      "epoch": 125.37878787878788,
      "grad_norm": 0.042779311537742615,
      "learning_rate": 3.7310606060606066e-05,
      "loss": 0.0094,
      "step": 8275
    },
    {
      "epoch": 125.45454545454545,
      "grad_norm": 0.08304107189178467,
      "learning_rate": 3.7272727272727276e-05,
      "loss": 0.0115,
      "step": 8280
    },
    {
      "epoch": 125.53030303030303,
      "grad_norm": 0.11674903333187103,
      "learning_rate": 3.723484848484849e-05,
      "loss": 0.0101,
      "step": 8285
    },
    {
      "epoch": 125.60606060606061,
      "grad_norm": 0.05261913314461708,
      "learning_rate": 3.7196969696969695e-05,
      "loss": 0.0103,
      "step": 8290
    },
    {
      "epoch": 125.68181818181819,
      "grad_norm": 0.05260763689875603,
      "learning_rate": 3.715909090909091e-05,
      "loss": 0.0126,
      "step": 8295
    },
    {
      "epoch": 125.75757575757575,
      "grad_norm": 0.07079128921031952,
      "learning_rate": 3.712121212121212e-05,
      "loss": 0.0084,
      "step": 8300
    },
    {
      "epoch": 125.83333333333333,
      "grad_norm": 0.13524267077445984,
      "learning_rate": 3.708333333333334e-05,
      "loss": 0.0104,
      "step": 8305
    },
    {
      "epoch": 125.9090909090909,
      "grad_norm": 0.039092425256967545,
      "learning_rate": 3.704545454545455e-05,
      "loss": 0.009,
      "step": 8310
    },
    {
      "epoch": 125.98484848484848,
      "grad_norm": 0.04737089201807976,
      "learning_rate": 3.700757575757576e-05,
      "loss": 0.0088,
      "step": 8315
    },
    {
      "epoch": 126.0,
      "eval_loss": 0.011696246452629566,
      "eval_mean_accuracy": 0.9557720251913508,
      "eval_mean_iou": 0.9254455155795881,
      "eval_runtime": 34.2258,
      "eval_samples_per_second": 6.866,
      "eval_steps_per_second": 0.877,
      "step": 8316
    },
    {
      "epoch": 126.06060606060606,
      "grad_norm": 0.07828246057033539,
      "learning_rate": 3.6969696969696974e-05,
      "loss": 0.0095,
      "step": 8320
    },
    {
      "epoch": 126.13636363636364,
      "grad_norm": 0.06909701973199844,
      "learning_rate": 3.6931818181818184e-05,
      "loss": 0.009,
      "step": 8325
    },
    {
      "epoch": 126.21212121212122,
      "grad_norm": 0.10240132361650467,
      "learning_rate": 3.68939393939394e-05,
      "loss": 0.0085,
      "step": 8330
    },
    {
      "epoch": 126.28787878787878,
      "grad_norm": 0.04975391924381256,
      "learning_rate": 3.68560606060606e-05,
      "loss": 0.0121,
      "step": 8335
    },
    {
      "epoch": 126.36363636363636,
      "grad_norm": 0.05571567639708519,
      "learning_rate": 3.681818181818182e-05,
      "loss": 0.0092,
      "step": 8340
    },
    {
      "epoch": 126.43939393939394,
      "grad_norm": 0.04559260234236717,
      "learning_rate": 3.678030303030303e-05,
      "loss": 0.0125,
      "step": 8345
    },
    {
      "epoch": 126.51515151515152,
      "grad_norm": 0.15386709570884705,
      "learning_rate": 3.6742424242424246e-05,
      "loss": 0.0081,
      "step": 8350
    },
    {
      "epoch": 126.5909090909091,
      "grad_norm": 0.035543836653232574,
      "learning_rate": 3.6704545454545456e-05,
      "loss": 0.0107,
      "step": 8355
    },
    {
      "epoch": 126.66666666666667,
      "grad_norm": 0.037887975573539734,
      "learning_rate": 3.6666666666666666e-05,
      "loss": 0.0073,
      "step": 8360
    },
    {
      "epoch": 126.74242424242425,
      "grad_norm": 0.04327854886651039,
      "learning_rate": 3.662878787878788e-05,
      "loss": 0.0117,
      "step": 8365
    },
    {
      "epoch": 126.81818181818181,
      "grad_norm": 0.04229844734072685,
      "learning_rate": 3.659090909090909e-05,
      "loss": 0.0091,
      "step": 8370
    },
    {
      "epoch": 126.89393939393939,
      "grad_norm": 0.0372186005115509,
      "learning_rate": 3.655303030303031e-05,
      "loss": 0.0108,
      "step": 8375
    },
    {
      "epoch": 126.96969696969697,
      "grad_norm": 0.08369189500808716,
      "learning_rate": 3.651515151515152e-05,
      "loss": 0.0119,
      "step": 8380
    },
    {
      "epoch": 127.0,
      "eval_loss": 0.011482066474854946,
      "eval_mean_accuracy": 0.9556383660074533,
      "eval_mean_iou": 0.9264753379120841,
      "eval_runtime": 35.3872,
      "eval_samples_per_second": 6.641,
      "eval_steps_per_second": 0.848,
      "step": 8382
    },
    {
      "epoch": 127.04545454545455,
      "grad_norm": 0.036103442311286926,
      "learning_rate": 3.647727272727273e-05,
      "loss": 0.013,
      "step": 8385
    },
    {
      "epoch": 127.12121212121212,
      "grad_norm": 0.10117600858211517,
      "learning_rate": 3.643939393939394e-05,
      "loss": 0.0129,
      "step": 8390
    },
    {
      "epoch": 127.1969696969697,
      "grad_norm": 0.039770618081092834,
      "learning_rate": 3.6401515151515154e-05,
      "loss": 0.0085,
      "step": 8395
    },
    {
      "epoch": 127.27272727272727,
      "grad_norm": 0.05471545457839966,
      "learning_rate": 3.6363636363636364e-05,
      "loss": 0.0094,
      "step": 8400
    },
    {
      "epoch": 127.34848484848484,
      "grad_norm": 0.05394561588764191,
      "learning_rate": 3.632575757575758e-05,
      "loss": 0.012,
      "step": 8405
    },
    {
      "epoch": 127.42424242424242,
      "grad_norm": 0.04250704497098923,
      "learning_rate": 3.628787878787879e-05,
      "loss": 0.0097,
      "step": 8410
    },
    {
      "epoch": 127.5,
      "grad_norm": 0.04030847176909447,
      "learning_rate": 3.625e-05,
      "loss": 0.0071,
      "step": 8415
    },
    {
      "epoch": 127.57575757575758,
      "grad_norm": 0.051118411123752594,
      "learning_rate": 3.621212121212122e-05,
      "loss": 0.0091,
      "step": 8420
    },
    {
      "epoch": 127.65151515151516,
      "grad_norm": 0.03594302013516426,
      "learning_rate": 3.6174242424242427e-05,
      "loss": 0.0125,
      "step": 8425
    },
    {
      "epoch": 127.72727272727273,
      "grad_norm": 0.025613170117139816,
      "learning_rate": 3.613636363636364e-05,
      "loss": 0.0094,
      "step": 8430
    },
    {
      "epoch": 127.8030303030303,
      "grad_norm": 0.06229902058839798,
      "learning_rate": 3.6098484848484846e-05,
      "loss": 0.0098,
      "step": 8435
    },
    {
      "epoch": 127.87878787878788,
      "grad_norm": 0.04644754156470299,
      "learning_rate": 3.606060606060606e-05,
      "loss": 0.0098,
      "step": 8440
    },
    {
      "epoch": 127.95454545454545,
      "grad_norm": 0.05122085288167,
      "learning_rate": 3.602272727272727e-05,
      "loss": 0.0088,
      "step": 8445
    },
    {
      "epoch": 128.0,
      "eval_loss": 0.01208034809678793,
      "eval_mean_accuracy": 0.9412263422602472,
      "eval_mean_iou": 0.9179101556222182,
      "eval_runtime": 35.1208,
      "eval_samples_per_second": 6.691,
      "eval_steps_per_second": 0.854,
      "step": 8448
    },
    {
      "epoch": 128.03030303030303,
      "grad_norm": 0.09724091738462448,
      "learning_rate": 3.598484848484849e-05,
      "loss": 0.0089,
      "step": 8450
    },
    {
      "epoch": 128.1060606060606,
      "grad_norm": 0.044216983020305634,
      "learning_rate": 3.59469696969697e-05,
      "loss": 0.0108,
      "step": 8455
    },
    {
      "epoch": 128.1818181818182,
      "grad_norm": 0.1320667713880539,
      "learning_rate": 3.590909090909091e-05,
      "loss": 0.0088,
      "step": 8460
    },
    {
      "epoch": 128.25757575757575,
      "grad_norm": 0.03729567676782608,
      "learning_rate": 3.5871212121212125e-05,
      "loss": 0.0091,
      "step": 8465
    },
    {
      "epoch": 128.33333333333334,
      "grad_norm": 0.0361250676214695,
      "learning_rate": 3.5833333333333335e-05,
      "loss": 0.009,
      "step": 8470
    },
    {
      "epoch": 128.4090909090909,
      "grad_norm": 0.02596472203731537,
      "learning_rate": 3.579545454545455e-05,
      "loss": 0.0122,
      "step": 8475
    },
    {
      "epoch": 128.4848484848485,
      "grad_norm": 0.09995855391025543,
      "learning_rate": 3.575757575757576e-05,
      "loss": 0.0121,
      "step": 8480
    },
    {
      "epoch": 128.56060606060606,
      "grad_norm": 0.04827670380473137,
      "learning_rate": 3.571969696969697e-05,
      "loss": 0.0105,
      "step": 8485
    },
    {
      "epoch": 128.63636363636363,
      "grad_norm": 0.03461725264787674,
      "learning_rate": 3.568181818181818e-05,
      "loss": 0.0083,
      "step": 8490
    },
    {
      "epoch": 128.71212121212122,
      "grad_norm": 0.03628242015838623,
      "learning_rate": 3.56439393939394e-05,
      "loss": 0.0079,
      "step": 8495
    },
    {
      "epoch": 128.78787878787878,
      "grad_norm": 0.04277709871530533,
      "learning_rate": 3.560606060606061e-05,
      "loss": 0.0096,
      "step": 8500
    },
    {
      "epoch": 128.86363636363637,
      "grad_norm": 0.07530226558446884,
      "learning_rate": 3.5568181818181817e-05,
      "loss": 0.0094,
      "step": 8505
    },
    {
      "epoch": 128.93939393939394,
      "grad_norm": 0.04459000751376152,
      "learning_rate": 3.553030303030303e-05,
      "loss": 0.0093,
      "step": 8510
    },
    {
      "epoch": 129.0,
      "eval_loss": 0.011603708378970623,
      "eval_mean_accuracy": 0.9490118024541309,
      "eval_mean_iou": 0.9238850890221331,
      "eval_runtime": 34.9804,
      "eval_samples_per_second": 6.718,
      "eval_steps_per_second": 0.858,
      "step": 8514
    },
    {
      "epoch": 129.0151515151515,
      "grad_norm": 0.04359501227736473,
      "learning_rate": 3.549242424242424e-05,
      "loss": 0.0131,
      "step": 8515
    },
    {
      "epoch": 129.0909090909091,
      "grad_norm": 0.02928740344941616,
      "learning_rate": 3.545454545454546e-05,
      "loss": 0.009,
      "step": 8520
    },
    {
      "epoch": 129.16666666666666,
      "grad_norm": 0.06293297559022903,
      "learning_rate": 3.541666666666667e-05,
      "loss": 0.0097,
      "step": 8525
    },
    {
      "epoch": 129.24242424242425,
      "grad_norm": 0.07199506461620331,
      "learning_rate": 3.537878787878788e-05,
      "loss": 0.0092,
      "step": 8530
    },
    {
      "epoch": 129.3181818181818,
      "grad_norm": 0.07345488667488098,
      "learning_rate": 3.534090909090909e-05,
      "loss": 0.0121,
      "step": 8535
    },
    {
      "epoch": 129.3939393939394,
      "grad_norm": 0.04610205814242363,
      "learning_rate": 3.5303030303030305e-05,
      "loss": 0.0082,
      "step": 8540
    },
    {
      "epoch": 129.46969696969697,
      "grad_norm": 0.04513859748840332,
      "learning_rate": 3.5265151515151515e-05,
      "loss": 0.0063,
      "step": 8545
    },
    {
      "epoch": 129.54545454545453,
      "grad_norm": 0.03400440514087677,
      "learning_rate": 3.522727272727273e-05,
      "loss": 0.0076,
      "step": 8550
    },
    {
      "epoch": 129.62121212121212,
      "grad_norm": 0.03500162810087204,
      "learning_rate": 3.518939393939394e-05,
      "loss": 0.0068,
      "step": 8555
    },
    {
      "epoch": 129.6969696969697,
      "grad_norm": 0.03735074773430824,
      "learning_rate": 3.515151515151515e-05,
      "loss": 0.0081,
      "step": 8560
    },
    {
      "epoch": 129.77272727272728,
      "grad_norm": 0.06178868189454079,
      "learning_rate": 3.511363636363637e-05,
      "loss": 0.0117,
      "step": 8565
    },
    {
      "epoch": 129.84848484848484,
      "grad_norm": 0.0311241764575243,
      "learning_rate": 3.507575757575758e-05,
      "loss": 0.01,
      "step": 8570
    },
    {
      "epoch": 129.92424242424244,
      "grad_norm": 0.08507522195577621,
      "learning_rate": 3.5037878787878794e-05,
      "loss": 0.0145,
      "step": 8575
    },
    {
      "epoch": 130.0,
      "grad_norm": 0.05748141556978226,
      "learning_rate": 3.5e-05,
      "loss": 0.0097,
      "step": 8580
    },
    {
      "epoch": 130.0,
      "eval_loss": 0.011471419595181942,
      "eval_mean_accuracy": 0.9511098017517172,
      "eval_mean_iou": 0.9245490415572023,
      "eval_runtime": 37.5119,
      "eval_samples_per_second": 6.265,
      "eval_steps_per_second": 0.8,
      "step": 8580
    },
    {
      "epoch": 130.07575757575756,
      "grad_norm": 0.08225303143262863,
      "learning_rate": 3.4962121212121213e-05,
      "loss": 0.0116,
      "step": 8585
    },
    {
      "epoch": 130.15151515151516,
      "grad_norm": 0.05481057986617088,
      "learning_rate": 3.492424242424242e-05,
      "loss": 0.0107,
      "step": 8590
    },
    {
      "epoch": 130.22727272727272,
      "grad_norm": 0.05367141589522362,
      "learning_rate": 3.488636363636364e-05,
      "loss": 0.0108,
      "step": 8595
    },
    {
      "epoch": 130.3030303030303,
      "grad_norm": 0.0409327894449234,
      "learning_rate": 3.484848484848485e-05,
      "loss": 0.0113,
      "step": 8600
    },
    {
      "epoch": 130.37878787878788,
      "grad_norm": 0.061954010277986526,
      "learning_rate": 3.481060606060606e-05,
      "loss": 0.0105,
      "step": 8605
    },
    {
      "epoch": 130.45454545454547,
      "grad_norm": 0.10364287346601486,
      "learning_rate": 3.4772727272727276e-05,
      "loss": 0.0112,
      "step": 8610
    },
    {
      "epoch": 130.53030303030303,
      "grad_norm": 0.039972711354494095,
      "learning_rate": 3.4734848484848486e-05,
      "loss": 0.0081,
      "step": 8615
    },
    {
      "epoch": 130.6060606060606,
      "grad_norm": 0.11094771325588226,
      "learning_rate": 3.46969696969697e-05,
      "loss": 0.012,
      "step": 8620
    },
    {
      "epoch": 130.6818181818182,
      "grad_norm": 0.028931086882948875,
      "learning_rate": 3.465909090909091e-05,
      "loss": 0.0077,
      "step": 8625
    },
    {
      "epoch": 130.75757575757575,
      "grad_norm": 0.04306592419743538,
      "learning_rate": 3.462121212121212e-05,
      "loss": 0.0073,
      "step": 8630
    },
    {
      "epoch": 130.83333333333334,
      "grad_norm": 0.030020417645573616,
      "learning_rate": 3.458333333333333e-05,
      "loss": 0.0111,
      "step": 8635
    },
    {
      "epoch": 130.9090909090909,
      "grad_norm": 0.049590595066547394,
      "learning_rate": 3.454545454545455e-05,
      "loss": 0.0094,
      "step": 8640
    },
    {
      "epoch": 130.9848484848485,
      "grad_norm": 0.026402123272418976,
      "learning_rate": 3.450757575757576e-05,
      "loss": 0.0073,
      "step": 8645
    },
    {
      "epoch": 131.0,
      "eval_loss": 0.01160680502653122,
      "eval_mean_accuracy": 0.9485720717477798,
      "eval_mean_iou": 0.9237098617817533,
      "eval_runtime": 34.3222,
      "eval_samples_per_second": 6.847,
      "eval_steps_per_second": 0.874,
      "step": 8646
    },
    {
      "epoch": 131.06060606060606,
      "grad_norm": 0.01677548512816429,
      "learning_rate": 3.4469696969696974e-05,
      "loss": 0.0101,
      "step": 8650
    },
    {
      "epoch": 131.13636363636363,
      "grad_norm": 0.030371837317943573,
      "learning_rate": 3.4431818181818184e-05,
      "loss": 0.0082,
      "step": 8655
    },
    {
      "epoch": 131.21212121212122,
      "grad_norm": 0.032993610948324203,
      "learning_rate": 3.4393939393939394e-05,
      "loss": 0.0103,
      "step": 8660
    },
    {
      "epoch": 131.28787878787878,
      "grad_norm": 0.024894731119275093,
      "learning_rate": 3.435606060606061e-05,
      "loss": 0.0103,
      "step": 8665
    },
    {
      "epoch": 131.36363636363637,
      "grad_norm": 0.13170059025287628,
      "learning_rate": 3.431818181818182e-05,
      "loss": 0.0123,
      "step": 8670
    },
    {
      "epoch": 131.43939393939394,
      "grad_norm": 0.11395426094532013,
      "learning_rate": 3.428030303030303e-05,
      "loss": 0.009,
      "step": 8675
    },
    {
      "epoch": 131.5151515151515,
      "grad_norm": 0.06485287845134735,
      "learning_rate": 3.424242424242424e-05,
      "loss": 0.0087,
      "step": 8680
    },
    {
      "epoch": 131.5909090909091,
      "grad_norm": 0.08698050677776337,
      "learning_rate": 3.4204545454545456e-05,
      "loss": 0.0102,
      "step": 8685
    },
    {
      "epoch": 131.66666666666666,
      "grad_norm": 0.02075636200606823,
      "learning_rate": 3.4166666666666666e-05,
      "loss": 0.0087,
      "step": 8690
    },
    {
      "epoch": 131.74242424242425,
      "grad_norm": 0.09352346509695053,
      "learning_rate": 3.412878787878788e-05,
      "loss": 0.0106,
      "step": 8695
    },
    {
      "epoch": 131.8181818181818,
      "grad_norm": 0.037382129579782486,
      "learning_rate": 3.409090909090909e-05,
      "loss": 0.0096,
      "step": 8700
    },
    {
      "epoch": 131.8939393939394,
      "grad_norm": 0.07882483303546906,
      "learning_rate": 3.40530303030303e-05,
      "loss": 0.0089,
      "step": 8705
    },
    {
      "epoch": 131.96969696969697,
      "grad_norm": 0.11329109966754913,
      "learning_rate": 3.401515151515152e-05,
      "loss": 0.0099,
      "step": 8710
    },
    {
      "epoch": 132.0,
      "eval_loss": 0.011413846164941788,
      "eval_mean_accuracy": 0.9601387010043393,
      "eval_mean_iou": 0.9271655665314481,
      "eval_runtime": 35.2717,
      "eval_samples_per_second": 6.663,
      "eval_steps_per_second": 0.851,
      "step": 8712
    },
    {
      "epoch": 132.04545454545453,
      "grad_norm": 0.034428004175424576,
      "learning_rate": 3.397727272727273e-05,
      "loss": 0.0104,
      "step": 8715
    },
    {
      "epoch": 132.12121212121212,
      "grad_norm": 0.023165065795183182,
      "learning_rate": 3.3939393939393945e-05,
      "loss": 0.0086,
      "step": 8720
    },
    {
      "epoch": 132.1969696969697,
      "grad_norm": 0.045800983905792236,
      "learning_rate": 3.390151515151515e-05,
      "loss": 0.0092,
      "step": 8725
    },
    {
      "epoch": 132.27272727272728,
      "grad_norm": 0.03724659979343414,
      "learning_rate": 3.3863636363636364e-05,
      "loss": 0.0085,
      "step": 8730
    },
    {
      "epoch": 132.34848484848484,
      "grad_norm": 0.06768983602523804,
      "learning_rate": 3.3825757575757574e-05,
      "loss": 0.0116,
      "step": 8735
    },
    {
      "epoch": 132.42424242424244,
      "grad_norm": 0.029391763731837273,
      "learning_rate": 3.378787878787879e-05,
      "loss": 0.01,
      "step": 8740
    },
    {
      "epoch": 132.5,
      "grad_norm": 0.07768172025680542,
      "learning_rate": 3.375000000000001e-05,
      "loss": 0.0108,
      "step": 8745
    },
    {
      "epoch": 132.57575757575756,
      "grad_norm": 0.04634859040379524,
      "learning_rate": 3.371212121212121e-05,
      "loss": 0.0092,
      "step": 8750
    },
    {
      "epoch": 132.65151515151516,
      "grad_norm": 0.09306085854768753,
      "learning_rate": 3.367424242424243e-05,
      "loss": 0.0083,
      "step": 8755
    },
    {
      "epoch": 132.72727272727272,
      "grad_norm": 0.03383093327283859,
      "learning_rate": 3.3636363636363636e-05,
      "loss": 0.0076,
      "step": 8760
    },
    {
      "epoch": 132.8030303030303,
      "grad_norm": 0.029636047780513763,
      "learning_rate": 3.359848484848485e-05,
      "loss": 0.0091,
      "step": 8765
    },
    {
      "epoch": 132.87878787878788,
      "grad_norm": 0.035899270325899124,
      "learning_rate": 3.356060606060606e-05,
      "loss": 0.0105,
      "step": 8770
    },
    {
      "epoch": 132.95454545454547,
      "grad_norm": 0.03174450248479843,
      "learning_rate": 3.352272727272727e-05,
      "loss": 0.0105,
      "step": 8775
    },
    {
      "epoch": 133.0,
      "eval_loss": 0.0115585345774889,
      "eval_mean_accuracy": 0.9509786151644244,
      "eval_mean_iou": 0.9253765435279201,
      "eval_runtime": 34.6909,
      "eval_samples_per_second": 6.774,
      "eval_steps_per_second": 0.865,
      "step": 8778
    },
    {
      "epoch": 133.03030303030303,
      "grad_norm": 0.020252712070941925,
      "learning_rate": 3.348484848484848e-05,
      "loss": 0.0115,
      "step": 8780
    },
    {
      "epoch": 133.1060606060606,
      "grad_norm": 0.04375099390745163,
      "learning_rate": 3.34469696969697e-05,
      "loss": 0.0083,
      "step": 8785
    },
    {
      "epoch": 133.1818181818182,
      "grad_norm": 0.10234501212835312,
      "learning_rate": 3.3409090909090915e-05,
      "loss": 0.0115,
      "step": 8790
    },
    {
      "epoch": 133.25757575757575,
      "grad_norm": 0.03203010931611061,
      "learning_rate": 3.3371212121212125e-05,
      "loss": 0.0091,
      "step": 8795
    },
    {
      "epoch": 133.33333333333334,
      "grad_norm": 0.06114264577627182,
      "learning_rate": 3.3333333333333335e-05,
      "loss": 0.0087,
      "step": 8800
    },
    {
      "epoch": 133.4090909090909,
      "grad_norm": 0.034767214208841324,
      "learning_rate": 3.3295454545454545e-05,
      "loss": 0.0107,
      "step": 8805
    },
    {
      "epoch": 133.4848484848485,
      "grad_norm": 0.0632830485701561,
      "learning_rate": 3.325757575757576e-05,
      "loss": 0.0099,
      "step": 8810
    },
    {
      "epoch": 133.56060606060606,
      "grad_norm": 0.04133511334657669,
      "learning_rate": 3.321969696969697e-05,
      "loss": 0.0098,
      "step": 8815
    },
    {
      "epoch": 133.63636363636363,
      "grad_norm": 0.022309722378849983,
      "learning_rate": 3.318181818181819e-05,
      "loss": 0.0078,
      "step": 8820
    },
    {
      "epoch": 133.71212121212122,
      "grad_norm": 0.07877352833747864,
      "learning_rate": 3.314393939393939e-05,
      "loss": 0.0104,
      "step": 8825
    },
    {
      "epoch": 133.78787878787878,
      "grad_norm": 0.05790197104215622,
      "learning_rate": 3.310606060606061e-05,
      "loss": 0.0127,
      "step": 8830
    },
    {
      "epoch": 133.86363636363637,
      "grad_norm": 0.06557337194681168,
      "learning_rate": 3.3068181818181824e-05,
      "loss": 0.0097,
      "step": 8835
    },
    {
      "epoch": 133.93939393939394,
      "grad_norm": 0.04801876097917557,
      "learning_rate": 3.303030303030303e-05,
      "loss": 0.009,
      "step": 8840
    },
    {
      "epoch": 134.0,
      "eval_loss": 0.011321503669023514,
      "eval_mean_accuracy": 0.9559894810266513,
      "eval_mean_iou": 0.9272887505983404,
      "eval_runtime": 34.3295,
      "eval_samples_per_second": 6.845,
      "eval_steps_per_second": 0.874,
      "step": 8844
    },
    {
      "epoch": 134.0151515151515,
      "grad_norm": 0.03230193257331848,
      "learning_rate": 3.299242424242424e-05,
      "loss": 0.0086,
      "step": 8845
    },
    {
      "epoch": 134.0909090909091,
      "grad_norm": 0.04430897906422615,
      "learning_rate": 3.295454545454545e-05,
      "loss": 0.0082,
      "step": 8850
    },
    {
      "epoch": 134.16666666666666,
      "grad_norm": 0.055584728717803955,
      "learning_rate": 3.291666666666667e-05,
      "loss": 0.0104,
      "step": 8855
    },
    {
      "epoch": 134.24242424242425,
      "grad_norm": 0.0530477836728096,
      "learning_rate": 3.287878787878788e-05,
      "loss": 0.0091,
      "step": 8860
    },
    {
      "epoch": 134.3181818181818,
      "grad_norm": 0.02086392603814602,
      "learning_rate": 3.2840909090909096e-05,
      "loss": 0.0071,
      "step": 8865
    },
    {
      "epoch": 134.3939393939394,
      "grad_norm": 0.049056071788072586,
      "learning_rate": 3.2803030303030305e-05,
      "loss": 0.0093,
      "step": 8870
    },
    {
      "epoch": 134.46969696969697,
      "grad_norm": 0.044120270758867264,
      "learning_rate": 3.2765151515151515e-05,
      "loss": 0.0077,
      "step": 8875
    },
    {
      "epoch": 134.54545454545453,
      "grad_norm": 0.05047740414738655,
      "learning_rate": 3.272727272727273e-05,
      "loss": 0.0108,
      "step": 8880
    },
    {
      "epoch": 134.62121212121212,
      "grad_norm": 0.04228570684790611,
      "learning_rate": 3.268939393939394e-05,
      "loss": 0.0113,
      "step": 8885
    },
    {
      "epoch": 134.6969696969697,
      "grad_norm": 0.05611094459891319,
      "learning_rate": 3.265151515151516e-05,
      "loss": 0.0088,
      "step": 8890
    },
    {
      "epoch": 134.77272727272728,
      "grad_norm": 0.03632834926247597,
      "learning_rate": 3.261363636363636e-05,
      "loss": 0.0082,
      "step": 8895
    },
    {
      "epoch": 134.84848484848484,
      "grad_norm": 0.032114434987306595,
      "learning_rate": 3.257575757575758e-05,
      "loss": 0.0109,
      "step": 8900
    },
    {
      "epoch": 134.92424242424244,
      "grad_norm": 0.05100783333182335,
      "learning_rate": 3.253787878787879e-05,
      "loss": 0.0138,
      "step": 8905
    },
    {
      "epoch": 135.0,
      "grad_norm": 0.15643560886383057,
      "learning_rate": 3.2500000000000004e-05,
      "loss": 0.0152,
      "step": 8910
    },
    {
      "epoch": 135.0,
      "eval_loss": 0.01141317654401064,
      "eval_mean_accuracy": 0.9558845456943931,
      "eval_mean_iou": 0.9270431823948725,
      "eval_runtime": 35.9872,
      "eval_samples_per_second": 6.53,
      "eval_steps_per_second": 0.834,
      "step": 8910
    },
    {
      "epoch": 135.07575757575756,
      "grad_norm": 0.03925289213657379,
      "learning_rate": 3.2462121212121214e-05,
      "loss": 0.0117,
      "step": 8915
    },
    {
      "epoch": 135.15151515151516,
      "grad_norm": 0.09307269006967545,
      "learning_rate": 3.2424242424242423e-05,
      "loss": 0.0118,
      "step": 8920
    },
    {
      "epoch": 135.22727272727272,
      "grad_norm": 0.030813489109277725,
      "learning_rate": 3.238636363636364e-05,
      "loss": 0.0088,
      "step": 8925
    },
    {
      "epoch": 135.3030303030303,
      "grad_norm": 0.02784663625061512,
      "learning_rate": 3.234848484848485e-05,
      "loss": 0.0073,
      "step": 8930
    },
    {
      "epoch": 135.37878787878788,
      "grad_norm": 0.032904066145420074,
      "learning_rate": 3.2310606060606066e-05,
      "loss": 0.0072,
      "step": 8935
    },
    {
      "epoch": 135.45454545454547,
      "grad_norm": 0.03006020560860634,
      "learning_rate": 3.2272727272727276e-05,
      "loss": 0.0101,
      "step": 8940
    },
    {
      "epoch": 135.53030303030303,
      "grad_norm": 0.029239676892757416,
      "learning_rate": 3.2234848484848486e-05,
      "loss": 0.0106,
      "step": 8945
    },
    {
      "epoch": 135.6060606060606,
      "grad_norm": 0.024123111739754677,
      "learning_rate": 3.2196969696969696e-05,
      "loss": 0.0093,
      "step": 8950
    },
    {
      "epoch": 135.6818181818182,
      "grad_norm": 0.022710217162966728,
      "learning_rate": 3.215909090909091e-05,
      "loss": 0.0097,
      "step": 8955
    },
    {
      "epoch": 135.75757575757575,
      "grad_norm": 0.024024497717618942,
      "learning_rate": 3.212121212121212e-05,
      "loss": 0.0105,
      "step": 8960
    },
    {
      "epoch": 135.83333333333334,
      "grad_norm": 0.043535154312849045,
      "learning_rate": 3.208333333333334e-05,
      "loss": 0.0091,
      "step": 8965
    },
    {
      "epoch": 135.9090909090909,
      "grad_norm": 0.061916157603263855,
      "learning_rate": 3.204545454545455e-05,
      "loss": 0.009,
      "step": 8970
    },
    {
      "epoch": 135.9848484848485,
      "grad_norm": 0.07204847782850266,
      "learning_rate": 3.200757575757576e-05,
      "loss": 0.0083,
      "step": 8975
    },
    {
      "epoch": 136.0,
      "eval_loss": 0.011458001099526882,
      "eval_mean_accuracy": 0.9489084392520933,
      "eval_mean_iou": 0.924620993781628,
      "eval_runtime": 35.5413,
      "eval_samples_per_second": 6.612,
      "eval_steps_per_second": 0.844,
      "step": 8976
    },
    {
      "epoch": 136.06060606060606,
      "grad_norm": 0.09752221405506134,
      "learning_rate": 3.1969696969696974e-05,
      "loss": 0.0102,
      "step": 8980
    },
    {
      "epoch": 136.13636363636363,
      "grad_norm": 0.04896555468440056,
      "learning_rate": 3.1931818181818184e-05,
      "loss": 0.0101,
      "step": 8985
    },
    {
      "epoch": 136.21212121212122,
      "grad_norm": 0.07691677659749985,
      "learning_rate": 3.18939393939394e-05,
      "loss": 0.0121,
      "step": 8990
    },
    {
      "epoch": 136.28787878787878,
      "grad_norm": 0.03542190417647362,
      "learning_rate": 3.1856060606060604e-05,
      "loss": 0.0091,
      "step": 8995
    },
    {
      "epoch": 136.36363636363637,
      "grad_norm": 0.038343992084264755,
      "learning_rate": 3.181818181818182e-05,
      "loss": 0.0077,
      "step": 9000
    },
    {
      "epoch": 136.43939393939394,
      "grad_norm": 0.07743646204471588,
      "learning_rate": 3.178030303030303e-05,
      "loss": 0.0084,
      "step": 9005
    },
    {
      "epoch": 136.5151515151515,
      "grad_norm": 0.021942174062132835,
      "learning_rate": 3.174242424242425e-05,
      "loss": 0.0126,
      "step": 9010
    },
    {
      "epoch": 136.5909090909091,
      "grad_norm": 0.07081823796033859,
      "learning_rate": 3.1704545454545456e-05,
      "loss": 0.0076,
      "step": 9015
    },
    {
      "epoch": 136.66666666666666,
      "grad_norm": 0.029726307839155197,
      "learning_rate": 3.1666666666666666e-05,
      "loss": 0.0083,
      "step": 9020
    },
    {
      "epoch": 136.74242424242425,
      "grad_norm": 0.052832573652267456,
      "learning_rate": 3.162878787878788e-05,
      "loss": 0.0114,
      "step": 9025
    },
    {
      "epoch": 136.8181818181818,
      "grad_norm": 0.029708823189139366,
      "learning_rate": 3.159090909090909e-05,
      "loss": 0.0084,
      "step": 9030
    },
    {
      "epoch": 136.8939393939394,
      "grad_norm": 0.07680356502532959,
      "learning_rate": 3.155303030303031e-05,
      "loss": 0.0098,
      "step": 9035
    },
    {
      "epoch": 136.96969696969697,
      "grad_norm": 0.030426310375332832,
      "learning_rate": 3.151515151515151e-05,
      "loss": 0.0083,
      "step": 9040
    },
    {
      "epoch": 137.0,
      "eval_loss": 0.0112299844622612,
      "eval_mean_accuracy": 0.9603647881408847,
      "eval_mean_iou": 0.9287105977320885,
      "eval_runtime": 35.3819,
      "eval_samples_per_second": 6.642,
      "eval_steps_per_second": 0.848,
      "step": 9042
    },
    {
      "epoch": 137.04545454545453,
      "grad_norm": 0.04114415869116783,
      "learning_rate": 3.147727272727273e-05,
      "loss": 0.0092,
      "step": 9045
    },
    {
      "epoch": 137.12121212121212,
      "grad_norm": 0.015342568047344685,
      "learning_rate": 3.143939393939394e-05,
      "loss": 0.0074,
      "step": 9050
    },
    {
      "epoch": 137.1969696969697,
      "grad_norm": 0.030284738168120384,
      "learning_rate": 3.1401515151515155e-05,
      "loss": 0.0088,
      "step": 9055
    },
    {
      "epoch": 137.27272727272728,
      "grad_norm": 0.02703307382762432,
      "learning_rate": 3.1363636363636365e-05,
      "loss": 0.0112,
      "step": 9060
    },
    {
      "epoch": 137.34848484848484,
      "grad_norm": 0.054635338485240936,
      "learning_rate": 3.1325757575757574e-05,
      "loss": 0.0101,
      "step": 9065
    },
    {
      "epoch": 137.42424242424244,
      "grad_norm": 0.04962631314992905,
      "learning_rate": 3.128787878787879e-05,
      "loss": 0.0093,
      "step": 9070
    },
    {
      "epoch": 137.5,
      "grad_norm": 0.04999437555670738,
      "learning_rate": 3.125e-05,
      "loss": 0.0097,
      "step": 9075
    },
    {
      "epoch": 137.57575757575756,
      "grad_norm": 0.023156099021434784,
      "learning_rate": 3.121212121212122e-05,
      "loss": 0.0098,
      "step": 9080
    },
    {
      "epoch": 137.65151515151516,
      "grad_norm": 0.029950710013508797,
      "learning_rate": 3.117424242424243e-05,
      "loss": 0.0082,
      "step": 9085
    },
    {
      "epoch": 137.72727272727272,
      "grad_norm": 0.0806482806801796,
      "learning_rate": 3.113636363636364e-05,
      "loss": 0.0104,
      "step": 9090
    },
    {
      "epoch": 137.8030303030303,
      "grad_norm": 0.03174377605319023,
      "learning_rate": 3.1098484848484846e-05,
      "loss": 0.0094,
      "step": 9095
    },
    {
      "epoch": 137.87878787878788,
      "grad_norm": 0.017493611201643944,
      "learning_rate": 3.106060606060606e-05,
      "loss": 0.0087,
      "step": 9100
    },
    {
      "epoch": 137.95454545454547,
      "grad_norm": 0.07741950452327728,
      "learning_rate": 3.102272727272727e-05,
      "loss": 0.009,
      "step": 9105
    },
    {
      "epoch": 138.0,
      "eval_loss": 0.011622000485658646,
      "eval_mean_accuracy": 0.9490641832208571,
      "eval_mean_iou": 0.9239742911588195,
      "eval_runtime": 34.8464,
      "eval_samples_per_second": 6.744,
      "eval_steps_per_second": 0.861,
      "step": 9108
    },
    {
      "epoch": 138.03030303030303,
      "grad_norm": 0.05129273608326912,
      "learning_rate": 3.098484848484849e-05,
      "loss": 0.0082,
      "step": 9110
    },
    {
      "epoch": 138.1060606060606,
      "grad_norm": 0.07347065210342407,
      "learning_rate": 3.09469696969697e-05,
      "loss": 0.0087,
      "step": 9115
    },
    {
      "epoch": 138.1818181818182,
      "grad_norm": 0.037973109632730484,
      "learning_rate": 3.090909090909091e-05,
      "loss": 0.0109,
      "step": 9120
    },
    {
      "epoch": 138.25757575757575,
      "grad_norm": 0.041690196841955185,
      "learning_rate": 3.0871212121212125e-05,
      "loss": 0.0098,
      "step": 9125
    },
    {
      "epoch": 138.33333333333334,
      "grad_norm": 0.04626249521970749,
      "learning_rate": 3.0833333333333335e-05,
      "loss": 0.0127,
      "step": 9130
    },
    {
      "epoch": 138.4090909090909,
      "grad_norm": 0.029777755960822105,
      "learning_rate": 3.079545454545455e-05,
      "loss": 0.009,
      "step": 9135
    },
    {
      "epoch": 138.4848484848485,
      "grad_norm": 0.0491754375398159,
      "learning_rate": 3.0757575757575755e-05,
      "loss": 0.0077,
      "step": 9140
    },
    {
      "epoch": 138.56060606060606,
      "grad_norm": 0.0240919329226017,
      "learning_rate": 3.071969696969697e-05,
      "loss": 0.0089,
      "step": 9145
    },
    {
      "epoch": 138.63636363636363,
      "grad_norm": 0.07376453280448914,
      "learning_rate": 3.068181818181818e-05,
      "loss": 0.0107,
      "step": 9150
    },
    {
      "epoch": 138.71212121212122,
      "grad_norm": 0.03935198113322258,
      "learning_rate": 3.06439393939394e-05,
      "loss": 0.0076,
      "step": 9155
    },
    {
      "epoch": 138.78787878787878,
      "grad_norm": 0.019614601507782936,
      "learning_rate": 3.060606060606061e-05,
      "loss": 0.0079,
      "step": 9160
    },
    {
      "epoch": 138.86363636363637,
      "grad_norm": 0.035364095121622086,
      "learning_rate": 3.056818181818182e-05,
      "loss": 0.009,
      "step": 9165
    },
    {
      "epoch": 138.93939393939394,
      "grad_norm": 0.024053087458014488,
      "learning_rate": 3.0530303030303034e-05,
      "loss": 0.0111,
      "step": 9170
    },
    {
      "epoch": 139.0,
      "eval_loss": 0.01123303733766079,
      "eval_mean_accuracy": 0.956451915357795,
      "eval_mean_iou": 0.9278483557434325,
      "eval_runtime": 42.7628,
      "eval_samples_per_second": 5.495,
      "eval_steps_per_second": 0.702,
      "step": 9174
    },
    {
      "epoch": 139.0151515151515,
      "grad_norm": 0.02032369188964367,
      "learning_rate": 3.0492424242424243e-05,
      "loss": 0.008,
      "step": 9175
    },
    {
      "epoch": 139.0909090909091,
      "grad_norm": 0.019847478717565536,
      "learning_rate": 3.0454545454545456e-05,
      "loss": 0.0081,
      "step": 9180
    },
    {
      "epoch": 139.16666666666666,
      "grad_norm": 0.04704216495156288,
      "learning_rate": 3.0416666666666666e-05,
      "loss": 0.0118,
      "step": 9185
    },
    {
      "epoch": 139.24242424242425,
      "grad_norm": 0.05798988416790962,
      "learning_rate": 3.037878787878788e-05,
      "loss": 0.0086,
      "step": 9190
    },
    {
      "epoch": 139.3181818181818,
      "grad_norm": 0.036973986774683,
      "learning_rate": 3.0340909090909093e-05,
      "loss": 0.0083,
      "step": 9195
    },
    {
      "epoch": 139.3939393939394,
      "grad_norm": 0.09129863232374191,
      "learning_rate": 3.0303030303030306e-05,
      "loss": 0.0103,
      "step": 9200
    },
    {
      "epoch": 139.46969696969697,
      "grad_norm": 0.05164320394396782,
      "learning_rate": 3.026515151515152e-05,
      "loss": 0.0099,
      "step": 9205
    },
    {
      "epoch": 139.54545454545453,
      "grad_norm": 0.05359017103910446,
      "learning_rate": 3.0227272727272725e-05,
      "loss": 0.0081,
      "step": 9210
    },
    {
      "epoch": 139.62121212121212,
      "grad_norm": 0.07858824729919434,
      "learning_rate": 3.018939393939394e-05,
      "loss": 0.0106,
      "step": 9215
    },
    {
      "epoch": 139.6969696969697,
      "grad_norm": 0.03882090747356415,
      "learning_rate": 3.015151515151515e-05,
      "loss": 0.0092,
      "step": 9220
    },
    {
      "epoch": 139.77272727272728,
      "grad_norm": 0.04158050939440727,
      "learning_rate": 3.0113636363636365e-05,
      "loss": 0.0092,
      "step": 9225
    },
    {
      "epoch": 139.84848484848484,
      "grad_norm": 0.034930627793073654,
      "learning_rate": 3.0075757575757578e-05,
      "loss": 0.0082,
      "step": 9230
    },
    {
      "epoch": 139.92424242424244,
      "grad_norm": 0.03474419564008713,
      "learning_rate": 3.0037878787878788e-05,
      "loss": 0.0097,
      "step": 9235
    },
    {
      "epoch": 140.0,
      "grad_norm": 0.04605212062597275,
      "learning_rate": 3e-05,
      "loss": 0.0099,
      "step": 9240
    },
    {
      "epoch": 140.0,
      "eval_loss": 0.011149872094392776,
      "eval_mean_accuracy": 0.9580674016617928,
      "eval_mean_iou": 0.9284994993680876,
      "eval_runtime": 34.7152,
      "eval_samples_per_second": 6.769,
      "eval_steps_per_second": 0.864,
      "step": 9240
    },
    {
      "epoch": 140.07575757575756,
      "grad_norm": 0.021637775003910065,
      "learning_rate": 2.9962121212121214e-05,
      "loss": 0.0096,
      "step": 9245
    },
    {
      "epoch": 140.15151515151516,
      "grad_norm": 0.039745599031448364,
      "learning_rate": 2.9924242424242427e-05,
      "loss": 0.0098,
      "step": 9250
    },
    {
      "epoch": 140.22727272727272,
      "grad_norm": 0.05335872247815132,
      "learning_rate": 2.988636363636364e-05,
      "loss": 0.0083,
      "step": 9255
    },
    {
      "epoch": 140.3030303030303,
      "grad_norm": 0.04503278061747551,
      "learning_rate": 2.9848484848484847e-05,
      "loss": 0.011,
      "step": 9260
    },
    {
      "epoch": 140.37878787878788,
      "grad_norm": 0.033141400665044785,
      "learning_rate": 2.981060606060606e-05,
      "loss": 0.0087,
      "step": 9265
    },
    {
      "epoch": 140.45454545454547,
      "grad_norm": 0.04414450749754906,
      "learning_rate": 2.9772727272727273e-05,
      "loss": 0.0098,
      "step": 9270
    },
    {
      "epoch": 140.53030303030303,
      "grad_norm": 0.05212988704442978,
      "learning_rate": 2.9734848484848486e-05,
      "loss": 0.0064,
      "step": 9275
    },
    {
      "epoch": 140.6060606060606,
      "grad_norm": 0.06106264144182205,
      "learning_rate": 2.96969696969697e-05,
      "loss": 0.0069,
      "step": 9280
    },
    {
      "epoch": 140.6818181818182,
      "grad_norm": 0.038654688745737076,
      "learning_rate": 2.965909090909091e-05,
      "loss": 0.0092,
      "step": 9285
    },
    {
      "epoch": 140.75757575757575,
      "grad_norm": 0.02784941904246807,
      "learning_rate": 2.9621212121212122e-05,
      "loss": 0.0086,
      "step": 9290
    },
    {
      "epoch": 140.83333333333334,
      "grad_norm": 0.06511247158050537,
      "learning_rate": 2.9583333333333335e-05,
      "loss": 0.009,
      "step": 9295
    },
    {
      "epoch": 140.9090909090909,
      "grad_norm": 0.04713728651404381,
      "learning_rate": 2.954545454545455e-05,
      "loss": 0.0092,
      "step": 9300
    },
    {
      "epoch": 140.9848484848485,
      "grad_norm": 0.029808275401592255,
      "learning_rate": 2.950757575757576e-05,
      "loss": 0.0112,
      "step": 9305
    },
    {
      "epoch": 141.0,
      "eval_loss": 0.01148268859833479,
      "eval_mean_accuracy": 0.954552537532519,
      "eval_mean_iou": 0.9271898936389262,
      "eval_runtime": 35.1311,
      "eval_samples_per_second": 6.689,
      "eval_steps_per_second": 0.854,
      "step": 9306
    },
    {
      "epoch": 141.06060606060606,
      "grad_norm": 0.05896470695734024,
      "learning_rate": 2.9469696969696968e-05,
      "loss": 0.009,
      "step": 9310
    },
    {
      "epoch": 141.13636363636363,
      "grad_norm": 0.0359964519739151,
      "learning_rate": 2.943181818181818e-05,
      "loss": 0.0106,
      "step": 9315
    },
    {
      "epoch": 141.21212121212122,
      "grad_norm": 0.06325055658817291,
      "learning_rate": 2.9393939393939394e-05,
      "loss": 0.0117,
      "step": 9320
    },
    {
      "epoch": 141.28787878787878,
      "grad_norm": 0.039637427777051926,
      "learning_rate": 2.935606060606061e-05,
      "loss": 0.0112,
      "step": 9325
    },
    {
      "epoch": 141.36363636363637,
      "grad_norm": 0.048491694033145905,
      "learning_rate": 2.9318181818181817e-05,
      "loss": 0.0086,
      "step": 9330
    },
    {
      "epoch": 141.43939393939394,
      "grad_norm": 0.030787263065576553,
      "learning_rate": 2.928030303030303e-05,
      "loss": 0.0089,
      "step": 9335
    },
    {
      "epoch": 141.5151515151515,
      "grad_norm": 0.03217754513025284,
      "learning_rate": 2.9242424242424243e-05,
      "loss": 0.0112,
      "step": 9340
    },
    {
      "epoch": 141.5909090909091,
      "grad_norm": 0.030675780028104782,
      "learning_rate": 2.9204545454545457e-05,
      "loss": 0.0076,
      "step": 9345
    },
    {
      "epoch": 141.66666666666666,
      "grad_norm": 0.05709864944219589,
      "learning_rate": 2.916666666666667e-05,
      "loss": 0.0103,
      "step": 9350
    },
    {
      "epoch": 141.74242424242425,
      "grad_norm": 0.04584331810474396,
      "learning_rate": 2.9128787878787876e-05,
      "loss": 0.0087,
      "step": 9355
    },
    {
      "epoch": 141.8181818181818,
      "grad_norm": 0.04584889858961105,
      "learning_rate": 2.909090909090909e-05,
      "loss": 0.0086,
      "step": 9360
    },
    {
      "epoch": 141.8939393939394,
      "grad_norm": 0.033714793622493744,
      "learning_rate": 2.9053030303030306e-05,
      "loss": 0.0091,
      "step": 9365
    },
    {
      "epoch": 141.96969696969697,
      "grad_norm": 0.025855572894215584,
      "learning_rate": 2.901515151515152e-05,
      "loss": 0.0081,
      "step": 9370
    },
    {
      "epoch": 142.0,
      "eval_loss": 0.011119426228106022,
      "eval_mean_accuracy": 0.9607400295876299,
      "eval_mean_iou": 0.9293187884547999,
      "eval_runtime": 35.0053,
      "eval_samples_per_second": 6.713,
      "eval_steps_per_second": 0.857,
      "step": 9372
    },
    {
      "epoch": 142.04545454545453,
      "grad_norm": 0.0324908047914505,
      "learning_rate": 2.8977272727272732e-05,
      "loss": 0.0082,
      "step": 9375
    },
    {
      "epoch": 142.12121212121212,
      "grad_norm": 0.05079766735434532,
      "learning_rate": 2.893939393939394e-05,
      "loss": 0.0098,
      "step": 9380
    },
    {
      "epoch": 142.1969696969697,
      "grad_norm": 0.044437699019908905,
      "learning_rate": 2.890151515151515e-05,
      "loss": 0.0097,
      "step": 9385
    },
    {
      "epoch": 142.27272727272728,
      "grad_norm": 0.02495563216507435,
      "learning_rate": 2.8863636363636365e-05,
      "loss": 0.0083,
      "step": 9390
    },
    {
      "epoch": 142.34848484848484,
      "grad_norm": 0.04471038654446602,
      "learning_rate": 2.8825757575757578e-05,
      "loss": 0.0108,
      "step": 9395
    },
    {
      "epoch": 142.42424242424244,
      "grad_norm": 0.03874368220567703,
      "learning_rate": 2.878787878787879e-05,
      "loss": 0.0086,
      "step": 9400
    },
    {
      "epoch": 142.5,
      "grad_norm": 0.03519868105649948,
      "learning_rate": 2.8749999999999997e-05,
      "loss": 0.0089,
      "step": 9405
    },
    {
      "epoch": 142.57575757575756,
      "grad_norm": 0.0460781492292881,
      "learning_rate": 2.8712121212121214e-05,
      "loss": 0.0118,
      "step": 9410
    },
    {
      "epoch": 142.65151515151516,
      "grad_norm": 0.09168718755245209,
      "learning_rate": 2.8674242424242427e-05,
      "loss": 0.0081,
      "step": 9415
    },
    {
      "epoch": 142.72727272727272,
      "grad_norm": 0.03923609107732773,
      "learning_rate": 2.863636363636364e-05,
      "loss": 0.0108,
      "step": 9420
    },
    {
      "epoch": 142.8030303030303,
      "grad_norm": 0.036094486713409424,
      "learning_rate": 2.8598484848484853e-05,
      "loss": 0.0073,
      "step": 9425
    },
    {
      "epoch": 142.87878787878788,
      "grad_norm": 0.03948888182640076,
      "learning_rate": 2.856060606060606e-05,
      "loss": 0.0078,
      "step": 9430
    },
    {
      "epoch": 142.95454545454547,
      "grad_norm": 0.0396638922393322,
      "learning_rate": 2.8522727272727273e-05,
      "loss": 0.0101,
      "step": 9435
    },
    {
      "epoch": 143.0,
      "eval_loss": 0.011552204377949238,
      "eval_mean_accuracy": 0.9530592189555996,
      "eval_mean_iou": 0.9263534866024284,
      "eval_runtime": 34.5176,
      "eval_samples_per_second": 6.808,
      "eval_steps_per_second": 0.869,
      "step": 9438
    },
    {
      "epoch": 143.03030303030303,
      "grad_norm": 0.03308232128620148,
      "learning_rate": 2.8484848484848486e-05,
      "loss": 0.0114,
      "step": 9440
    },
    {
      "epoch": 143.1060606060606,
      "grad_norm": 0.05524769425392151,
      "learning_rate": 2.84469696969697e-05,
      "loss": 0.0125,
      "step": 9445
    },
    {
      "epoch": 143.1818181818182,
      "grad_norm": 0.07751823961734772,
      "learning_rate": 2.8409090909090912e-05,
      "loss": 0.007,
      "step": 9450
    },
    {
      "epoch": 143.25757575757575,
      "grad_norm": 0.09663157165050507,
      "learning_rate": 2.8371212121212122e-05,
      "loss": 0.0088,
      "step": 9455
    },
    {
      "epoch": 143.33333333333334,
      "grad_norm": 0.04683161899447441,
      "learning_rate": 2.8333333333333335e-05,
      "loss": 0.0094,
      "step": 9460
    },
    {
      "epoch": 143.4090909090909,
      "grad_norm": 0.04813295602798462,
      "learning_rate": 2.829545454545455e-05,
      "loss": 0.0114,
      "step": 9465
    },
    {
      "epoch": 143.4848484848485,
      "grad_norm": 0.03996315598487854,
      "learning_rate": 2.825757575757576e-05,
      "loss": 0.0069,
      "step": 9470
    },
    {
      "epoch": 143.56060606060606,
      "grad_norm": 0.049014125019311905,
      "learning_rate": 2.8219696969696975e-05,
      "loss": 0.0114,
      "step": 9475
    },
    {
      "epoch": 143.63636363636363,
      "grad_norm": 0.049554113298654556,
      "learning_rate": 2.818181818181818e-05,
      "loss": 0.0092,
      "step": 9480
    },
    {
      "epoch": 143.71212121212122,
      "grad_norm": 0.05448229983448982,
      "learning_rate": 2.8143939393939394e-05,
      "loss": 0.0089,
      "step": 9485
    },
    {
      "epoch": 143.78787878787878,
      "grad_norm": 0.08751636743545532,
      "learning_rate": 2.8106060606060607e-05,
      "loss": 0.0103,
      "step": 9490
    },
    {
      "epoch": 143.86363636363637,
      "grad_norm": 0.032574355602264404,
      "learning_rate": 2.806818181818182e-05,
      "loss": 0.0067,
      "step": 9495
    },
    {
      "epoch": 143.93939393939394,
      "grad_norm": 0.03482254967093468,
      "learning_rate": 2.803030303030303e-05,
      "loss": 0.0105,
      "step": 9500
    },
    {
      "epoch": 144.0,
      "eval_loss": 0.011357367970049381,
      "eval_mean_accuracy": 0.9534591800801878,
      "eval_mean_iou": 0.9263249426507678,
      "eval_runtime": 35.3722,
      "eval_samples_per_second": 6.644,
      "eval_steps_per_second": 0.848,
      "step": 9504
    },
    {
      "epoch": 144.0151515151515,
      "grad_norm": 0.02069547213613987,
      "learning_rate": 2.7992424242424244e-05,
      "loss": 0.0086,
      "step": 9505
    },
    {
      "epoch": 144.0909090909091,
      "grad_norm": 0.06391298770904541,
      "learning_rate": 2.7954545454545457e-05,
      "loss": 0.0082,
      "step": 9510
    },
    {
      "epoch": 144.16666666666666,
      "grad_norm": 0.07772085070610046,
      "learning_rate": 2.791666666666667e-05,
      "loss": 0.0084,
      "step": 9515
    },
    {
      "epoch": 144.24242424242425,
      "grad_norm": 0.03711830824613571,
      "learning_rate": 2.7878787878787883e-05,
      "loss": 0.0078,
      "step": 9520
    },
    {
      "epoch": 144.3181818181818,
      "grad_norm": 0.029929134994745255,
      "learning_rate": 2.784090909090909e-05,
      "loss": 0.0111,
      "step": 9525
    },
    {
      "epoch": 144.3939393939394,
      "grad_norm": 0.01979893073439598,
      "learning_rate": 2.7803030303030303e-05,
      "loss": 0.0088,
      "step": 9530
    },
    {
      "epoch": 144.46969696969697,
      "grad_norm": 0.031434886157512665,
      "learning_rate": 2.7765151515151516e-05,
      "loss": 0.0107,
      "step": 9535
    },
    {
      "epoch": 144.54545454545453,
      "grad_norm": 0.032604802399873734,
      "learning_rate": 2.772727272727273e-05,
      "loss": 0.0092,
      "step": 9540
    },
    {
      "epoch": 144.62121212121212,
      "grad_norm": 0.020434832200407982,
      "learning_rate": 2.7689393939393942e-05,
      "loss": 0.0091,
      "step": 9545
    },
    {
      "epoch": 144.6969696969697,
      "grad_norm": 0.03851557895541191,
      "learning_rate": 2.7651515151515152e-05,
      "loss": 0.0109,
      "step": 9550
    },
    {
      "epoch": 144.77272727272728,
      "grad_norm": 0.052175335586071014,
      "learning_rate": 2.7613636363636365e-05,
      "loss": 0.0112,
      "step": 9555
    },
    {
      "epoch": 144.84848484848484,
      "grad_norm": 0.07969200611114502,
      "learning_rate": 2.7575757575757578e-05,
      "loss": 0.0114,
      "step": 9560
    },
    {
      "epoch": 144.92424242424244,
      "grad_norm": 0.06430736929178238,
      "learning_rate": 2.753787878787879e-05,
      "loss": 0.009,
      "step": 9565
    },
    {
      "epoch": 145.0,
      "grad_norm": 0.06336387246847153,
      "learning_rate": 2.7500000000000004e-05,
      "loss": 0.0091,
      "step": 9570
    },
    {
      "epoch": 145.0,
      "eval_loss": 0.011331645771861076,
      "eval_mean_accuracy": 0.9570791081719084,
      "eval_mean_iou": 0.9272519678560452,
      "eval_runtime": 34.195,
      "eval_samples_per_second": 6.872,
      "eval_steps_per_second": 0.877,
      "step": 9570
    },
    {
      "epoch": 145.07575757575756,
      "grad_norm": 0.06045173853635788,
      "learning_rate": 2.746212121212121e-05,
      "loss": 0.0095,
      "step": 9575
    },
    {
      "epoch": 145.15151515151516,
      "grad_norm": 0.03139245882630348,
      "learning_rate": 2.7424242424242424e-05,
      "loss": 0.0108,
      "step": 9580
    },
    {
      "epoch": 145.22727272727272,
      "grad_norm": 0.05391242355108261,
      "learning_rate": 2.7386363636363637e-05,
      "loss": 0.0087,
      "step": 9585
    },
    {
      "epoch": 145.3030303030303,
      "grad_norm": 0.04533889889717102,
      "learning_rate": 2.734848484848485e-05,
      "loss": 0.0094,
      "step": 9590
    },
    {
      "epoch": 145.37878787878788,
      "grad_norm": 0.041605543345212936,
      "learning_rate": 2.7310606060606063e-05,
      "loss": 0.009,
      "step": 9595
    },
    {
      "epoch": 145.45454545454547,
      "grad_norm": 0.03732568398118019,
      "learning_rate": 2.7272727272727273e-05,
      "loss": 0.0079,
      "step": 9600
    },
    {
      "epoch": 145.53030303030303,
      "grad_norm": 0.0638463944196701,
      "learning_rate": 2.7234848484848486e-05,
      "loss": 0.0109,
      "step": 9605
    },
    {
      "epoch": 145.6060606060606,
      "grad_norm": 0.07542566955089569,
      "learning_rate": 2.71969696969697e-05,
      "loss": 0.0087,
      "step": 9610
    },
    {
      "epoch": 145.6818181818182,
      "grad_norm": 0.03826601430773735,
      "learning_rate": 2.7159090909090913e-05,
      "loss": 0.0076,
      "step": 9615
    },
    {
      "epoch": 145.75757575757575,
      "grad_norm": 0.060041509568691254,
      "learning_rate": 2.7121212121212126e-05,
      "loss": 0.0103,
      "step": 9620
    },
    {
      "epoch": 145.83333333333334,
      "grad_norm": 0.06424738466739655,
      "learning_rate": 2.7083333333333332e-05,
      "loss": 0.0101,
      "step": 9625
    },
    {
      "epoch": 145.9090909090909,
      "grad_norm": 0.025485126301646233,
      "learning_rate": 2.7045454545454545e-05,
      "loss": 0.0081,
      "step": 9630
    },
    {
      "epoch": 145.9848484848485,
      "grad_norm": 0.04055999219417572,
      "learning_rate": 2.700757575757576e-05,
      "loss": 0.0121,
      "step": 9635
    },
    {
      "epoch": 146.0,
      "eval_loss": 0.011180225759744644,
      "eval_mean_accuracy": 0.9584970332176911,
      "eval_mean_iou": 0.9279655990146072,
      "eval_runtime": 35.1292,
      "eval_samples_per_second": 6.69,
      "eval_steps_per_second": 0.854,
      "step": 9636
    },
    {
      "epoch": 146.06060606060606,
      "grad_norm": 0.04757855832576752,
      "learning_rate": 2.696969696969697e-05,
      "loss": 0.0112,
      "step": 9640
    },
    {
      "epoch": 146.13636363636363,
      "grad_norm": 0.029668770730495453,
      "learning_rate": 2.6931818181818185e-05,
      "loss": 0.0119,
      "step": 9645
    },
    {
      "epoch": 146.21212121212122,
      "grad_norm": 0.031325813382864,
      "learning_rate": 2.6893939393939394e-05,
      "loss": 0.0067,
      "step": 9650
    },
    {
      "epoch": 146.28787878787878,
      "grad_norm": 0.02415735274553299,
      "learning_rate": 2.6856060606060608e-05,
      "loss": 0.0108,
      "step": 9655
    },
    {
      "epoch": 146.36363636363637,
      "grad_norm": 0.03624691814184189,
      "learning_rate": 2.681818181818182e-05,
      "loss": 0.0082,
      "step": 9660
    },
    {
      "epoch": 146.43939393939394,
      "grad_norm": 0.04097117856144905,
      "learning_rate": 2.6780303030303034e-05,
      "loss": 0.0099,
      "step": 9665
    },
    {
      "epoch": 146.5151515151515,
      "grad_norm": 0.02718796767294407,
      "learning_rate": 2.674242424242424e-05,
      "loss": 0.0116,
      "step": 9670
    },
    {
      "epoch": 146.5909090909091,
      "grad_norm": 0.0513654462993145,
      "learning_rate": 2.6704545454545453e-05,
      "loss": 0.0074,
      "step": 9675
    },
    {
      "epoch": 146.66666666666666,
      "grad_norm": 0.07524017989635468,
      "learning_rate": 2.6666666666666667e-05,
      "loss": 0.0118,
      "step": 9680
    },
    {
      "epoch": 146.74242424242425,
      "grad_norm": 0.030494479462504387,
      "learning_rate": 2.662878787878788e-05,
      "loss": 0.0084,
      "step": 9685
    },
    {
      "epoch": 146.8181818181818,
      "grad_norm": 0.0624314546585083,
      "learning_rate": 2.6590909090909093e-05,
      "loss": 0.0105,
      "step": 9690
    },
    {
      "epoch": 146.8939393939394,
      "grad_norm": 0.02995832823216915,
      "learning_rate": 2.6553030303030303e-05,
      "loss": 0.0087,
      "step": 9695
    },
    {
      "epoch": 146.96969696969697,
      "grad_norm": 0.026138320565223694,
      "learning_rate": 2.6515151515151516e-05,
      "loss": 0.0061,
      "step": 9700
    },
    {
      "epoch": 147.0,
      "eval_loss": 0.011477365158498287,
      "eval_mean_accuracy": 0.9537119101764765,
      "eval_mean_iou": 0.925751299414095,
      "eval_runtime": 34.1534,
      "eval_samples_per_second": 6.881,
      "eval_steps_per_second": 0.878,
      "step": 9702
    },
    {
      "epoch": 147.04545454545453,
      "grad_norm": 0.034230612218379974,
      "learning_rate": 2.647727272727273e-05,
      "loss": 0.0114,
      "step": 9705
    },
    {
      "epoch": 147.12121212121212,
      "grad_norm": 0.08833404630422592,
      "learning_rate": 2.6439393939393942e-05,
      "loss": 0.0094,
      "step": 9710
    },
    {
      "epoch": 147.1969696969697,
      "grad_norm": 0.06377340853214264,
      "learning_rate": 2.6401515151515155e-05,
      "loss": 0.0089,
      "step": 9715
    },
    {
      "epoch": 147.27272727272728,
      "grad_norm": 0.05134362727403641,
      "learning_rate": 2.636363636363636e-05,
      "loss": 0.0091,
      "step": 9720
    },
    {
      "epoch": 147.34848484848484,
      "grad_norm": 0.04731402173638344,
      "learning_rate": 2.6325757575757575e-05,
      "loss": 0.0083,
      "step": 9725
    },
    {
      "epoch": 147.42424242424244,
      "grad_norm": 0.08717456459999084,
      "learning_rate": 2.6287878787878788e-05,
      "loss": 0.0112,
      "step": 9730
    },
    {
      "epoch": 147.5,
      "grad_norm": 0.09381306171417236,
      "learning_rate": 2.625e-05,
      "loss": 0.0104,
      "step": 9735
    },
    {
      "epoch": 147.57575757575756,
      "grad_norm": 0.03365860506892204,
      "learning_rate": 2.6212121212121214e-05,
      "loss": 0.009,
      "step": 9740
    },
    {
      "epoch": 147.65151515151516,
      "grad_norm": 0.04399186000227928,
      "learning_rate": 2.6174242424242424e-05,
      "loss": 0.0089,
      "step": 9745
    },
    {
      "epoch": 147.72727272727272,
      "grad_norm": 0.03230700269341469,
      "learning_rate": 2.6136363636363637e-05,
      "loss": 0.0085,
      "step": 9750
    },
    {
      "epoch": 147.8030303030303,
      "grad_norm": 0.03984415903687477,
      "learning_rate": 2.609848484848485e-05,
      "loss": 0.0086,
      "step": 9755
    },
    {
      "epoch": 147.87878787878788,
      "grad_norm": 0.03905673325061798,
      "learning_rate": 2.6060606060606063e-05,
      "loss": 0.0108,
      "step": 9760
    },
    {
      "epoch": 147.95454545454547,
      "grad_norm": 0.033622827380895615,
      "learning_rate": 2.6022727272727277e-05,
      "loss": 0.0089,
      "step": 9765
    },
    {
      "epoch": 148.0,
      "eval_loss": 0.011325567029416561,
      "eval_mean_accuracy": 0.9588516339191774,
      "eval_mean_iou": 0.9282313447937139,
      "eval_runtime": 36.4579,
      "eval_samples_per_second": 6.446,
      "eval_steps_per_second": 0.823,
      "step": 9768
    },
    {
      "epoch": 148.03030303030303,
      "grad_norm": 0.019589023664593697,
      "learning_rate": 2.5984848484848483e-05,
      "loss": 0.0083,
      "step": 9770
    },
    {
      "epoch": 148.1060606060606,
      "grad_norm": 0.038650788366794586,
      "learning_rate": 2.5946969696969696e-05,
      "loss": 0.0086,
      "step": 9775
    },
    {
      "epoch": 148.1818181818182,
      "grad_norm": 0.026437709107995033,
      "learning_rate": 2.590909090909091e-05,
      "loss": 0.0082,
      "step": 9780
    },
    {
      "epoch": 148.25757575757575,
      "grad_norm": 0.06517606973648071,
      "learning_rate": 2.5871212121212122e-05,
      "loss": 0.0108,
      "step": 9785
    },
    {
      "epoch": 148.33333333333334,
      "grad_norm": 0.032494932413101196,
      "learning_rate": 2.5833333333333336e-05,
      "loss": 0.0074,
      "step": 9790
    },
    {
      "epoch": 148.4090909090909,
      "grad_norm": 0.0408414863049984,
      "learning_rate": 2.5795454545454545e-05,
      "loss": 0.0081,
      "step": 9795
    },
    {
      "epoch": 148.4848484848485,
      "grad_norm": 0.017063258215785027,
      "learning_rate": 2.575757575757576e-05,
      "loss": 0.0092,
      "step": 9800
    },
    {
      "epoch": 148.56060606060606,
      "grad_norm": 0.07835640013217926,
      "learning_rate": 2.571969696969697e-05,
      "loss": 0.0109,
      "step": 9805
    },
    {
      "epoch": 148.63636363636363,
      "grad_norm": 0.03404843062162399,
      "learning_rate": 2.5681818181818185e-05,
      "loss": 0.0087,
      "step": 9810
    },
    {
      "epoch": 148.71212121212122,
      "grad_norm": 0.026407428085803986,
      "learning_rate": 2.5643939393939398e-05,
      "loss": 0.0132,
      "step": 9815
    },
    {
      "epoch": 148.78787878787878,
      "grad_norm": 0.06405722349882126,
      "learning_rate": 2.5606060606060604e-05,
      "loss": 0.0097,
      "step": 9820
    },
    {
      "epoch": 148.86363636363637,
      "grad_norm": 0.039846234023571014,
      "learning_rate": 2.5568181818181817e-05,
      "loss": 0.0076,
      "step": 9825
    },
    {
      "epoch": 148.93939393939394,
      "grad_norm": 0.04348624870181084,
      "learning_rate": 2.553030303030303e-05,
      "loss": 0.0098,
      "step": 9830
    },
    {
      "epoch": 149.0,
      "eval_loss": 0.011515659280121326,
      "eval_mean_accuracy": 0.9526248456777645,
      "eval_mean_iou": 0.9260448003183491,
      "eval_runtime": 34.5001,
      "eval_samples_per_second": 6.812,
      "eval_steps_per_second": 0.87,
      "step": 9834
    },
    {
      "epoch": 149.0151515151515,
      "grad_norm": 0.04746094346046448,
      "learning_rate": 2.5492424242424244e-05,
      "loss": 0.0071,
      "step": 9835
    },
    {
      "epoch": 149.0909090909091,
      "grad_norm": 0.06539618968963623,
      "learning_rate": 2.5454545454545454e-05,
      "loss": 0.0101,
      "step": 9840
    },
    {
      "epoch": 149.16666666666666,
      "grad_norm": 0.04768550395965576,
      "learning_rate": 2.5416666666666667e-05,
      "loss": 0.0104,
      "step": 9845
    },
    {
      "epoch": 149.24242424242425,
      "grad_norm": 0.03970608860254288,
      "learning_rate": 2.537878787878788e-05,
      "loss": 0.0088,
      "step": 9850
    },
    {
      "epoch": 149.3181818181818,
      "grad_norm": 0.02308507449924946,
      "learning_rate": 2.5340909090909093e-05,
      "loss": 0.0073,
      "step": 9855
    },
    {
      "epoch": 149.3939393939394,
      "grad_norm": 0.01706257089972496,
      "learning_rate": 2.5303030303030306e-05,
      "loss": 0.0076,
      "step": 9860
    },
    {
      "epoch": 149.46969696969697,
      "grad_norm": 0.06288671493530273,
      "learning_rate": 2.5265151515151513e-05,
      "loss": 0.0115,
      "step": 9865
    },
    {
      "epoch": 149.54545454545453,
      "grad_norm": 0.042919714003801346,
      "learning_rate": 2.5227272727272726e-05,
      "loss": 0.0082,
      "step": 9870
    },
    {
      "epoch": 149.62121212121212,
      "grad_norm": 0.022503746673464775,
      "learning_rate": 2.518939393939394e-05,
      "loss": 0.0091,
      "step": 9875
    },
    {
      "epoch": 149.6969696969697,
      "grad_norm": 0.055113665759563446,
      "learning_rate": 2.5151515151515155e-05,
      "loss": 0.0091,
      "step": 9880
    },
    {
      "epoch": 149.77272727272728,
      "grad_norm": 0.03529473766684532,
      "learning_rate": 2.511363636363637e-05,
      "loss": 0.0076,
      "step": 9885
    },
    {
      "epoch": 149.84848484848484,
      "grad_norm": 0.08923030644655228,
      "learning_rate": 2.5075757575757575e-05,
      "loss": 0.0097,
      "step": 9890
    },
    {
      "epoch": 149.92424242424244,
      "grad_norm": 0.026113077998161316,
      "learning_rate": 2.5037878787878788e-05,
      "loss": 0.0122,
      "step": 9895
    },
    {
      "epoch": 150.0,
      "grad_norm": 0.05537215620279312,
      "learning_rate": 2.5e-05,
      "loss": 0.009,
      "step": 9900
    },
    {
      "epoch": 150.0,
      "eval_loss": 0.011325099505484104,
      "eval_mean_accuracy": 0.9525917695397959,
      "eval_mean_iou": 0.9262480495296138,
      "eval_runtime": 35.2213,
      "eval_samples_per_second": 6.672,
      "eval_steps_per_second": 0.852,
      "step": 9900
    },
    {
      "epoch": 150.07575757575756,
      "grad_norm": 0.053478632122278214,
      "learning_rate": 2.4962121212121214e-05,
      "loss": 0.0064,
      "step": 9905
    },
    {
      "epoch": 150.15151515151516,
      "grad_norm": 0.04678941145539284,
      "learning_rate": 2.4924242424242424e-05,
      "loss": 0.0101,
      "step": 9910
    },
    {
      "epoch": 150.22727272727272,
      "grad_norm": 0.09064902365207672,
      "learning_rate": 2.4886363636363637e-05,
      "loss": 0.0099,
      "step": 9915
    },
    {
      "epoch": 150.3030303030303,
      "grad_norm": 0.10668646544218063,
      "learning_rate": 2.4848484848484847e-05,
      "loss": 0.0108,
      "step": 9920
    },
    {
      "epoch": 150.37878787878788,
      "grad_norm": 0.09310371428728104,
      "learning_rate": 2.4810606060606064e-05,
      "loss": 0.0111,
      "step": 9925
    },
    {
      "epoch": 150.45454545454547,
      "grad_norm": 0.04042200744152069,
      "learning_rate": 2.4772727272727277e-05,
      "loss": 0.0091,
      "step": 9930
    },
    {
      "epoch": 150.53030303030303,
      "grad_norm": 0.05566850304603577,
      "learning_rate": 2.4734848484848486e-05,
      "loss": 0.0107,
      "step": 9935
    },
    {
      "epoch": 150.6060606060606,
      "grad_norm": 0.05805859714746475,
      "learning_rate": 2.46969696969697e-05,
      "loss": 0.0075,
      "step": 9940
    },
    {
      "epoch": 150.6818181818182,
      "grad_norm": 0.044396255165338516,
      "learning_rate": 2.465909090909091e-05,
      "loss": 0.0095,
      "step": 9945
    },
    {
      "epoch": 150.75757575757575,
      "grad_norm": 0.03019944205880165,
      "learning_rate": 2.4621212121212123e-05,
      "loss": 0.0101,
      "step": 9950
    },
    {
      "epoch": 150.83333333333334,
      "grad_norm": 0.038101717829704285,
      "learning_rate": 2.4583333333333332e-05,
      "loss": 0.011,
      "step": 9955
    },
    {
      "epoch": 150.9090909090909,
      "grad_norm": 0.02200940251350403,
      "learning_rate": 2.4545454545454545e-05,
      "loss": 0.0079,
      "step": 9960
    },
    {
      "epoch": 150.9848484848485,
      "grad_norm": 0.050113022327423096,
      "learning_rate": 2.450757575757576e-05,
      "loss": 0.0082,
      "step": 9965
    },
    {
      "epoch": 151.0,
      "eval_loss": 0.011452735401690006,
      "eval_mean_accuracy": 0.9544279499663108,
      "eval_mean_iou": 0.9265112025434052,
      "eval_runtime": 34.183,
      "eval_samples_per_second": 6.875,
      "eval_steps_per_second": 0.878,
      "step": 9966
    },
    {
      "epoch": 151.06060606060606,
      "grad_norm": 0.04733998328447342,
      "learning_rate": 2.4469696969696972e-05,
      "loss": 0.0082,
      "step": 9970
    },
    {
      "epoch": 151.13636363636363,
      "grad_norm": 0.09643007069826126,
      "learning_rate": 2.4431818181818185e-05,
      "loss": 0.0089,
      "step": 9975
    },
    {
      "epoch": 151.21212121212122,
      "grad_norm": 0.04186355322599411,
      "learning_rate": 2.4393939393939395e-05,
      "loss": 0.0089,
      "step": 9980
    },
    {
      "epoch": 151.28787878787878,
      "grad_norm": 0.03889001905918121,
      "learning_rate": 2.4356060606060608e-05,
      "loss": 0.0102,
      "step": 9985
    },
    {
      "epoch": 151.36363636363637,
      "grad_norm": 0.04066826030611992,
      "learning_rate": 2.431818181818182e-05,
      "loss": 0.0087,
      "step": 9990
    },
    {
      "epoch": 151.43939393939394,
      "grad_norm": 0.08011851459741592,
      "learning_rate": 2.428030303030303e-05,
      "loss": 0.0089,
      "step": 9995
    },
    {
      "epoch": 151.5151515151515,
      "grad_norm": 0.02497289702296257,
      "learning_rate": 2.4242424242424244e-05,
      "loss": 0.0087,
      "step": 10000
    },
    {
      "epoch": 151.5909090909091,
      "grad_norm": 0.08738812059164047,
      "learning_rate": 2.4204545454545454e-05,
      "loss": 0.0111,
      "step": 10005
    },
    {
      "epoch": 151.66666666666666,
      "grad_norm": 0.03942103683948517,
      "learning_rate": 2.4166666666666667e-05,
      "loss": 0.0077,
      "step": 10010
    },
    {
      "epoch": 151.74242424242425,
      "grad_norm": 0.0332251600921154,
      "learning_rate": 2.412878787878788e-05,
      "loss": 0.0081,
      "step": 10015
    },
    {
      "epoch": 151.8181818181818,
      "grad_norm": 0.06483544409275055,
      "learning_rate": 2.4090909090909093e-05,
      "loss": 0.0111,
      "step": 10020
    },
    {
      "epoch": 151.8939393939394,
      "grad_norm": 0.024844318628311157,
      "learning_rate": 2.4053030303030306e-05,
      "loss": 0.009,
      "step": 10025
    },
    {
      "epoch": 151.96969696969697,
      "grad_norm": 0.046690989285707474,
      "learning_rate": 2.4015151515151516e-05,
      "loss": 0.0083,
      "step": 10030
    },
    {
      "epoch": 152.0,
      "eval_loss": 0.011377842165529728,
      "eval_mean_accuracy": 0.9534133521408762,
      "eval_mean_iou": 0.925840327717752,
      "eval_runtime": 36.6915,
      "eval_samples_per_second": 6.405,
      "eval_steps_per_second": 0.818,
      "step": 10032
    },
    {
      "epoch": 152.04545454545453,
      "grad_norm": 0.04895268380641937,
      "learning_rate": 2.397727272727273e-05,
      "loss": 0.011,
      "step": 10035
    },
    {
      "epoch": 152.12121212121212,
      "grad_norm": 0.03939783573150635,
      "learning_rate": 2.393939393939394e-05,
      "loss": 0.0066,
      "step": 10040
    },
    {
      "epoch": 152.1969696969697,
      "grad_norm": 0.03606943041086197,
      "learning_rate": 2.3901515151515152e-05,
      "loss": 0.0094,
      "step": 10045
    },
    {
      "epoch": 152.27272727272728,
      "grad_norm": 0.03762936219573021,
      "learning_rate": 2.3863636363636365e-05,
      "loss": 0.0082,
      "step": 10050
    },
    {
      "epoch": 152.34848484848484,
      "grad_norm": 0.05530097708106041,
      "learning_rate": 2.3825757575757575e-05,
      "loss": 0.0113,
      "step": 10055
    },
    {
      "epoch": 152.42424242424244,
      "grad_norm": 0.03165353089570999,
      "learning_rate": 2.3787878787878788e-05,
      "loss": 0.0109,
      "step": 10060
    },
    {
      "epoch": 152.5,
      "grad_norm": 0.03188694268465042,
      "learning_rate": 2.375e-05,
      "loss": 0.009,
      "step": 10065
    },
    {
      "epoch": 152.57575757575756,
      "grad_norm": 0.03723525628447533,
      "learning_rate": 2.3712121212121214e-05,
      "loss": 0.0083,
      "step": 10070
    },
    {
      "epoch": 152.65151515151516,
      "grad_norm": 0.030428113415837288,
      "learning_rate": 2.3674242424242428e-05,
      "loss": 0.0099,
      "step": 10075
    },
    {
      "epoch": 152.72727272727272,
      "grad_norm": 0.022990161553025246,
      "learning_rate": 2.3636363636363637e-05,
      "loss": 0.0116,
      "step": 10080
    },
    {
      "epoch": 152.8030303030303,
      "grad_norm": 0.10713709145784378,
      "learning_rate": 2.359848484848485e-05,
      "loss": 0.0094,
      "step": 10085
    },
    {
      "epoch": 152.87878787878788,
      "grad_norm": 0.030324656516313553,
      "learning_rate": 2.356060606060606e-05,
      "loss": 0.0074,
      "step": 10090
    },
    {
      "epoch": 152.95454545454547,
      "grad_norm": 0.035537876188755035,
      "learning_rate": 2.3522727272727273e-05,
      "loss": 0.013,
      "step": 10095
    },
    {
      "epoch": 153.0,
      "eval_loss": 0.011314652860164642,
      "eval_mean_accuracy": 0.9618376655228115,
      "eval_mean_iou": 0.928575550480813,
      "eval_runtime": 34.734,
      "eval_samples_per_second": 6.766,
      "eval_steps_per_second": 0.864,
      "step": 10098
    },
    {
      "epoch": 153.03030303030303,
      "grad_norm": 0.07702215015888214,
      "learning_rate": 2.3484848484848487e-05,
      "loss": 0.0077,
      "step": 10100
    },
    {
      "epoch": 153.1060606060606,
      "grad_norm": 0.04306747391819954,
      "learning_rate": 2.3446969696969696e-05,
      "loss": 0.0108,
      "step": 10105
    },
    {
      "epoch": 153.1818181818182,
      "grad_norm": 0.08003927022218704,
      "learning_rate": 2.340909090909091e-05,
      "loss": 0.0103,
      "step": 10110
    },
    {
      "epoch": 153.25757575757575,
      "grad_norm": 0.06024370715022087,
      "learning_rate": 2.3371212121212123e-05,
      "loss": 0.009,
      "step": 10115
    },
    {
      "epoch": 153.33333333333334,
      "grad_norm": 0.03968359902501106,
      "learning_rate": 2.3333333333333336e-05,
      "loss": 0.009,
      "step": 10120
    },
    {
      "epoch": 153.4090909090909,
      "grad_norm": 0.05242961272597313,
      "learning_rate": 2.3295454545454546e-05,
      "loss": 0.0081,
      "step": 10125
    },
    {
      "epoch": 153.4848484848485,
      "grad_norm": 0.04464291036128998,
      "learning_rate": 2.325757575757576e-05,
      "loss": 0.0073,
      "step": 10130
    },
    {
      "epoch": 153.56060606060606,
      "grad_norm": 0.1240650862455368,
      "learning_rate": 2.3219696969696972e-05,
      "loss": 0.0105,
      "step": 10135
    },
    {
      "epoch": 153.63636363636363,
      "grad_norm": 0.049628887325525284,
      "learning_rate": 2.318181818181818e-05,
      "loss": 0.0105,
      "step": 10140
    },
    {
      "epoch": 153.71212121212122,
      "grad_norm": 0.049718987196683884,
      "learning_rate": 2.3143939393939395e-05,
      "loss": 0.012,
      "step": 10145
    },
    {
      "epoch": 153.78787878787878,
      "grad_norm": 0.018459497019648552,
      "learning_rate": 2.3106060606060605e-05,
      "loss": 0.0067,
      "step": 10150
    },
    {
      "epoch": 153.86363636363637,
      "grad_norm": 0.03318193182349205,
      "learning_rate": 2.3068181818181818e-05,
      "loss": 0.0088,
      "step": 10155
    },
    {
      "epoch": 153.93939393939394,
      "grad_norm": 0.049415312707424164,
      "learning_rate": 2.3030303030303034e-05,
      "loss": 0.0079,
      "step": 10160
    },
    {
      "epoch": 154.0,
      "eval_loss": 0.011402730830013752,
      "eval_mean_accuracy": 0.9517347098238232,
      "eval_mean_iou": 0.9257909069472344,
      "eval_runtime": 33.8904,
      "eval_samples_per_second": 6.934,
      "eval_steps_per_second": 0.885,
      "step": 10164
    },
    {
      "epoch": 154.0151515151515,
      "grad_norm": 0.03453397750854492,
      "learning_rate": 2.2992424242424244e-05,
      "loss": 0.0087,
      "step": 10165
    },
    {
      "epoch": 154.0909090909091,
      "grad_norm": 0.04217781871557236,
      "learning_rate": 2.2954545454545457e-05,
      "loss": 0.0106,
      "step": 10170
    },
    {
      "epoch": 154.16666666666666,
      "grad_norm": 0.02656628005206585,
      "learning_rate": 2.2916666666666667e-05,
      "loss": 0.0076,
      "step": 10175
    },
    {
      "epoch": 154.24242424242425,
      "grad_norm": 0.04012560099363327,
      "learning_rate": 2.287878787878788e-05,
      "loss": 0.0076,
      "step": 10180
    },
    {
      "epoch": 154.3181818181818,
      "grad_norm": 0.03208084776997566,
      "learning_rate": 2.2840909090909093e-05,
      "loss": 0.009,
      "step": 10185
    },
    {
      "epoch": 154.3939393939394,
      "grad_norm": 0.046852413564920425,
      "learning_rate": 2.2803030303030303e-05,
      "loss": 0.0104,
      "step": 10190
    },
    {
      "epoch": 154.46969696969697,
      "grad_norm": 0.041021399199962616,
      "learning_rate": 2.2765151515151516e-05,
      "loss": 0.0093,
      "step": 10195
    },
    {
      "epoch": 154.54545454545453,
      "grad_norm": 0.034494008868932724,
      "learning_rate": 2.272727272727273e-05,
      "loss": 0.0098,
      "step": 10200
    },
    {
      "epoch": 154.62121212121212,
      "grad_norm": 0.04534567520022392,
      "learning_rate": 2.2689393939393942e-05,
      "loss": 0.0078,
      "step": 10205
    },
    {
      "epoch": 154.6969696969697,
      "grad_norm": 0.032059196382761,
      "learning_rate": 2.2651515151515152e-05,
      "loss": 0.0092,
      "step": 10210
    },
    {
      "epoch": 154.77272727272728,
      "grad_norm": 0.057678911834955215,
      "learning_rate": 2.2613636363636365e-05,
      "loss": 0.0127,
      "step": 10215
    },
    {
      "epoch": 154.84848484848484,
      "grad_norm": 0.019214339554309845,
      "learning_rate": 2.257575757575758e-05,
      "loss": 0.0084,
      "step": 10220
    },
    {
      "epoch": 154.92424242424244,
      "grad_norm": 0.051868971437215805,
      "learning_rate": 2.2537878787878788e-05,
      "loss": 0.0075,
      "step": 10225
    },
    {
      "epoch": 155.0,
      "grad_norm": 0.018364429473876953,
      "learning_rate": 2.25e-05,
      "loss": 0.0083,
      "step": 10230
    },
    {
      "epoch": 155.0,
      "eval_loss": 0.011282687075436115,
      "eval_mean_accuracy": 0.9580683972259559,
      "eval_mean_iou": 0.9276679294527471,
      "eval_runtime": 34.579,
      "eval_samples_per_second": 6.796,
      "eval_steps_per_second": 0.868,
      "step": 10230
    },
    {
      "epoch": 155.07575757575756,
      "grad_norm": 0.058082520961761475,
      "learning_rate": 2.246212121212121e-05,
      "loss": 0.0081,
      "step": 10235
    },
    {
      "epoch": 155.15151515151516,
      "grad_norm": 0.058119162917137146,
      "learning_rate": 2.2424242424242424e-05,
      "loss": 0.0112,
      "step": 10240
    },
    {
      "epoch": 155.22727272727272,
      "grad_norm": 0.045244764536619186,
      "learning_rate": 2.2386363636363637e-05,
      "loss": 0.0092,
      "step": 10245
    },
    {
      "epoch": 155.3030303030303,
      "grad_norm": 0.03140517696738243,
      "learning_rate": 2.234848484848485e-05,
      "loss": 0.009,
      "step": 10250
    },
    {
      "epoch": 155.37878787878788,
      "grad_norm": 0.019806994125247,
      "learning_rate": 2.2310606060606064e-05,
      "loss": 0.0072,
      "step": 10255
    },
    {
      "epoch": 155.45454545454547,
      "grad_norm": 0.04551331698894501,
      "learning_rate": 2.2272727272727274e-05,
      "loss": 0.0087,
      "step": 10260
    },
    {
      "epoch": 155.53030303030303,
      "grad_norm": 0.055262498557567596,
      "learning_rate": 2.2234848484848487e-05,
      "loss": 0.0105,
      "step": 10265
    },
    {
      "epoch": 155.6060606060606,
      "grad_norm": 0.11201003938913345,
      "learning_rate": 2.21969696969697e-05,
      "loss": 0.009,
      "step": 10270
    },
    {
      "epoch": 155.6818181818182,
      "grad_norm": 0.04275002330541611,
      "learning_rate": 2.215909090909091e-05,
      "loss": 0.0114,
      "step": 10275
    },
    {
      "epoch": 155.75757575757575,
      "grad_norm": 0.060243960469961166,
      "learning_rate": 2.2121212121212123e-05,
      "loss": 0.009,
      "step": 10280
    },
    {
      "epoch": 155.83333333333334,
      "grad_norm": 0.04212094470858574,
      "learning_rate": 2.2083333333333333e-05,
      "loss": 0.0084,
      "step": 10285
    },
    {
      "epoch": 155.9090909090909,
      "grad_norm": 0.019151635468006134,
      "learning_rate": 2.2045454545454546e-05,
      "loss": 0.007,
      "step": 10290
    },
    {
      "epoch": 155.9848484848485,
      "grad_norm": 0.058512941002845764,
      "learning_rate": 2.200757575757576e-05,
      "loss": 0.0103,
      "step": 10295
    },
    {
      "epoch": 156.0,
      "eval_loss": 0.011413468979299068,
      "eval_mean_accuracy": 0.9519878559377025,
      "eval_mean_iou": 0.9259959464810392,
      "eval_runtime": 36.1353,
      "eval_samples_per_second": 6.503,
      "eval_steps_per_second": 0.83,
      "step": 10296
    },
    {
      "epoch": 156.06060606060606,
      "grad_norm": 0.03750451281666756,
      "learning_rate": 2.1969696969696972e-05,
      "loss": 0.0077,
      "step": 10300
    },
    {
      "epoch": 156.13636363636363,
      "grad_norm": 0.030226558446884155,
      "learning_rate": 2.1931818181818185e-05,
      "loss": 0.0076,
      "step": 10305
    },
    {
      "epoch": 156.21212121212122,
      "grad_norm": 0.03709398955106735,
      "learning_rate": 2.1893939393939395e-05,
      "loss": 0.0097,
      "step": 10310
    },
    {
      "epoch": 156.28787878787878,
      "grad_norm": 0.04763031005859375,
      "learning_rate": 2.1856060606060608e-05,
      "loss": 0.0092,
      "step": 10315
    },
    {
      "epoch": 156.36363636363637,
      "grad_norm": 0.047432832419872284,
      "learning_rate": 2.1818181818181818e-05,
      "loss": 0.0089,
      "step": 10320
    },
    {
      "epoch": 156.43939393939394,
      "grad_norm": 0.0626806691288948,
      "learning_rate": 2.178030303030303e-05,
      "loss": 0.0121,
      "step": 10325
    },
    {
      "epoch": 156.5151515151515,
      "grad_norm": 0.079598069190979,
      "learning_rate": 2.1742424242424244e-05,
      "loss": 0.0085,
      "step": 10330
    },
    {
      "epoch": 156.5909090909091,
      "grad_norm": 0.0388801284134388,
      "learning_rate": 2.1704545454545454e-05,
      "loss": 0.007,
      "step": 10335
    },
    {
      "epoch": 156.66666666666666,
      "grad_norm": 0.09725102037191391,
      "learning_rate": 2.1666666666666667e-05,
      "loss": 0.0116,
      "step": 10340
    },
    {
      "epoch": 156.74242424242425,
      "grad_norm": 0.029241764917969704,
      "learning_rate": 2.162878787878788e-05,
      "loss": 0.0105,
      "step": 10345
    },
    {
      "epoch": 156.8181818181818,
      "grad_norm": 0.03782626613974571,
      "learning_rate": 2.1590909090909093e-05,
      "loss": 0.0066,
      "step": 10350
    },
    {
      "epoch": 156.8939393939394,
      "grad_norm": 0.04052730277180672,
      "learning_rate": 2.1553030303030303e-05,
      "loss": 0.007,
      "step": 10355
    },
    {
      "epoch": 156.96969696969697,
      "grad_norm": 0.04589178040623665,
      "learning_rate": 2.1515151515151516e-05,
      "loss": 0.0083,
      "step": 10360
    },
    {
      "epoch": 157.0,
      "eval_loss": 0.011354141868650913,
      "eval_mean_accuracy": 0.9540401495253553,
      "eval_mean_iou": 0.9269613600321168,
      "eval_runtime": 37.5338,
      "eval_samples_per_second": 6.261,
      "eval_steps_per_second": 0.799,
      "step": 10362
    },
    {
      "epoch": 157.04545454545453,
      "grad_norm": 0.027843600139021873,
      "learning_rate": 2.147727272727273e-05,
      "loss": 0.0118,
      "step": 10365
    },
    {
      "epoch": 157.12121212121212,
      "grad_norm": 0.028804434463381767,
      "learning_rate": 2.143939393939394e-05,
      "loss": 0.0069,
      "step": 10370
    },
    {
      "epoch": 157.1969696969697,
      "grad_norm": 0.034283895045518875,
      "learning_rate": 2.1401515151515152e-05,
      "loss": 0.0092,
      "step": 10375
    },
    {
      "epoch": 157.27272727272728,
      "grad_norm": 0.018444275483489037,
      "learning_rate": 2.1363636363636362e-05,
      "loss": 0.0091,
      "step": 10380
    },
    {
      "epoch": 157.34848484848484,
      "grad_norm": 0.05327335745096207,
      "learning_rate": 2.1325757575757575e-05,
      "loss": 0.0089,
      "step": 10385
    },
    {
      "epoch": 157.42424242424244,
      "grad_norm": 0.05062629655003548,
      "learning_rate": 2.128787878787879e-05,
      "loss": 0.0078,
      "step": 10390
    },
    {
      "epoch": 157.5,
      "grad_norm": 0.03726862743496895,
      "learning_rate": 2.125e-05,
      "loss": 0.0089,
      "step": 10395
    },
    {
      "epoch": 157.57575757575756,
      "grad_norm": 0.08047214150428772,
      "learning_rate": 2.1212121212121215e-05,
      "loss": 0.0101,
      "step": 10400
    },
    {
      "epoch": 157.65151515151516,
      "grad_norm": 0.02738412655889988,
      "learning_rate": 2.1174242424242424e-05,
      "loss": 0.0076,
      "step": 10405
    },
    {
      "epoch": 157.72727272727272,
      "grad_norm": 0.08308125287294388,
      "learning_rate": 2.1136363636363638e-05,
      "loss": 0.0101,
      "step": 10410
    },
    {
      "epoch": 157.8030303030303,
      "grad_norm": 0.04785419628024101,
      "learning_rate": 2.109848484848485e-05,
      "loss": 0.0102,
      "step": 10415
    },
    {
      "epoch": 157.87878787878788,
      "grad_norm": 0.026348695158958435,
      "learning_rate": 2.106060606060606e-05,
      "loss": 0.0095,
      "step": 10420
    },
    {
      "epoch": 157.95454545454547,
      "grad_norm": 0.060041945427656174,
      "learning_rate": 2.1022727272727274e-05,
      "loss": 0.0094,
      "step": 10425
    },
    {
      "epoch": 158.0,
      "eval_loss": 0.011280659586191177,
      "eval_mean_accuracy": 0.95704198258835,
      "eval_mean_iou": 0.927645488069253,
      "eval_runtime": 35.9693,
      "eval_samples_per_second": 6.533,
      "eval_steps_per_second": 0.834,
      "step": 10428
    },
    {
      "epoch": 158.03030303030303,
      "grad_norm": 0.0795101597905159,
      "learning_rate": 2.0984848484848483e-05,
      "loss": 0.01,
      "step": 10430
    },
    {
      "epoch": 158.1060606060606,
      "grad_norm": 0.06699241697788239,
      "learning_rate": 2.0946969696969697e-05,
      "loss": 0.0083,
      "step": 10435
    },
    {
      "epoch": 158.1818181818182,
      "grad_norm": 0.0274868942797184,
      "learning_rate": 2.090909090909091e-05,
      "loss": 0.008,
      "step": 10440
    },
    {
      "epoch": 158.25757575757575,
      "grad_norm": 0.027851689606904984,
      "learning_rate": 2.0871212121212123e-05,
      "loss": 0.0086,
      "step": 10445
    },
    {
      "epoch": 158.33333333333334,
      "grad_norm": 0.07611910253763199,
      "learning_rate": 2.0833333333333336e-05,
      "loss": 0.0117,
      "step": 10450
    },
    {
      "epoch": 158.4090909090909,
      "grad_norm": 0.05738113075494766,
      "learning_rate": 2.0795454545454546e-05,
      "loss": 0.0077,
      "step": 10455
    },
    {
      "epoch": 158.4848484848485,
      "grad_norm": 0.02926352061331272,
      "learning_rate": 2.075757575757576e-05,
      "loss": 0.0107,
      "step": 10460
    },
    {
      "epoch": 158.56060606060606,
      "grad_norm": 0.023429399356245995,
      "learning_rate": 2.071969696969697e-05,
      "loss": 0.0113,
      "step": 10465
    },
    {
      "epoch": 158.63636363636363,
      "grad_norm": 0.06089494377374649,
      "learning_rate": 2.0681818181818182e-05,
      "loss": 0.0073,
      "step": 10470
    },
    {
      "epoch": 158.71212121212122,
      "grad_norm": 0.04005968198180199,
      "learning_rate": 2.0643939393939395e-05,
      "loss": 0.0083,
      "step": 10475
    },
    {
      "epoch": 158.78787878787878,
      "grad_norm": 0.030611805617809296,
      "learning_rate": 2.0606060606060608e-05,
      "loss": 0.008,
      "step": 10480
    },
    {
      "epoch": 158.86363636363637,
      "grad_norm": 0.025063475593924522,
      "learning_rate": 2.056818181818182e-05,
      "loss": 0.0079,
      "step": 10485
    },
    {
      "epoch": 158.93939393939394,
      "grad_norm": 0.07822442054748535,
      "learning_rate": 2.053030303030303e-05,
      "loss": 0.0083,
      "step": 10490
    },
    {
      "epoch": 159.0,
      "eval_loss": 0.011247639544308186,
      "eval_mean_accuracy": 0.9579039446651599,
      "eval_mean_iou": 0.9282750901393824,
      "eval_runtime": 35.5873,
      "eval_samples_per_second": 6.603,
      "eval_steps_per_second": 0.843,
      "step": 10494
    },
    {
      "epoch": 159.0151515151515,
      "grad_norm": 0.05568947270512581,
      "learning_rate": 2.0492424242424244e-05,
      "loss": 0.0173,
      "step": 10495
    },
    {
      "epoch": 159.0909090909091,
      "grad_norm": 0.0869707465171814,
      "learning_rate": 2.0454545454545457e-05,
      "loss": 0.008,
      "step": 10500
    },
    {
      "epoch": 159.16666666666666,
      "grad_norm": 0.02542329765856266,
      "learning_rate": 2.0416666666666667e-05,
      "loss": 0.0106,
      "step": 10505
    },
    {
      "epoch": 159.24242424242425,
      "grad_norm": 0.04913695529103279,
      "learning_rate": 2.037878787878788e-05,
      "loss": 0.0104,
      "step": 10510
    },
    {
      "epoch": 159.3181818181818,
      "grad_norm": 0.1075977012515068,
      "learning_rate": 2.034090909090909e-05,
      "loss": 0.0117,
      "step": 10515
    },
    {
      "epoch": 159.3939393939394,
      "grad_norm": 0.06269510835409164,
      "learning_rate": 2.0303030303030303e-05,
      "loss": 0.0079,
      "step": 10520
    },
    {
      "epoch": 159.46969696969697,
      "grad_norm": 0.06338796019554138,
      "learning_rate": 2.0265151515151516e-05,
      "loss": 0.0094,
      "step": 10525
    },
    {
      "epoch": 159.54545454545453,
      "grad_norm": 0.03351595252752304,
      "learning_rate": 2.022727272727273e-05,
      "loss": 0.008,
      "step": 10530
    },
    {
      "epoch": 159.62121212121212,
      "grad_norm": 0.044570211321115494,
      "learning_rate": 2.0189393939393943e-05,
      "loss": 0.0077,
      "step": 10535
    },
    {
      "epoch": 159.6969696969697,
      "grad_norm": 0.08738235384225845,
      "learning_rate": 2.0151515151515152e-05,
      "loss": 0.0104,
      "step": 10540
    },
    {
      "epoch": 159.77272727272728,
      "grad_norm": 0.023128705099225044,
      "learning_rate": 2.0113636363636366e-05,
      "loss": 0.0089,
      "step": 10545
    },
    {
      "epoch": 159.84848484848484,
      "grad_norm": 0.04126572608947754,
      "learning_rate": 2.0075757575757575e-05,
      "loss": 0.0071,
      "step": 10550
    },
    {
      "epoch": 159.92424242424244,
      "grad_norm": 0.043159693479537964,
      "learning_rate": 2.003787878787879e-05,
      "loss": 0.0116,
      "step": 10555
    },
    {
      "epoch": 160.0,
      "grad_norm": 0.02562563307583332,
      "learning_rate": 2e-05,
      "loss": 0.0092,
      "step": 10560
    },
    {
      "epoch": 160.0,
      "eval_loss": 0.011324673891067505,
      "eval_mean_accuracy": 0.9544352474698604,
      "eval_mean_iou": 0.9266360237457131,
      "eval_runtime": 35.3299,
      "eval_samples_per_second": 6.652,
      "eval_steps_per_second": 0.849,
      "step": 10560
    },
    {
      "epoch": 160.07575757575756,
      "grad_norm": 0.06445985287427902,
      "learning_rate": 1.996212121212121e-05,
      "loss": 0.0094,
      "step": 10565
    },
    {
      "epoch": 160.15151515151516,
      "grad_norm": 0.027885334566235542,
      "learning_rate": 1.9924242424242425e-05,
      "loss": 0.0101,
      "step": 10570
    },
    {
      "epoch": 160.22727272727272,
      "grad_norm": 0.04759266600012779,
      "learning_rate": 1.9886363636363638e-05,
      "loss": 0.0105,
      "step": 10575
    },
    {
      "epoch": 160.3030303030303,
      "grad_norm": 0.0749398022890091,
      "learning_rate": 1.984848484848485e-05,
      "loss": 0.0092,
      "step": 10580
    },
    {
      "epoch": 160.37878787878788,
      "grad_norm": 0.020334230735898018,
      "learning_rate": 1.9810606060606064e-05,
      "loss": 0.0075,
      "step": 10585
    },
    {
      "epoch": 160.45454545454547,
      "grad_norm": 0.027339350432157516,
      "learning_rate": 1.9772727272727274e-05,
      "loss": 0.0069,
      "step": 10590
    },
    {
      "epoch": 160.53030303030303,
      "grad_norm": 0.031346794217824936,
      "learning_rate": 1.9734848484848487e-05,
      "loss": 0.0101,
      "step": 10595
    },
    {
      "epoch": 160.6060606060606,
      "grad_norm": 0.1405462920665741,
      "learning_rate": 1.9696969696969697e-05,
      "loss": 0.0102,
      "step": 10600
    },
    {
      "epoch": 160.6818181818182,
      "grad_norm": 0.041002895683050156,
      "learning_rate": 1.965909090909091e-05,
      "loss": 0.0082,
      "step": 10605
    },
    {
      "epoch": 160.75757575757575,
      "grad_norm": 0.03083837777376175,
      "learning_rate": 1.962121212121212e-05,
      "loss": 0.0079,
      "step": 10610
    },
    {
      "epoch": 160.83333333333334,
      "grad_norm": 0.18179790675640106,
      "learning_rate": 1.9583333333333333e-05,
      "loss": 0.0115,
      "step": 10615
    },
    {
      "epoch": 160.9090909090909,
      "grad_norm": 0.014926694333553314,
      "learning_rate": 1.9545454545454546e-05,
      "loss": 0.0079,
      "step": 10620
    },
    {
      "epoch": 160.9848484848485,
      "grad_norm": 0.08189494162797928,
      "learning_rate": 1.950757575757576e-05,
      "loss": 0.0102,
      "step": 10625
    },
    {
      "epoch": 161.0,
      "eval_loss": 0.011284000240266323,
      "eval_mean_accuracy": 0.9559465856291709,
      "eval_mean_iou": 0.9277009684926304,
      "eval_runtime": 36.024,
      "eval_samples_per_second": 6.523,
      "eval_steps_per_second": 0.833,
      "step": 10626
    },
    {
      "epoch": 161.06060606060606,
      "grad_norm": 0.038384877145290375,
      "learning_rate": 1.9469696969696972e-05,
      "loss": 0.0083,
      "step": 10630
    },
    {
      "epoch": 161.13636363636363,
      "grad_norm": 0.018902190029621124,
      "learning_rate": 1.9431818181818182e-05,
      "loss": 0.0092,
      "step": 10635
    },
    {
      "epoch": 161.21212121212122,
      "grad_norm": 0.048388756811618805,
      "learning_rate": 1.9393939393939395e-05,
      "loss": 0.0092,
      "step": 10640
    },
    {
      "epoch": 161.28787878787878,
      "grad_norm": 0.029563002288341522,
      "learning_rate": 1.9356060606060608e-05,
      "loss": 0.0085,
      "step": 10645
    },
    {
      "epoch": 161.36363636363637,
      "grad_norm": 0.04581021890044212,
      "learning_rate": 1.9318181818181818e-05,
      "loss": 0.0114,
      "step": 10650
    },
    {
      "epoch": 161.43939393939394,
      "grad_norm": 0.07129979133605957,
      "learning_rate": 1.928030303030303e-05,
      "loss": 0.0097,
      "step": 10655
    },
    {
      "epoch": 161.5151515151515,
      "grad_norm": 0.06146996095776558,
      "learning_rate": 1.924242424242424e-05,
      "loss": 0.0077,
      "step": 10660
    },
    {
      "epoch": 161.5909090909091,
      "grad_norm": 0.03913751244544983,
      "learning_rate": 1.9204545454545454e-05,
      "loss": 0.0086,
      "step": 10665
    },
    {
      "epoch": 161.66666666666666,
      "grad_norm": 0.03823217377066612,
      "learning_rate": 1.9166666666666667e-05,
      "loss": 0.0106,
      "step": 10670
    },
    {
      "epoch": 161.74242424242425,
      "grad_norm": 0.024882428348064423,
      "learning_rate": 1.912878787878788e-05,
      "loss": 0.0077,
      "step": 10675
    },
    {
      "epoch": 161.8181818181818,
      "grad_norm": 0.022445356473326683,
      "learning_rate": 1.9090909090909094e-05,
      "loss": 0.0075,
      "step": 10680
    },
    {
      "epoch": 161.8939393939394,
      "grad_norm": 0.03439823165535927,
      "learning_rate": 1.9053030303030303e-05,
      "loss": 0.0105,
      "step": 10685
    },
    {
      "epoch": 161.96969696969697,
      "grad_norm": 0.02401348389685154,
      "learning_rate": 1.9015151515151516e-05,
      "loss": 0.0089,
      "step": 10690
    },
    {
      "epoch": 162.0,
      "eval_loss": 0.011278776451945305,
      "eval_mean_accuracy": 0.954323995407559,
      "eval_mean_iou": 0.9273379984950426,
      "eval_runtime": 37.8714,
      "eval_samples_per_second": 6.205,
      "eval_steps_per_second": 0.792,
      "step": 10692
    },
    {
      "epoch": 162.04545454545453,
      "grad_norm": 0.0607118122279644,
      "learning_rate": 1.8977272727272726e-05,
      "loss": 0.0086,
      "step": 10695
    },
    {
      "epoch": 162.12121212121212,
      "grad_norm": 0.030251163989305496,
      "learning_rate": 1.893939393939394e-05,
      "loss": 0.0086,
      "step": 10700
    },
    {
      "epoch": 162.1969696969697,
      "grad_norm": 0.01807577535510063,
      "learning_rate": 1.8901515151515153e-05,
      "loss": 0.0079,
      "step": 10705
    },
    {
      "epoch": 162.27272727272728,
      "grad_norm": 0.04539410397410393,
      "learning_rate": 1.8863636363636362e-05,
      "loss": 0.0111,
      "step": 10710
    },
    {
      "epoch": 162.34848484848484,
      "grad_norm": 0.09482821822166443,
      "learning_rate": 1.882575757575758e-05,
      "loss": 0.0097,
      "step": 10715
    },
    {
      "epoch": 162.42424242424244,
      "grad_norm": 0.04523533955216408,
      "learning_rate": 1.878787878787879e-05,
      "loss": 0.0087,
      "step": 10720
    },
    {
      "epoch": 162.5,
      "grad_norm": 0.03613491728901863,
      "learning_rate": 1.8750000000000002e-05,
      "loss": 0.0087,
      "step": 10725
    },
    {
      "epoch": 162.57575757575756,
      "grad_norm": 0.053956788033246994,
      "learning_rate": 1.8712121212121215e-05,
      "loss": 0.0072,
      "step": 10730
    },
    {
      "epoch": 162.65151515151516,
      "grad_norm": 0.0297955721616745,
      "learning_rate": 1.8674242424242425e-05,
      "loss": 0.0095,
      "step": 10735
    },
    {
      "epoch": 162.72727272727272,
      "grad_norm": 0.04248766601085663,
      "learning_rate": 1.8636363636363638e-05,
      "loss": 0.0106,
      "step": 10740
    },
    {
      "epoch": 162.8030303030303,
      "grad_norm": 0.07496894150972366,
      "learning_rate": 1.8598484848484848e-05,
      "loss": 0.0101,
      "step": 10745
    },
    {
      "epoch": 162.87878787878788,
      "grad_norm": 0.047424256801605225,
      "learning_rate": 1.856060606060606e-05,
      "loss": 0.0089,
      "step": 10750
    },
    {
      "epoch": 162.95454545454547,
      "grad_norm": 0.04810931533575058,
      "learning_rate": 1.8522727272727274e-05,
      "loss": 0.0086,
      "step": 10755
    },
    {
      "epoch": 163.0,
      "eval_loss": 0.011151275597512722,
      "eval_mean_accuracy": 0.9557917451663948,
      "eval_mean_iou": 0.9281644297531706,
      "eval_runtime": 37.1805,
      "eval_samples_per_second": 6.321,
      "eval_steps_per_second": 0.807,
      "step": 10758
    },
    {
      "epoch": 163.03030303030303,
      "grad_norm": 0.0387769378721714,
      "learning_rate": 1.8484848484848487e-05,
      "loss": 0.0061,
      "step": 10760
    },
    {
      "epoch": 163.1060606060606,
      "grad_norm": 0.052900392562150955,
      "learning_rate": 1.84469696969697e-05,
      "loss": 0.0096,
      "step": 10765
    },
    {
      "epoch": 163.1818181818182,
      "grad_norm": 0.07133764028549194,
      "learning_rate": 1.840909090909091e-05,
      "loss": 0.01,
      "step": 10770
    },
    {
      "epoch": 163.25757575757575,
      "grad_norm": 0.054321520030498505,
      "learning_rate": 1.8371212121212123e-05,
      "loss": 0.0095,
      "step": 10775
    },
    {
      "epoch": 163.33333333333334,
      "grad_norm": 0.02339097484946251,
      "learning_rate": 1.8333333333333333e-05,
      "loss": 0.0119,
      "step": 10780
    },
    {
      "epoch": 163.4090909090909,
      "grad_norm": 0.030062464997172356,
      "learning_rate": 1.8295454545454546e-05,
      "loss": 0.0085,
      "step": 10785
    },
    {
      "epoch": 163.4848484848485,
      "grad_norm": 0.03471481427550316,
      "learning_rate": 1.825757575757576e-05,
      "loss": 0.007,
      "step": 10790
    },
    {
      "epoch": 163.56060606060606,
      "grad_norm": 0.016919812187552452,
      "learning_rate": 1.821969696969697e-05,
      "loss": 0.0086,
      "step": 10795
    },
    {
      "epoch": 163.63636363636363,
      "grad_norm": 0.053019456565380096,
      "learning_rate": 1.8181818181818182e-05,
      "loss": 0.0106,
      "step": 10800
    },
    {
      "epoch": 163.71212121212122,
      "grad_norm": 0.04763426259160042,
      "learning_rate": 1.8143939393939395e-05,
      "loss": 0.0069,
      "step": 10805
    },
    {
      "epoch": 163.78787878787878,
      "grad_norm": 0.03805103525519371,
      "learning_rate": 1.810606060606061e-05,
      "loss": 0.0101,
      "step": 10810
    },
    {
      "epoch": 163.86363636363637,
      "grad_norm": 0.02324060909450054,
      "learning_rate": 1.806818181818182e-05,
      "loss": 0.0096,
      "step": 10815
    },
    {
      "epoch": 163.93939393939394,
      "grad_norm": 0.01340368390083313,
      "learning_rate": 1.803030303030303e-05,
      "loss": 0.0076,
      "step": 10820
    },
    {
      "epoch": 164.0,
      "eval_loss": 0.01131797581911087,
      "eval_mean_accuracy": 0.9570391496444034,
      "eval_mean_iou": 0.9278879961639578,
      "eval_runtime": 37.4835,
      "eval_samples_per_second": 6.269,
      "eval_steps_per_second": 0.8,
      "step": 10824
    },
    {
      "epoch": 164.0151515151515,
      "grad_norm": 0.04041508957743645,
      "learning_rate": 1.7992424242424244e-05,
      "loss": 0.0086,
      "step": 10825
    },
    {
      "epoch": 164.0909090909091,
      "grad_norm": 0.049855444580316544,
      "learning_rate": 1.7954545454545454e-05,
      "loss": 0.0107,
      "step": 10830
    },
    {
      "epoch": 164.16666666666666,
      "grad_norm": 0.06599489599466324,
      "learning_rate": 1.7916666666666667e-05,
      "loss": 0.0093,
      "step": 10835
    },
    {
      "epoch": 164.24242424242425,
      "grad_norm": 0.04243709146976471,
      "learning_rate": 1.787878787878788e-05,
      "loss": 0.0089,
      "step": 10840
    },
    {
      "epoch": 164.3181818181818,
      "grad_norm": 0.09043898433446884,
      "learning_rate": 1.784090909090909e-05,
      "loss": 0.0106,
      "step": 10845
    },
    {
      "epoch": 164.3939393939394,
      "grad_norm": 0.028482165187597275,
      "learning_rate": 1.7803030303030303e-05,
      "loss": 0.0074,
      "step": 10850
    },
    {
      "epoch": 164.46969696969697,
      "grad_norm": 0.05633034184575081,
      "learning_rate": 1.7765151515151517e-05,
      "loss": 0.0094,
      "step": 10855
    },
    {
      "epoch": 164.54545454545453,
      "grad_norm": 0.03428157418966293,
      "learning_rate": 1.772727272727273e-05,
      "loss": 0.0093,
      "step": 10860
    },
    {
      "epoch": 164.62121212121212,
      "grad_norm": 0.041122227907180786,
      "learning_rate": 1.768939393939394e-05,
      "loss": 0.0115,
      "step": 10865
    },
    {
      "epoch": 164.6969696969697,
      "grad_norm": 0.02417839504778385,
      "learning_rate": 1.7651515151515153e-05,
      "loss": 0.0082,
      "step": 10870
    },
    {
      "epoch": 164.77272727272728,
      "grad_norm": 0.023925330489873886,
      "learning_rate": 1.7613636363636366e-05,
      "loss": 0.007,
      "step": 10875
    },
    {
      "epoch": 164.84848484848484,
      "grad_norm": 0.02939862385392189,
      "learning_rate": 1.7575757575757576e-05,
      "loss": 0.0093,
      "step": 10880
    },
    {
      "epoch": 164.92424242424244,
      "grad_norm": 0.022345630452036858,
      "learning_rate": 1.753787878787879e-05,
      "loss": 0.0079,
      "step": 10885
    },
    {
      "epoch": 165.0,
      "grad_norm": 0.052027229219675064,
      "learning_rate": 1.75e-05,
      "loss": 0.0089,
      "step": 10890
    },
    {
      "epoch": 165.0,
      "eval_loss": 0.011224967427551746,
      "eval_mean_accuracy": 0.9583174247734384,
      "eval_mean_iou": 0.9283040965490467,
      "eval_runtime": 38.7645,
      "eval_samples_per_second": 6.062,
      "eval_steps_per_second": 0.774,
      "step": 10890
    },
    {
      "epoch": 165.07575757575756,
      "grad_norm": 0.02571563422679901,
      "learning_rate": 1.746212121212121e-05,
      "loss": 0.0069,
      "step": 10895
    },
    {
      "epoch": 165.15151515151516,
      "grad_norm": 0.023892438039183617,
      "learning_rate": 1.7424242424242425e-05,
      "loss": 0.0089,
      "step": 10900
    },
    {
      "epoch": 165.22727272727272,
      "grad_norm": 0.05152574181556702,
      "learning_rate": 1.7386363636363638e-05,
      "loss": 0.0076,
      "step": 10905
    },
    {
      "epoch": 165.3030303030303,
      "grad_norm": 0.04756501689553261,
      "learning_rate": 1.734848484848485e-05,
      "loss": 0.0114,
      "step": 10910
    },
    {
      "epoch": 165.37878787878788,
      "grad_norm": 0.06286393851041794,
      "learning_rate": 1.731060606060606e-05,
      "loss": 0.0097,
      "step": 10915
    },
    {
      "epoch": 165.45454545454547,
      "grad_norm": 0.030574548989534378,
      "learning_rate": 1.7272727272727274e-05,
      "loss": 0.0071,
      "step": 10920
    },
    {
      "epoch": 165.53030303030303,
      "grad_norm": 0.017590459436178207,
      "learning_rate": 1.7234848484848487e-05,
      "loss": 0.0084,
      "step": 10925
    },
    {
      "epoch": 165.6060606060606,
      "grad_norm": 0.018039289861917496,
      "learning_rate": 1.7196969696969697e-05,
      "loss": 0.0093,
      "step": 10930
    },
    {
      "epoch": 165.6818181818182,
      "grad_norm": 0.0484914630651474,
      "learning_rate": 1.715909090909091e-05,
      "loss": 0.0113,
      "step": 10935
    },
    {
      "epoch": 165.75757575757575,
      "grad_norm": 0.03142247721552849,
      "learning_rate": 1.712121212121212e-05,
      "loss": 0.0059,
      "step": 10940
    },
    {
      "epoch": 165.83333333333334,
      "grad_norm": 0.027634793892502785,
      "learning_rate": 1.7083333333333333e-05,
      "loss": 0.0103,
      "step": 10945
    },
    {
      "epoch": 165.9090909090909,
      "grad_norm": 0.07303141057491302,
      "learning_rate": 1.7045454545454546e-05,
      "loss": 0.0122,
      "step": 10950
    },
    {
      "epoch": 165.9848484848485,
      "grad_norm": 0.021265963092446327,
      "learning_rate": 1.700757575757576e-05,
      "loss": 0.0084,
      "step": 10955
    },
    {
      "epoch": 166.0,
      "eval_loss": 0.011246977373957634,
      "eval_mean_accuracy": 0.9571884571592087,
      "eval_mean_iou": 0.9280327994356395,
      "eval_runtime": 34.7324,
      "eval_samples_per_second": 6.766,
      "eval_steps_per_second": 0.864,
      "step": 10956
    },
    {
      "epoch": 166.06060606060606,
      "grad_norm": 0.03065045364201069,
      "learning_rate": 1.6969696969696972e-05,
      "loss": 0.0082,
      "step": 10960
    },
    {
      "epoch": 166.13636363636363,
      "grad_norm": 0.0379100926220417,
      "learning_rate": 1.6931818181818182e-05,
      "loss": 0.0101,
      "step": 10965
    },
    {
      "epoch": 166.21212121212122,
      "grad_norm": 0.03359740599989891,
      "learning_rate": 1.6893939393939395e-05,
      "loss": 0.0082,
      "step": 10970
    },
    {
      "epoch": 166.28787878787878,
      "grad_norm": 0.03783692419528961,
      "learning_rate": 1.6856060606060605e-05,
      "loss": 0.0112,
      "step": 10975
    },
    {
      "epoch": 166.36363636363637,
      "grad_norm": 0.038849737495183945,
      "learning_rate": 1.6818181818181818e-05,
      "loss": 0.0108,
      "step": 10980
    },
    {
      "epoch": 166.43939393939394,
      "grad_norm": 0.042607422918081284,
      "learning_rate": 1.678030303030303e-05,
      "loss": 0.0071,
      "step": 10985
    },
    {
      "epoch": 166.5151515151515,
      "grad_norm": 0.056289996951818466,
      "learning_rate": 1.674242424242424e-05,
      "loss": 0.011,
      "step": 10990
    },
    {
      "epoch": 166.5909090909091,
      "grad_norm": 0.021373853087425232,
      "learning_rate": 1.6704545454545458e-05,
      "loss": 0.0091,
      "step": 10995
    },
    {
      "epoch": 166.66666666666666,
      "grad_norm": 0.03538452461361885,
      "learning_rate": 1.6666666666666667e-05,
      "loss": 0.0092,
      "step": 11000
    },
    {
      "epoch": 166.74242424242425,
      "grad_norm": 0.07104497402906418,
      "learning_rate": 1.662878787878788e-05,
      "loss": 0.0078,
      "step": 11005
    },
    {
      "epoch": 166.8181818181818,
      "grad_norm": 0.03861116245388985,
      "learning_rate": 1.6590909090909094e-05,
      "loss": 0.0065,
      "step": 11010
    },
    {
      "epoch": 166.8939393939394,
      "grad_norm": 0.030134856700897217,
      "learning_rate": 1.6553030303030304e-05,
      "loss": 0.0108,
      "step": 11015
    },
    {
      "epoch": 166.96969696969697,
      "grad_norm": 0.03871239349246025,
      "learning_rate": 1.6515151515151517e-05,
      "loss": 0.0088,
      "step": 11020
    },
    {
      "epoch": 167.0,
      "eval_loss": 0.011325166560709476,
      "eval_mean_accuracy": 0.9567643358482604,
      "eval_mean_iou": 0.9274264637320916,
      "eval_runtime": 37.4538,
      "eval_samples_per_second": 6.274,
      "eval_steps_per_second": 0.801,
      "step": 11022
    },
    {
      "epoch": 167.04545454545453,
      "grad_norm": 0.03437385335564613,
      "learning_rate": 1.6477272727272726e-05,
      "loss": 0.006,
      "step": 11025
    },
    {
      "epoch": 167.12121212121212,
      "grad_norm": 0.019675759598612785,
      "learning_rate": 1.643939393939394e-05,
      "loss": 0.0088,
      "step": 11030
    },
    {
      "epoch": 167.1969696969697,
      "grad_norm": 0.05714665725827217,
      "learning_rate": 1.6401515151515153e-05,
      "loss": 0.0094,
      "step": 11035
    },
    {
      "epoch": 167.27272727272728,
      "grad_norm": 0.02804950624704361,
      "learning_rate": 1.6363636363636366e-05,
      "loss": 0.0078,
      "step": 11040
    },
    {
      "epoch": 167.34848484848484,
      "grad_norm": 0.04505068436264992,
      "learning_rate": 1.632575757575758e-05,
      "loss": 0.0084,
      "step": 11045
    },
    {
      "epoch": 167.42424242424244,
      "grad_norm": 0.05408765375614166,
      "learning_rate": 1.628787878787879e-05,
      "loss": 0.0083,
      "step": 11050
    },
    {
      "epoch": 167.5,
      "grad_norm": 0.04443540796637535,
      "learning_rate": 1.6250000000000002e-05,
      "loss": 0.0118,
      "step": 11055
    },
    {
      "epoch": 167.57575757575756,
      "grad_norm": 0.06730034202337265,
      "learning_rate": 1.6212121212121212e-05,
      "loss": 0.0084,
      "step": 11060
    },
    {
      "epoch": 167.65151515151516,
      "grad_norm": 0.03343914449214935,
      "learning_rate": 1.6174242424242425e-05,
      "loss": 0.0081,
      "step": 11065
    },
    {
      "epoch": 167.72727272727272,
      "grad_norm": 0.0924658328294754,
      "learning_rate": 1.6136363636363638e-05,
      "loss": 0.0085,
      "step": 11070
    },
    {
      "epoch": 167.8030303030303,
      "grad_norm": 0.017449328675866127,
      "learning_rate": 1.6098484848484848e-05,
      "loss": 0.0112,
      "step": 11075
    },
    {
      "epoch": 167.87878787878788,
      "grad_norm": 0.02669735625386238,
      "learning_rate": 1.606060606060606e-05,
      "loss": 0.0103,
      "step": 11080
    },
    {
      "epoch": 167.95454545454547,
      "grad_norm": 0.038670070469379425,
      "learning_rate": 1.6022727272727274e-05,
      "loss": 0.0071,
      "step": 11085
    },
    {
      "epoch": 168.0,
      "eval_loss": 0.011157882399857044,
      "eval_mean_accuracy": 0.9562893107924199,
      "eval_mean_iou": 0.9285062320778114,
      "eval_runtime": 37.9708,
      "eval_samples_per_second": 6.189,
      "eval_steps_per_second": 0.79,
      "step": 11088
    },
    {
      "epoch": 168.03030303030303,
      "grad_norm": 0.02428271248936653,
      "learning_rate": 1.5984848484848487e-05,
      "loss": 0.0094,
      "step": 11090
    },
    {
      "epoch": 168.1060606060606,
      "grad_norm": 0.047002509236335754,
      "learning_rate": 1.59469696969697e-05,
      "loss": 0.0079,
      "step": 11095
    },
    {
      "epoch": 168.1818181818182,
      "grad_norm": 0.05363202467560768,
      "learning_rate": 1.590909090909091e-05,
      "loss": 0.0088,
      "step": 11100
    },
    {
      "epoch": 168.25757575757575,
      "grad_norm": 0.032988596707582474,
      "learning_rate": 1.5871212121212123e-05,
      "loss": 0.0098,
      "step": 11105
    },
    {
      "epoch": 168.33333333333334,
      "grad_norm": 0.06102301552891731,
      "learning_rate": 1.5833333333333333e-05,
      "loss": 0.0098,
      "step": 11110
    },
    {
      "epoch": 168.4090909090909,
      "grad_norm": 0.1341535747051239,
      "learning_rate": 1.5795454545454546e-05,
      "loss": 0.0101,
      "step": 11115
    },
    {
      "epoch": 168.4848484848485,
      "grad_norm": 0.03458147868514061,
      "learning_rate": 1.5757575757575756e-05,
      "loss": 0.0092,
      "step": 11120
    },
    {
      "epoch": 168.56060606060606,
      "grad_norm": 0.06516727060079575,
      "learning_rate": 1.571969696969697e-05,
      "loss": 0.0074,
      "step": 11125
    },
    {
      "epoch": 168.63636363636363,
      "grad_norm": 0.06574384868144989,
      "learning_rate": 1.5681818181818182e-05,
      "loss": 0.0122,
      "step": 11130
    },
    {
      "epoch": 168.71212121212122,
      "grad_norm": 0.075789675116539,
      "learning_rate": 1.5643939393939395e-05,
      "loss": 0.0098,
      "step": 11135
    },
    {
      "epoch": 168.78787878787878,
      "grad_norm": 0.03797443211078644,
      "learning_rate": 1.560606060606061e-05,
      "loss": 0.0088,
      "step": 11140
    },
    {
      "epoch": 168.86363636363637,
      "grad_norm": 0.027371611446142197,
      "learning_rate": 1.556818181818182e-05,
      "loss": 0.0095,
      "step": 11145
    },
    {
      "epoch": 168.93939393939394,
      "grad_norm": 0.05595516413450241,
      "learning_rate": 1.553030303030303e-05,
      "loss": 0.01,
      "step": 11150
    },
    {
      "epoch": 169.0,
      "eval_loss": 0.011500139720737934,
      "eval_mean_accuracy": 0.9509908758899823,
      "eval_mean_iou": 0.9250559968106625,
      "eval_runtime": 38.2542,
      "eval_samples_per_second": 6.143,
      "eval_steps_per_second": 0.784,
      "step": 11154
    },
    {
      "epoch": 169.0151515151515,
      "grad_norm": 0.12889814376831055,
      "learning_rate": 1.5492424242424245e-05,
      "loss": 0.0116,
      "step": 11155
    },
    {
      "epoch": 169.0909090909091,
      "grad_norm": 0.032219287008047104,
      "learning_rate": 1.5454545454545454e-05,
      "loss": 0.0103,
      "step": 11160
    },
    {
      "epoch": 169.16666666666666,
      "grad_norm": 0.02340775728225708,
      "learning_rate": 1.5416666666666668e-05,
      "loss": 0.0102,
      "step": 11165
    },
    {
      "epoch": 169.24242424242425,
      "grad_norm": 0.031015045940876007,
      "learning_rate": 1.5378787878787877e-05,
      "loss": 0.0069,
      "step": 11170
    },
    {
      "epoch": 169.3181818181818,
      "grad_norm": 0.02201847732067108,
      "learning_rate": 1.534090909090909e-05,
      "loss": 0.0075,
      "step": 11175
    },
    {
      "epoch": 169.3939393939394,
      "grad_norm": 0.01951393112540245,
      "learning_rate": 1.5303030303030304e-05,
      "loss": 0.008,
      "step": 11180
    },
    {
      "epoch": 169.46969696969697,
      "grad_norm": 0.07388915866613388,
      "learning_rate": 1.5265151515151517e-05,
      "loss": 0.0109,
      "step": 11185
    },
    {
      "epoch": 169.54545454545453,
      "grad_norm": 0.031158551573753357,
      "learning_rate": 1.5227272727272728e-05,
      "loss": 0.0068,
      "step": 11190
    },
    {
      "epoch": 169.62121212121212,
      "grad_norm": 0.061049774289131165,
      "learning_rate": 1.518939393939394e-05,
      "loss": 0.0085,
      "step": 11195
    },
    {
      "epoch": 169.6969696969697,
      "grad_norm": 0.037523239850997925,
      "learning_rate": 1.5151515151515153e-05,
      "loss": 0.0092,
      "step": 11200
    },
    {
      "epoch": 169.77272727272728,
      "grad_norm": 0.04240166023373604,
      "learning_rate": 1.5113636363636363e-05,
      "loss": 0.0092,
      "step": 11205
    },
    {
      "epoch": 169.84848484848484,
      "grad_norm": 0.027436353266239166,
      "learning_rate": 1.5075757575757576e-05,
      "loss": 0.0074,
      "step": 11210
    },
    {
      "epoch": 169.92424242424244,
      "grad_norm": 0.034957773983478546,
      "learning_rate": 1.5037878787878789e-05,
      "loss": 0.0086,
      "step": 11215
    },
    {
      "epoch": 170.0,
      "grad_norm": 0.11183241009712219,
      "learning_rate": 1.5e-05,
      "loss": 0.0095,
      "step": 11220
    },
    {
      "epoch": 170.0,
      "eval_loss": 0.011326088570058346,
      "eval_mean_accuracy": 0.955543967492027,
      "eval_mean_iou": 0.9273207177426984,
      "eval_runtime": 38.1086,
      "eval_samples_per_second": 6.167,
      "eval_steps_per_second": 0.787,
      "step": 11220
    },
    {
      "epoch": 170.07575757575756,
      "grad_norm": 0.04534455016255379,
      "learning_rate": 1.4962121212121214e-05,
      "loss": 0.0078,
      "step": 11225
    },
    {
      "epoch": 170.15151515151516,
      "grad_norm": 0.03492847457528114,
      "learning_rate": 1.4924242424242423e-05,
      "loss": 0.0091,
      "step": 11230
    },
    {
      "epoch": 170.22727272727272,
      "grad_norm": 0.03179406374692917,
      "learning_rate": 1.4886363636363636e-05,
      "loss": 0.0086,
      "step": 11235
    },
    {
      "epoch": 170.3030303030303,
      "grad_norm": 0.033982716500759125,
      "learning_rate": 1.484848484848485e-05,
      "loss": 0.008,
      "step": 11240
    },
    {
      "epoch": 170.37878787878788,
      "grad_norm": 0.05597720667719841,
      "learning_rate": 1.4810606060606061e-05,
      "loss": 0.0103,
      "step": 11245
    },
    {
      "epoch": 170.45454545454547,
      "grad_norm": 0.02094489149749279,
      "learning_rate": 1.4772727272727274e-05,
      "loss": 0.0075,
      "step": 11250
    },
    {
      "epoch": 170.53030303030303,
      "grad_norm": 0.04556063190102577,
      "learning_rate": 1.4734848484848484e-05,
      "loss": 0.0083,
      "step": 11255
    },
    {
      "epoch": 170.6060606060606,
      "grad_norm": 0.06934504210948944,
      "learning_rate": 1.4696969696969697e-05,
      "loss": 0.0109,
      "step": 11260
    },
    {
      "epoch": 170.6818181818182,
      "grad_norm": 0.01845012977719307,
      "learning_rate": 1.4659090909090909e-05,
      "loss": 0.008,
      "step": 11265
    },
    {
      "epoch": 170.75757575757575,
      "grad_norm": 0.04757295548915863,
      "learning_rate": 1.4621212121212122e-05,
      "loss": 0.008,
      "step": 11270
    },
    {
      "epoch": 170.83333333333334,
      "grad_norm": 0.03461065888404846,
      "learning_rate": 1.4583333333333335e-05,
      "loss": 0.0081,
      "step": 11275
    },
    {
      "epoch": 170.9090909090909,
      "grad_norm": 0.04684995487332344,
      "learning_rate": 1.4545454545454545e-05,
      "loss": 0.012,
      "step": 11280
    },
    {
      "epoch": 170.9848484848485,
      "grad_norm": 0.038422126322984695,
      "learning_rate": 1.450757575757576e-05,
      "loss": 0.0094,
      "step": 11285
    },
    {
      "epoch": 171.0,
      "eval_loss": 0.011345348320901394,
      "eval_mean_accuracy": 0.9541298565585289,
      "eval_mean_iou": 0.92635607441308,
      "eval_runtime": 35.9794,
      "eval_samples_per_second": 6.532,
      "eval_steps_per_second": 0.834,
      "step": 11286
    },
    {
      "epoch": 171.06060606060606,
      "grad_norm": 0.04366075620055199,
      "learning_rate": 1.446969696969697e-05,
      "loss": 0.0092,
      "step": 11290
    },
    {
      "epoch": 171.13636363636363,
      "grad_norm": 0.04008825868368149,
      "learning_rate": 1.4431818181818182e-05,
      "loss": 0.0111,
      "step": 11295
    },
    {
      "epoch": 171.21212121212122,
      "grad_norm": 0.039515431970357895,
      "learning_rate": 1.4393939393939396e-05,
      "loss": 0.0071,
      "step": 11300
    },
    {
      "epoch": 171.28787878787878,
      "grad_norm": 0.05157947167754173,
      "learning_rate": 1.4356060606060607e-05,
      "loss": 0.0084,
      "step": 11305
    },
    {
      "epoch": 171.36363636363637,
      "grad_norm": 0.028227059170603752,
      "learning_rate": 1.431818181818182e-05,
      "loss": 0.0084,
      "step": 11310
    },
    {
      "epoch": 171.43939393939394,
      "grad_norm": 0.04063564911484718,
      "learning_rate": 1.428030303030303e-05,
      "loss": 0.0078,
      "step": 11315
    },
    {
      "epoch": 171.5151515151515,
      "grad_norm": 0.026288973167538643,
      "learning_rate": 1.4242424242424243e-05,
      "loss": 0.007,
      "step": 11320
    },
    {
      "epoch": 171.5909090909091,
      "grad_norm": 0.02598394826054573,
      "learning_rate": 1.4204545454545456e-05,
      "loss": 0.0103,
      "step": 11325
    },
    {
      "epoch": 171.66666666666666,
      "grad_norm": 0.049306727945804596,
      "learning_rate": 1.4166666666666668e-05,
      "loss": 0.0088,
      "step": 11330
    },
    {
      "epoch": 171.74242424242425,
      "grad_norm": 0.05288395285606384,
      "learning_rate": 1.412878787878788e-05,
      "loss": 0.0103,
      "step": 11335
    },
    {
      "epoch": 171.8181818181818,
      "grad_norm": 0.06107615306973457,
      "learning_rate": 1.409090909090909e-05,
      "loss": 0.0105,
      "step": 11340
    },
    {
      "epoch": 171.8939393939394,
      "grad_norm": 0.04270188882946968,
      "learning_rate": 1.4053030303030304e-05,
      "loss": 0.0074,
      "step": 11345
    },
    {
      "epoch": 171.96969696969697,
      "grad_norm": 0.05209590494632721,
      "learning_rate": 1.4015151515151515e-05,
      "loss": 0.0093,
      "step": 11350
    },
    {
      "epoch": 172.0,
      "eval_loss": 0.01116079743951559,
      "eval_mean_accuracy": 0.957750473683717,
      "eval_mean_iou": 0.9284483221704655,
      "eval_runtime": 36.0393,
      "eval_samples_per_second": 6.521,
      "eval_steps_per_second": 0.832,
      "step": 11352
    },
    {
      "epoch": 172.04545454545453,
      "grad_norm": 0.04235956445336342,
      "learning_rate": 1.3977272727272728e-05,
      "loss": 0.0087,
      "step": 11355
    },
    {
      "epoch": 172.12121212121212,
      "grad_norm": 0.046661555767059326,
      "learning_rate": 1.3939393939393942e-05,
      "loss": 0.009,
      "step": 11360
    },
    {
      "epoch": 172.1969696969697,
      "grad_norm": 0.030435634776949883,
      "learning_rate": 1.3901515151515151e-05,
      "loss": 0.0114,
      "step": 11365
    },
    {
      "epoch": 172.27272727272728,
      "grad_norm": 0.033255066722631454,
      "learning_rate": 1.3863636363636364e-05,
      "loss": 0.007,
      "step": 11370
    },
    {
      "epoch": 172.34848484848484,
      "grad_norm": 0.016576888039708138,
      "learning_rate": 1.3825757575757576e-05,
      "loss": 0.0074,
      "step": 11375
    },
    {
      "epoch": 172.42424242424244,
      "grad_norm": 0.029440293088555336,
      "learning_rate": 1.3787878787878789e-05,
      "loss": 0.0093,
      "step": 11380
    },
    {
      "epoch": 172.5,
      "grad_norm": 0.055812858045101166,
      "learning_rate": 1.3750000000000002e-05,
      "loss": 0.0075,
      "step": 11385
    },
    {
      "epoch": 172.57575757575756,
      "grad_norm": 0.03621271625161171,
      "learning_rate": 1.3712121212121212e-05,
      "loss": 0.0086,
      "step": 11390
    },
    {
      "epoch": 172.65151515151516,
      "grad_norm": 0.032026343047618866,
      "learning_rate": 1.3674242424242425e-05,
      "loss": 0.0078,
      "step": 11395
    },
    {
      "epoch": 172.72727272727272,
      "grad_norm": 0.08968789130449295,
      "learning_rate": 1.3636363636363637e-05,
      "loss": 0.0104,
      "step": 11400
    },
    {
      "epoch": 172.8030303030303,
      "grad_norm": 0.02212809957563877,
      "learning_rate": 1.359848484848485e-05,
      "loss": 0.0095,
      "step": 11405
    },
    {
      "epoch": 172.87878787878788,
      "grad_norm": 0.09911267459392548,
      "learning_rate": 1.3560606060606063e-05,
      "loss": 0.0084,
      "step": 11410
    },
    {
      "epoch": 172.95454545454547,
      "grad_norm": 0.036964889615774155,
      "learning_rate": 1.3522727272727273e-05,
      "loss": 0.0087,
      "step": 11415
    },
    {
      "epoch": 173.0,
      "eval_loss": 0.01126506645232439,
      "eval_mean_accuracy": 0.9545728112711981,
      "eval_mean_iou": 0.9272472696604868,
      "eval_runtime": 35.1782,
      "eval_samples_per_second": 6.68,
      "eval_steps_per_second": 0.853,
      "step": 11418
    },
    {
      "epoch": 173.03030303030303,
      "grad_norm": 0.01784163899719715,
      "learning_rate": 1.3484848484848486e-05,
      "loss": 0.0068,
      "step": 11420
    },
    {
      "epoch": 173.1060606060606,
      "grad_norm": 0.03315775841474533,
      "learning_rate": 1.3446969696969697e-05,
      "loss": 0.0065,
      "step": 11425
    },
    {
      "epoch": 173.1818181818182,
      "grad_norm": 0.034062206745147705,
      "learning_rate": 1.340909090909091e-05,
      "loss": 0.0075,
      "step": 11430
    },
    {
      "epoch": 173.25757575757575,
      "grad_norm": 0.020333610475063324,
      "learning_rate": 1.337121212121212e-05,
      "loss": 0.0085,
      "step": 11435
    },
    {
      "epoch": 173.33333333333334,
      "grad_norm": 0.03370313346385956,
      "learning_rate": 1.3333333333333333e-05,
      "loss": 0.0076,
      "step": 11440
    },
    {
      "epoch": 173.4090909090909,
      "grad_norm": 0.06935623288154602,
      "learning_rate": 1.3295454545454546e-05,
      "loss": 0.0116,
      "step": 11445
    },
    {
      "epoch": 173.4848484848485,
      "grad_norm": 0.04283108189702034,
      "learning_rate": 1.3257575757575758e-05,
      "loss": 0.0088,
      "step": 11450
    },
    {
      "epoch": 173.56060606060606,
      "grad_norm": 0.026409946382045746,
      "learning_rate": 1.3219696969696971e-05,
      "loss": 0.0091,
      "step": 11455
    },
    {
      "epoch": 173.63636363636363,
      "grad_norm": 0.03250478580594063,
      "learning_rate": 1.318181818181818e-05,
      "loss": 0.0084,
      "step": 11460
    },
    {
      "epoch": 173.71212121212122,
      "grad_norm": 0.04561053588986397,
      "learning_rate": 1.3143939393939394e-05,
      "loss": 0.0095,
      "step": 11465
    },
    {
      "epoch": 173.78787878787878,
      "grad_norm": 0.02712816558778286,
      "learning_rate": 1.3106060606060607e-05,
      "loss": 0.0106,
      "step": 11470
    },
    {
      "epoch": 173.86363636363637,
      "grad_norm": 0.04662031680345535,
      "learning_rate": 1.3068181818181819e-05,
      "loss": 0.0094,
      "step": 11475
    },
    {
      "epoch": 173.93939393939394,
      "grad_norm": 0.026631493121385574,
      "learning_rate": 1.3030303030303032e-05,
      "loss": 0.0072,
      "step": 11480
    },
    {
      "epoch": 174.0,
      "eval_loss": 0.011142564937472343,
      "eval_mean_accuracy": 0.9569586304631268,
      "eval_mean_iou": 0.9281292685791712,
      "eval_runtime": 35.65,
      "eval_samples_per_second": 6.592,
      "eval_steps_per_second": 0.842,
      "step": 11484
    },
    {
      "epoch": 174.0151515151515,
      "grad_norm": 0.03863002732396126,
      "learning_rate": 1.2992424242424241e-05,
      "loss": 0.0109,
      "step": 11485
    },
    {
      "epoch": 174.0909090909091,
      "grad_norm": 0.07671835273504257,
      "learning_rate": 1.2954545454545455e-05,
      "loss": 0.0102,
      "step": 11490
    },
    {
      "epoch": 174.16666666666666,
      "grad_norm": 0.045454081147909164,
      "learning_rate": 1.2916666666666668e-05,
      "loss": 0.007,
      "step": 11495
    },
    {
      "epoch": 174.24242424242425,
      "grad_norm": 0.024831268936395645,
      "learning_rate": 1.287878787878788e-05,
      "loss": 0.0065,
      "step": 11500
    },
    {
      "epoch": 174.3181818181818,
      "grad_norm": 0.028578517958521843,
      "learning_rate": 1.2840909090909092e-05,
      "loss": 0.0087,
      "step": 11505
    },
    {
      "epoch": 174.3939393939394,
      "grad_norm": 0.027024991810321808,
      "learning_rate": 1.2803030303030302e-05,
      "loss": 0.0081,
      "step": 11510
    },
    {
      "epoch": 174.46969696969697,
      "grad_norm": 0.04755948483943939,
      "learning_rate": 1.2765151515151515e-05,
      "loss": 0.0088,
      "step": 11515
    },
    {
      "epoch": 174.54545454545453,
      "grad_norm": 0.060758076608181,
      "learning_rate": 1.2727272727272727e-05,
      "loss": 0.0081,
      "step": 11520
    },
    {
      "epoch": 174.62121212121212,
      "grad_norm": 0.03178742155432701,
      "learning_rate": 1.268939393939394e-05,
      "loss": 0.0069,
      "step": 11525
    },
    {
      "epoch": 174.6969696969697,
      "grad_norm": 0.07118017226457596,
      "learning_rate": 1.2651515151515153e-05,
      "loss": 0.0103,
      "step": 11530
    },
    {
      "epoch": 174.77272727272728,
      "grad_norm": 0.05479759722948074,
      "learning_rate": 1.2613636363636363e-05,
      "loss": 0.009,
      "step": 11535
    },
    {
      "epoch": 174.84848484848484,
      "grad_norm": 0.028825460001826286,
      "learning_rate": 1.2575757575757578e-05,
      "loss": 0.0098,
      "step": 11540
    },
    {
      "epoch": 174.92424242424244,
      "grad_norm": 0.03640101104974747,
      "learning_rate": 1.2537878787878787e-05,
      "loss": 0.0109,
      "step": 11545
    },
    {
      "epoch": 175.0,
      "grad_norm": 0.11350501328706741,
      "learning_rate": 1.25e-05,
      "loss": 0.0123,
      "step": 11550
    },
    {
      "epoch": 175.0,
      "eval_loss": 0.011204613372683525,
      "eval_mean_accuracy": 0.956714517198038,
      "eval_mean_iou": 0.9278651315205799,
      "eval_runtime": 35.0213,
      "eval_samples_per_second": 6.71,
      "eval_steps_per_second": 0.857,
      "step": 11550
    },
    {
      "epoch": 175.07575757575756,
      "grad_norm": 0.03487880527973175,
      "learning_rate": 1.2462121212121212e-05,
      "loss": 0.0086,
      "step": 11555
    },
    {
      "epoch": 175.15151515151516,
      "grad_norm": 0.03202775865793228,
      "learning_rate": 1.2424242424242424e-05,
      "loss": 0.0103,
      "step": 11560
    },
    {
      "epoch": 175.22727272727272,
      "grad_norm": 0.06830263882875443,
      "learning_rate": 1.2386363636363638e-05,
      "loss": 0.0092,
      "step": 11565
    },
    {
      "epoch": 175.3030303030303,
      "grad_norm": 0.030505405738949776,
      "learning_rate": 1.234848484848485e-05,
      "loss": 0.0113,
      "step": 11570
    },
    {
      "epoch": 175.37878787878788,
      "grad_norm": 0.04032616317272186,
      "learning_rate": 1.2310606060606061e-05,
      "loss": 0.0082,
      "step": 11575
    },
    {
      "epoch": 175.45454545454547,
      "grad_norm": 0.07028012722730637,
      "learning_rate": 1.2272727272727273e-05,
      "loss": 0.01,
      "step": 11580
    },
    {
      "epoch": 175.53030303030303,
      "grad_norm": 0.08648121356964111,
      "learning_rate": 1.2234848484848486e-05,
      "loss": 0.0095,
      "step": 11585
    },
    {
      "epoch": 175.6060606060606,
      "grad_norm": 0.03604483604431152,
      "learning_rate": 1.2196969696969697e-05,
      "loss": 0.0078,
      "step": 11590
    },
    {
      "epoch": 175.6818181818182,
      "grad_norm": 0.053039781749248505,
      "learning_rate": 1.215909090909091e-05,
      "loss": 0.0085,
      "step": 11595
    },
    {
      "epoch": 175.75757575757575,
      "grad_norm": 0.0382864736020565,
      "learning_rate": 1.2121212121212122e-05,
      "loss": 0.0078,
      "step": 11600
    },
    {
      "epoch": 175.83333333333334,
      "grad_norm": 0.04862525314092636,
      "learning_rate": 1.2083333333333333e-05,
      "loss": 0.0082,
      "step": 11605
    },
    {
      "epoch": 175.9090909090909,
      "grad_norm": 0.0402299128472805,
      "learning_rate": 1.2045454545454547e-05,
      "loss": 0.0088,
      "step": 11610
    },
    {
      "epoch": 175.9848484848485,
      "grad_norm": 0.022995924577116966,
      "learning_rate": 1.2007575757575758e-05,
      "loss": 0.0092,
      "step": 11615
    },
    {
      "epoch": 176.0,
      "eval_loss": 0.011255434714257717,
      "eval_mean_accuracy": 0.9596094755036285,
      "eval_mean_iou": 0.9285293374553327,
      "eval_runtime": 35.2295,
      "eval_samples_per_second": 6.671,
      "eval_steps_per_second": 0.852,
      "step": 11616
    },
    {
      "epoch": 176.06060606060606,
      "grad_norm": 0.044564131647348404,
      "learning_rate": 1.196969696969697e-05,
      "loss": 0.0094,
      "step": 11620
    },
    {
      "epoch": 176.13636363636363,
      "grad_norm": 0.04968610033392906,
      "learning_rate": 1.1931818181818183e-05,
      "loss": 0.0077,
      "step": 11625
    },
    {
      "epoch": 176.21212121212122,
      "grad_norm": 0.029903171584010124,
      "learning_rate": 1.1893939393939394e-05,
      "loss": 0.0079,
      "step": 11630
    },
    {
      "epoch": 176.28787878787878,
      "grad_norm": 0.022929206490516663,
      "learning_rate": 1.1856060606060607e-05,
      "loss": 0.011,
      "step": 11635
    },
    {
      "epoch": 176.36363636363637,
      "grad_norm": 0.04649556428194046,
      "learning_rate": 1.1818181818181819e-05,
      "loss": 0.0082,
      "step": 11640
    },
    {
      "epoch": 176.43939393939394,
      "grad_norm": 0.014622013084590435,
      "learning_rate": 1.178030303030303e-05,
      "loss": 0.0082,
      "step": 11645
    },
    {
      "epoch": 176.5151515151515,
      "grad_norm": 0.06694382429122925,
      "learning_rate": 1.1742424242424243e-05,
      "loss": 0.0083,
      "step": 11650
    },
    {
      "epoch": 176.5909090909091,
      "grad_norm": 0.045661792159080505,
      "learning_rate": 1.1704545454545455e-05,
      "loss": 0.0087,
      "step": 11655
    },
    {
      "epoch": 176.66666666666666,
      "grad_norm": 0.03591826185584068,
      "learning_rate": 1.1666666666666668e-05,
      "loss": 0.0072,
      "step": 11660
    },
    {
      "epoch": 176.74242424242425,
      "grad_norm": 0.04481382295489311,
      "learning_rate": 1.162878787878788e-05,
      "loss": 0.0106,
      "step": 11665
    },
    {
      "epoch": 176.8181818181818,
      "grad_norm": 0.019147872924804688,
      "learning_rate": 1.159090909090909e-05,
      "loss": 0.0088,
      "step": 11670
    },
    {
      "epoch": 176.8939393939394,
      "grad_norm": 0.04664165899157524,
      "learning_rate": 1.1553030303030302e-05,
      "loss": 0.0096,
      "step": 11675
    },
    {
      "epoch": 176.96969696969697,
      "grad_norm": 0.01997470110654831,
      "learning_rate": 1.1515151515151517e-05,
      "loss": 0.0077,
      "step": 11680
    },
    {
      "epoch": 177.0,
      "eval_loss": 0.011250576935708523,
      "eval_mean_accuracy": 0.953890699985409,
      "eval_mean_iou": 0.9272599990439465,
      "eval_runtime": 36.3071,
      "eval_samples_per_second": 6.473,
      "eval_steps_per_second": 0.826,
      "step": 11682
    },
    {
      "epoch": 177.04545454545453,
      "grad_norm": 0.03645392134785652,
      "learning_rate": 1.1477272727272729e-05,
      "loss": 0.0121,
      "step": 11685
    },
    {
      "epoch": 177.12121212121212,
      "grad_norm": 0.04503287374973297,
      "learning_rate": 1.143939393939394e-05,
      "loss": 0.0096,
      "step": 11690
    },
    {
      "epoch": 177.1969696969697,
      "grad_norm": 0.028754903003573418,
      "learning_rate": 1.1401515151515151e-05,
      "loss": 0.0112,
      "step": 11695
    },
    {
      "epoch": 177.27272727272728,
      "grad_norm": 0.06462083011865616,
      "learning_rate": 1.1363636363636365e-05,
      "loss": 0.0073,
      "step": 11700
    },
    {
      "epoch": 177.34848484848484,
      "grad_norm": 0.04854431003332138,
      "learning_rate": 1.1325757575757576e-05,
      "loss": 0.0087,
      "step": 11705
    },
    {
      "epoch": 177.42424242424244,
      "grad_norm": 0.027262629941105843,
      "learning_rate": 1.128787878787879e-05,
      "loss": 0.0065,
      "step": 11710
    },
    {
      "epoch": 177.5,
      "grad_norm": 0.03888207674026489,
      "learning_rate": 1.125e-05,
      "loss": 0.0087,
      "step": 11715
    },
    {
      "epoch": 177.57575757575756,
      "grad_norm": 0.02712400257587433,
      "learning_rate": 1.1212121212121212e-05,
      "loss": 0.0089,
      "step": 11720
    },
    {
      "epoch": 177.65151515151516,
      "grad_norm": 0.023270603269338608,
      "learning_rate": 1.1174242424242425e-05,
      "loss": 0.0067,
      "step": 11725
    },
    {
      "epoch": 177.72727272727272,
      "grad_norm": 0.026186930015683174,
      "learning_rate": 1.1136363636363637e-05,
      "loss": 0.011,
      "step": 11730
    },
    {
      "epoch": 177.8030303030303,
      "grad_norm": 0.03569154068827629,
      "learning_rate": 1.109848484848485e-05,
      "loss": 0.0092,
      "step": 11735
    },
    {
      "epoch": 177.87878787878788,
      "grad_norm": 0.045189160853624344,
      "learning_rate": 1.1060606060606061e-05,
      "loss": 0.0076,
      "step": 11740
    },
    {
      "epoch": 177.95454545454547,
      "grad_norm": 0.0363650806248188,
      "learning_rate": 1.1022727272727273e-05,
      "loss": 0.0101,
      "step": 11745
    },
    {
      "epoch": 178.0,
      "eval_loss": 0.011188831180334091,
      "eval_mean_accuracy": 0.9579694332595524,
      "eval_mean_iou": 0.9288108783417747,
      "eval_runtime": 35.9898,
      "eval_samples_per_second": 6.53,
      "eval_steps_per_second": 0.834,
      "step": 11748
    },
    {
      "epoch": 178.03030303030303,
      "grad_norm": 0.060305628925561905,
      "learning_rate": 1.0984848484848486e-05,
      "loss": 0.0111,
      "step": 11750
    },
    {
      "epoch": 178.1060606060606,
      "grad_norm": 0.0155892139300704,
      "learning_rate": 1.0946969696969697e-05,
      "loss": 0.0098,
      "step": 11755
    },
    {
      "epoch": 178.1818181818182,
      "grad_norm": 0.045956745743751526,
      "learning_rate": 1.0909090909090909e-05,
      "loss": 0.0089,
      "step": 11760
    },
    {
      "epoch": 178.25757575757575,
      "grad_norm": 0.0799398273229599,
      "learning_rate": 1.0871212121212122e-05,
      "loss": 0.0081,
      "step": 11765
    },
    {
      "epoch": 178.33333333333334,
      "grad_norm": 0.02286676876246929,
      "learning_rate": 1.0833333333333334e-05,
      "loss": 0.0062,
      "step": 11770
    },
    {
      "epoch": 178.4090909090909,
      "grad_norm": 0.04599860683083534,
      "learning_rate": 1.0795454545454547e-05,
      "loss": 0.0103,
      "step": 11775
    },
    {
      "epoch": 178.4848484848485,
      "grad_norm": 0.0578174963593483,
      "learning_rate": 1.0757575757575758e-05,
      "loss": 0.0079,
      "step": 11780
    },
    {
      "epoch": 178.56060606060606,
      "grad_norm": 0.028259223327040672,
      "learning_rate": 1.071969696969697e-05,
      "loss": 0.0077,
      "step": 11785
    },
    {
      "epoch": 178.63636363636363,
      "grad_norm": 0.02801992930471897,
      "learning_rate": 1.0681818181818181e-05,
      "loss": 0.0086,
      "step": 11790
    },
    {
      "epoch": 178.71212121212122,
      "grad_norm": 0.02514462359249592,
      "learning_rate": 1.0643939393939394e-05,
      "loss": 0.01,
      "step": 11795
    },
    {
      "epoch": 178.78787878787878,
      "grad_norm": 0.03677919879555702,
      "learning_rate": 1.0606060606060607e-05,
      "loss": 0.0087,
      "step": 11800
    },
    {
      "epoch": 178.86363636363637,
      "grad_norm": 0.029220517724752426,
      "learning_rate": 1.0568181818181819e-05,
      "loss": 0.0107,
      "step": 11805
    },
    {
      "epoch": 178.93939393939394,
      "grad_norm": 0.030421458184719086,
      "learning_rate": 1.053030303030303e-05,
      "loss": 0.0077,
      "step": 11810
    },
    {
      "epoch": 179.0,
      "eval_loss": 0.011112677864730358,
      "eval_mean_accuracy": 0.9593417288708218,
      "eval_mean_iou": 0.9290787088512936,
      "eval_runtime": 35.9793,
      "eval_samples_per_second": 6.532,
      "eval_steps_per_second": 0.834,
      "step": 11814
    },
    {
      "epoch": 179.0151515151515,
      "grad_norm": 0.08287352323532104,
      "learning_rate": 1.0492424242424242e-05,
      "loss": 0.0077,
      "step": 11815
    },
    {
      "epoch": 179.0909090909091,
      "grad_norm": 0.05548910051584244,
      "learning_rate": 1.0454545454545455e-05,
      "loss": 0.0076,
      "step": 11820
    },
    {
      "epoch": 179.16666666666666,
      "grad_norm": 0.027239250019192696,
      "learning_rate": 1.0416666666666668e-05,
      "loss": 0.0076,
      "step": 11825
    },
    {
      "epoch": 179.24242424242425,
      "grad_norm": 0.040453922003507614,
      "learning_rate": 1.037878787878788e-05,
      "loss": 0.009,
      "step": 11830
    },
    {
      "epoch": 179.3181818181818,
      "grad_norm": 0.03803437575697899,
      "learning_rate": 1.0340909090909091e-05,
      "loss": 0.0087,
      "step": 11835
    },
    {
      "epoch": 179.3939393939394,
      "grad_norm": 0.026851365342736244,
      "learning_rate": 1.0303030303030304e-05,
      "loss": 0.0071,
      "step": 11840
    },
    {
      "epoch": 179.46969696969697,
      "grad_norm": 0.07290571928024292,
      "learning_rate": 1.0265151515151516e-05,
      "loss": 0.0103,
      "step": 11845
    },
    {
      "epoch": 179.54545454545453,
      "grad_norm": 0.03377348929643631,
      "learning_rate": 1.0227272727272729e-05,
      "loss": 0.0079,
      "step": 11850
    },
    {
      "epoch": 179.62121212121212,
      "grad_norm": 0.06372508406639099,
      "learning_rate": 1.018939393939394e-05,
      "loss": 0.0112,
      "step": 11855
    },
    {
      "epoch": 179.6969696969697,
      "grad_norm": 0.0579543262720108,
      "learning_rate": 1.0151515151515152e-05,
      "loss": 0.0097,
      "step": 11860
    },
    {
      "epoch": 179.77272727272728,
      "grad_norm": 0.027400167658925056,
      "learning_rate": 1.0113636363636365e-05,
      "loss": 0.0086,
      "step": 11865
    },
    {
      "epoch": 179.84848484848484,
      "grad_norm": 0.02889391779899597,
      "learning_rate": 1.0075757575757576e-05,
      "loss": 0.0087,
      "step": 11870
    },
    {
      "epoch": 179.92424242424244,
      "grad_norm": 0.03545737266540527,
      "learning_rate": 1.0037878787878788e-05,
      "loss": 0.0084,
      "step": 11875
    },
    {
      "epoch": 180.0,
      "grad_norm": 0.09649000316858292,
      "learning_rate": 1e-05,
      "loss": 0.0086,
      "step": 11880
    },
    {
      "epoch": 180.0,
      "eval_loss": 0.01128788199275732,
      "eval_mean_accuracy": 0.9559417771450205,
      "eval_mean_iou": 0.9274656142836041,
      "eval_runtime": 36.1909,
      "eval_samples_per_second": 6.493,
      "eval_steps_per_second": 0.829,
      "step": 11880
    },
    {
      "epoch": 180.07575757575756,
      "grad_norm": 0.03915990889072418,
      "learning_rate": 9.962121212121212e-06,
      "loss": 0.0086,
      "step": 11885
    },
    {
      "epoch": 180.15151515151516,
      "grad_norm": 0.04048236086964607,
      "learning_rate": 9.924242424242425e-06,
      "loss": 0.0088,
      "step": 11890
    },
    {
      "epoch": 180.22727272727272,
      "grad_norm": 0.040013451129198074,
      "learning_rate": 9.886363636363637e-06,
      "loss": 0.0104,
      "step": 11895
    },
    {
      "epoch": 180.3030303030303,
      "grad_norm": 0.02892131172120571,
      "learning_rate": 9.848484848484848e-06,
      "loss": 0.0076,
      "step": 11900
    },
    {
      "epoch": 180.37878787878788,
      "grad_norm": 0.03062748908996582,
      "learning_rate": 9.81060606060606e-06,
      "loss": 0.0075,
      "step": 11905
    },
    {
      "epoch": 180.45454545454547,
      "grad_norm": 0.050913941115140915,
      "learning_rate": 9.772727272727273e-06,
      "loss": 0.0097,
      "step": 11910
    },
    {
      "epoch": 180.53030303030303,
      "grad_norm": 0.035344310104846954,
      "learning_rate": 9.734848484848486e-06,
      "loss": 0.0091,
      "step": 11915
    },
    {
      "epoch": 180.6060606060606,
      "grad_norm": 0.03661605343222618,
      "learning_rate": 9.696969696969698e-06,
      "loss": 0.0072,
      "step": 11920
    },
    {
      "epoch": 180.6818181818182,
      "grad_norm": 0.058419909328222275,
      "learning_rate": 9.659090909090909e-06,
      "loss": 0.0096,
      "step": 11925
    },
    {
      "epoch": 180.75757575757575,
      "grad_norm": 0.05162007361650467,
      "learning_rate": 9.62121212121212e-06,
      "loss": 0.01,
      "step": 11930
    },
    {
      "epoch": 180.83333333333334,
      "grad_norm": 0.02515142597258091,
      "learning_rate": 9.583333333333334e-06,
      "loss": 0.011,
      "step": 11935
    },
    {
      "epoch": 180.9090909090909,
      "grad_norm": 0.030540144070982933,
      "learning_rate": 9.545454545454547e-06,
      "loss": 0.0094,
      "step": 11940
    },
    {
      "epoch": 180.9848484848485,
      "grad_norm": 0.12306870520114899,
      "learning_rate": 9.507575757575758e-06,
      "loss": 0.0058,
      "step": 11945
    },
    {
      "epoch": 181.0,
      "eval_loss": 0.011314013972878456,
      "eval_mean_accuracy": 0.9594978468169467,
      "eval_mean_iou": 0.9282566826723567,
      "eval_runtime": 36.326,
      "eval_samples_per_second": 6.469,
      "eval_steps_per_second": 0.826,
      "step": 11946
    },
    {
      "epoch": 181.06060606060606,
      "grad_norm": 0.025163128972053528,
      "learning_rate": 9.46969696969697e-06,
      "loss": 0.0094,
      "step": 11950
    },
    {
      "epoch": 181.13636363636363,
      "grad_norm": 0.036352213472127914,
      "learning_rate": 9.431818181818181e-06,
      "loss": 0.0076,
      "step": 11955
    },
    {
      "epoch": 181.21212121212122,
      "grad_norm": 0.03700150549411774,
      "learning_rate": 9.393939393939394e-06,
      "loss": 0.0086,
      "step": 11960
    },
    {
      "epoch": 181.28787878787878,
      "grad_norm": 0.026830105111002922,
      "learning_rate": 9.356060606060607e-06,
      "loss": 0.0078,
      "step": 11965
    },
    {
      "epoch": 181.36363636363637,
      "grad_norm": 0.05647367984056473,
      "learning_rate": 9.318181818181819e-06,
      "loss": 0.0112,
      "step": 11970
    },
    {
      "epoch": 181.43939393939394,
      "grad_norm": 0.020343322306871414,
      "learning_rate": 9.28030303030303e-06,
      "loss": 0.0098,
      "step": 11975
    },
    {
      "epoch": 181.5151515151515,
      "grad_norm": 0.04183444008231163,
      "learning_rate": 9.242424242424244e-06,
      "loss": 0.0109,
      "step": 11980
    },
    {
      "epoch": 181.5909090909091,
      "grad_norm": 0.05286508426070213,
      "learning_rate": 9.204545454545455e-06,
      "loss": 0.0082,
      "step": 11985
    },
    {
      "epoch": 181.66666666666666,
      "grad_norm": 0.02237076312303543,
      "learning_rate": 9.166666666666666e-06,
      "loss": 0.0094,
      "step": 11990
    },
    {
      "epoch": 181.74242424242425,
      "grad_norm": 0.02466491609811783,
      "learning_rate": 9.12878787878788e-06,
      "loss": 0.0079,
      "step": 11995
    },
    {
      "epoch": 181.8181818181818,
      "grad_norm": 0.04122363030910492,
      "learning_rate": 9.090909090909091e-06,
      "loss": 0.0073,
      "step": 12000
    },
    {
      "epoch": 181.8939393939394,
      "grad_norm": 0.029484273865818977,
      "learning_rate": 9.053030303030304e-06,
      "loss": 0.0074,
      "step": 12005
    },
    {
      "epoch": 181.96969696969697,
      "grad_norm": 0.03338354080915451,
      "learning_rate": 9.015151515151516e-06,
      "loss": 0.0093,
      "step": 12010
    },
    {
      "epoch": 182.0,
      "eval_loss": 0.011343506164848804,
      "eval_mean_accuracy": 0.95532167072284,
      "eval_mean_iou": 0.9274387315000941,
      "eval_runtime": 19.4204,
      "eval_samples_per_second": 12.101,
      "eval_steps_per_second": 1.545,
      "step": 12012
    },
    {
      "epoch": 182.04545454545453,
      "grad_norm": 0.025127874687314034,
      "learning_rate": 8.977272727272727e-06,
      "loss": 0.0109,
      "step": 12015
    },
    {
      "epoch": 182.12121212121212,
      "grad_norm": 0.044240739196538925,
      "learning_rate": 8.93939393939394e-06,
      "loss": 0.0077,
      "step": 12020
    },
    {
      "epoch": 182.1969696969697,
      "grad_norm": 0.03622879832983017,
      "learning_rate": 8.901515151515152e-06,
      "loss": 0.0094,
      "step": 12025
    },
    {
      "epoch": 182.27272727272728,
      "grad_norm": 0.046923790127038956,
      "learning_rate": 8.863636363636365e-06,
      "loss": 0.0087,
      "step": 12030
    },
    {
      "epoch": 182.34848484848484,
      "grad_norm": 0.025853004306554794,
      "learning_rate": 8.825757575757576e-06,
      "loss": 0.0075,
      "step": 12035
    },
    {
      "epoch": 182.42424242424244,
      "grad_norm": 0.06660354137420654,
      "learning_rate": 8.787878787878788e-06,
      "loss": 0.0081,
      "step": 12040
    },
    {
      "epoch": 182.5,
      "grad_norm": 0.02500092424452305,
      "learning_rate": 8.75e-06,
      "loss": 0.0091,
      "step": 12045
    },
    {
      "epoch": 182.57575757575756,
      "grad_norm": 0.026268115267157555,
      "learning_rate": 8.712121212121212e-06,
      "loss": 0.0086,
      "step": 12050
    },
    {
      "epoch": 182.65151515151516,
      "grad_norm": 0.038884539157152176,
      "learning_rate": 8.674242424242426e-06,
      "loss": 0.01,
      "step": 12055
    },
    {
      "epoch": 182.72727272727272,
      "grad_norm": 0.04669790342450142,
      "learning_rate": 8.636363636363637e-06,
      "loss": 0.0085,
      "step": 12060
    },
    {
      "epoch": 182.8030303030303,
      "grad_norm": 0.033831387758255005,
      "learning_rate": 8.598484848484848e-06,
      "loss": 0.0097,
      "step": 12065
    },
    {
      "epoch": 182.87878787878788,
      "grad_norm": 0.01813281700015068,
      "learning_rate": 8.56060606060606e-06,
      "loss": 0.0075,
      "step": 12070
    },
    {
      "epoch": 182.95454545454547,
      "grad_norm": 0.026086285710334778,
      "learning_rate": 8.522727272727273e-06,
      "loss": 0.0087,
      "step": 12075
    },
    {
      "epoch": 183.0,
      "eval_loss": 0.011236431077122688,
      "eval_mean_accuracy": 0.9585536598156721,
      "eval_mean_iou": 0.9286321217101063,
      "eval_runtime": 37.7201,
      "eval_samples_per_second": 6.23,
      "eval_steps_per_second": 0.795,
      "step": 12078
    },
    {
      "epoch": 183.03030303030303,
      "grad_norm": 0.027989424765110016,
      "learning_rate": 8.484848484848486e-06,
      "loss": 0.0068,
      "step": 12080
    },
    {
      "epoch": 183.1060606060606,
      "grad_norm": 0.018936146050691605,
      "learning_rate": 8.446969696969698e-06,
      "loss": 0.007,
      "step": 12085
    },
    {
      "epoch": 183.1818181818182,
      "grad_norm": 0.02619924210011959,
      "learning_rate": 8.409090909090909e-06,
      "loss": 0.0079,
      "step": 12090
    },
    {
      "epoch": 183.25757575757575,
      "grad_norm": 0.05630619451403618,
      "learning_rate": 8.37121212121212e-06,
      "loss": 0.0101,
      "step": 12095
    },
    {
      "epoch": 183.33333333333334,
      "grad_norm": 0.03212955594062805,
      "learning_rate": 8.333333333333334e-06,
      "loss": 0.0085,
      "step": 12100
    },
    {
      "epoch": 183.4090909090909,
      "grad_norm": 0.024414272978901863,
      "learning_rate": 8.295454545454547e-06,
      "loss": 0.0085,
      "step": 12105
    },
    {
      "epoch": 183.4848484848485,
      "grad_norm": 0.03410084545612335,
      "learning_rate": 8.257575757575758e-06,
      "loss": 0.0111,
      "step": 12110
    },
    {
      "epoch": 183.56060606060606,
      "grad_norm": 0.04830676317214966,
      "learning_rate": 8.21969696969697e-06,
      "loss": 0.0091,
      "step": 12115
    },
    {
      "epoch": 183.63636363636363,
      "grad_norm": 0.02695436403155327,
      "learning_rate": 8.181818181818183e-06,
      "loss": 0.0092,
      "step": 12120
    },
    {
      "epoch": 183.71212121212122,
      "grad_norm": 0.04681434854865074,
      "learning_rate": 8.143939393939394e-06,
      "loss": 0.0104,
      "step": 12125
    },
    {
      "epoch": 183.78787878787878,
      "grad_norm": 0.048279233276844025,
      "learning_rate": 8.106060606060606e-06,
      "loss": 0.0084,
      "step": 12130
    },
    {
      "epoch": 183.86363636363637,
      "grad_norm": 0.025165922939777374,
      "learning_rate": 8.068181818181819e-06,
      "loss": 0.0071,
      "step": 12135
    },
    {
      "epoch": 183.93939393939394,
      "grad_norm": 0.03776427358388901,
      "learning_rate": 8.03030303030303e-06,
      "loss": 0.0077,
      "step": 12140
    },
    {
      "epoch": 184.0,
      "eval_loss": 0.011312532238662243,
      "eval_mean_accuracy": 0.9536640162286774,
      "eval_mean_iou": 0.9269472572441212,
      "eval_runtime": 35.9876,
      "eval_samples_per_second": 6.53,
      "eval_steps_per_second": 0.834,
      "step": 12144
    },
    {
      "epoch": 184.0151515151515,
      "grad_norm": 0.02371867746114731,
      "learning_rate": 7.992424242424244e-06,
      "loss": 0.0066,
      "step": 12145
    },
    {
      "epoch": 184.0909090909091,
      "grad_norm": 0.025068456307053566,
      "learning_rate": 7.954545454545455e-06,
      "loss": 0.0071,
      "step": 12150
    },
    {
      "epoch": 184.16666666666666,
      "grad_norm": 0.02422492951154709,
      "learning_rate": 7.916666666666667e-06,
      "loss": 0.0088,
      "step": 12155
    },
    {
      "epoch": 184.24242424242425,
      "grad_norm": 0.04911724105477333,
      "learning_rate": 7.878787878787878e-06,
      "loss": 0.0095,
      "step": 12160
    },
    {
      "epoch": 184.3181818181818,
      "grad_norm": 0.07577892392873764,
      "learning_rate": 7.840909090909091e-06,
      "loss": 0.0083,
      "step": 12165
    },
    {
      "epoch": 184.3939393939394,
      "grad_norm": 0.04017915576696396,
      "learning_rate": 7.803030303030304e-06,
      "loss": 0.0098,
      "step": 12170
    },
    {
      "epoch": 184.46969696969697,
      "grad_norm": 0.05201975256204605,
      "learning_rate": 7.765151515151516e-06,
      "loss": 0.0069,
      "step": 12175
    },
    {
      "epoch": 184.54545454545453,
      "grad_norm": 0.03345292806625366,
      "learning_rate": 7.727272727272727e-06,
      "loss": 0.0071,
      "step": 12180
    },
    {
      "epoch": 184.62121212121212,
      "grad_norm": 0.03350428119301796,
      "learning_rate": 7.689393939393939e-06,
      "loss": 0.0089,
      "step": 12185
    },
    {
      "epoch": 184.6969696969697,
      "grad_norm": 0.027396343648433685,
      "learning_rate": 7.651515151515152e-06,
      "loss": 0.0077,
      "step": 12190
    },
    {
      "epoch": 184.77272727272728,
      "grad_norm": 0.07943537831306458,
      "learning_rate": 7.613636363636364e-06,
      "loss": 0.0099,
      "step": 12195
    },
    {
      "epoch": 184.84848484848484,
      "grad_norm": 0.050358302891254425,
      "learning_rate": 7.5757575757575764e-06,
      "loss": 0.0097,
      "step": 12200
    },
    {
      "epoch": 184.92424242424244,
      "grad_norm": 0.0380476638674736,
      "learning_rate": 7.537878787878788e-06,
      "loss": 0.0091,
      "step": 12205
    },
    {
      "epoch": 185.0,
      "grad_norm": 0.08742762356996536,
      "learning_rate": 7.5e-06,
      "loss": 0.0114,
      "step": 12210
    },
    {
      "epoch": 185.0,
      "eval_loss": 0.01115019153803587,
      "eval_mean_accuracy": 0.956834680740101,
      "eval_mean_iou": 0.9281878862396016,
      "eval_runtime": 36.6522,
      "eval_samples_per_second": 6.412,
      "eval_steps_per_second": 0.819,
      "step": 12210
    },
    {
      "epoch": 185.07575757575756,
      "grad_norm": 0.040985286235809326,
      "learning_rate": 7.462121212121212e-06,
      "loss": 0.0072,
      "step": 12215
    },
    {
      "epoch": 185.15151515151516,
      "grad_norm": 0.019717803224921227,
      "learning_rate": 7.424242424242425e-06,
      "loss": 0.0073,
      "step": 12220
    },
    {
      "epoch": 185.22727272727272,
      "grad_norm": 0.050606243312358856,
      "learning_rate": 7.386363636363637e-06,
      "loss": 0.0099,
      "step": 12225
    },
    {
      "epoch": 185.3030303030303,
      "grad_norm": 0.03211050108075142,
      "learning_rate": 7.3484848484848486e-06,
      "loss": 0.0089,
      "step": 12230
    },
    {
      "epoch": 185.37878787878788,
      "grad_norm": 0.03237175941467285,
      "learning_rate": 7.310606060606061e-06,
      "loss": 0.009,
      "step": 12235
    },
    {
      "epoch": 185.45454545454547,
      "grad_norm": 0.030280107632279396,
      "learning_rate": 7.272727272727272e-06,
      "loss": 0.0086,
      "step": 12240
    },
    {
      "epoch": 185.53030303030303,
      "grad_norm": 0.029135558754205704,
      "learning_rate": 7.234848484848485e-06,
      "loss": 0.0069,
      "step": 12245
    },
    {
      "epoch": 185.6060606060606,
      "grad_norm": 0.03906872496008873,
      "learning_rate": 7.196969696969698e-06,
      "loss": 0.0098,
      "step": 12250
    },
    {
      "epoch": 185.6818181818182,
      "grad_norm": 0.03595523536205292,
      "learning_rate": 7.15909090909091e-06,
      "loss": 0.0092,
      "step": 12255
    },
    {
      "epoch": 185.75757575757575,
      "grad_norm": 0.03849807381629944,
      "learning_rate": 7.1212121212121215e-06,
      "loss": 0.0098,
      "step": 12260
    },
    {
      "epoch": 185.83333333333334,
      "grad_norm": 0.21095554530620575,
      "learning_rate": 7.083333333333334e-06,
      "loss": 0.0109,
      "step": 12265
    },
    {
      "epoch": 185.9090909090909,
      "grad_norm": 0.029788479208946228,
      "learning_rate": 7.045454545454545e-06,
      "loss": 0.0101,
      "step": 12270
    },
    {
      "epoch": 185.9848484848485,
      "grad_norm": 0.030881615355610847,
      "learning_rate": 7.007575757575758e-06,
      "loss": 0.006,
      "step": 12275
    },
    {
      "epoch": 186.0,
      "eval_loss": 0.011262509040534496,
      "eval_mean_accuracy": 0.9567718431191073,
      "eval_mean_iou": 0.9283498841289592,
      "eval_runtime": 35.9844,
      "eval_samples_per_second": 6.531,
      "eval_steps_per_second": 0.834,
      "step": 12276
    },
    {
      "epoch": 186.06060606060606,
      "grad_norm": 0.02927226945757866,
      "learning_rate": 6.969696969696971e-06,
      "loss": 0.0063,
      "step": 12280
    },
    {
      "epoch": 186.13636363636363,
      "grad_norm": 0.030820056796073914,
      "learning_rate": 6.931818181818182e-06,
      "loss": 0.0074,
      "step": 12285
    },
    {
      "epoch": 186.21212121212122,
      "grad_norm": 0.028476925566792488,
      "learning_rate": 6.8939393939393945e-06,
      "loss": 0.0087,
      "step": 12290
    },
    {
      "epoch": 186.28787878787878,
      "grad_norm": 0.05091088265180588,
      "learning_rate": 6.856060606060606e-06,
      "loss": 0.0111,
      "step": 12295
    },
    {
      "epoch": 186.36363636363637,
      "grad_norm": 0.030546588823199272,
      "learning_rate": 6.818181818181818e-06,
      "loss": 0.0098,
      "step": 12300
    },
    {
      "epoch": 186.43939393939394,
      "grad_norm": 0.08301855623722076,
      "learning_rate": 6.7803030303030314e-06,
      "loss": 0.009,
      "step": 12305
    },
    {
      "epoch": 186.5151515151515,
      "grad_norm": 0.028771119192242622,
      "learning_rate": 6.742424242424243e-06,
      "loss": 0.0076,
      "step": 12310
    },
    {
      "epoch": 186.5909090909091,
      "grad_norm": 0.035652194172143936,
      "learning_rate": 6.704545454545455e-06,
      "loss": 0.0098,
      "step": 12315
    },
    {
      "epoch": 186.66666666666666,
      "grad_norm": 0.028551382943987846,
      "learning_rate": 6.666666666666667e-06,
      "loss": 0.0083,
      "step": 12320
    },
    {
      "epoch": 186.74242424242425,
      "grad_norm": 0.020514287054538727,
      "learning_rate": 6.628787878787879e-06,
      "loss": 0.007,
      "step": 12325
    },
    {
      "epoch": 186.8181818181818,
      "grad_norm": 0.0385904535651207,
      "learning_rate": 6.59090909090909e-06,
      "loss": 0.0071,
      "step": 12330
    },
    {
      "epoch": 186.8939393939394,
      "grad_norm": 0.02218129299581051,
      "learning_rate": 6.5530303030303036e-06,
      "loss": 0.0092,
      "step": 12335
    },
    {
      "epoch": 186.96969696969697,
      "grad_norm": 0.02668556571006775,
      "learning_rate": 6.515151515151516e-06,
      "loss": 0.0098,
      "step": 12340
    },
    {
      "epoch": 187.0,
      "eval_loss": 0.011296842247247696,
      "eval_mean_accuracy": 0.9549401159488268,
      "eval_mean_iou": 0.9276974363273727,
      "eval_runtime": 34.6621,
      "eval_samples_per_second": 6.78,
      "eval_steps_per_second": 0.865,
      "step": 12342
    },
    {
      "epoch": 187.04545454545453,
      "grad_norm": 0.04236734285950661,
      "learning_rate": 6.477272727272727e-06,
      "loss": 0.0105,
      "step": 12345
    },
    {
      "epoch": 187.12121212121212,
      "grad_norm": 0.040507350116968155,
      "learning_rate": 6.43939393939394e-06,
      "loss": 0.0094,
      "step": 12350
    },
    {
      "epoch": 187.1969696969697,
      "grad_norm": 0.05261850357055664,
      "learning_rate": 6.401515151515151e-06,
      "loss": 0.0076,
      "step": 12355
    },
    {
      "epoch": 187.27272727272728,
      "grad_norm": 0.029944254085421562,
      "learning_rate": 6.363636363636363e-06,
      "loss": 0.0076,
      "step": 12360
    },
    {
      "epoch": 187.34848484848484,
      "grad_norm": 0.03646131977438927,
      "learning_rate": 6.3257575757575765e-06,
      "loss": 0.0093,
      "step": 12365
    },
    {
      "epoch": 187.42424242424244,
      "grad_norm": 0.026762230321764946,
      "learning_rate": 6.287878787878789e-06,
      "loss": 0.0068,
      "step": 12370
    },
    {
      "epoch": 187.5,
      "grad_norm": 0.022791950032114983,
      "learning_rate": 6.25e-06,
      "loss": 0.0077,
      "step": 12375
    },
    {
      "epoch": 187.57575757575756,
      "grad_norm": 0.02913871593773365,
      "learning_rate": 6.212121212121212e-06,
      "loss": 0.0106,
      "step": 12380
    },
    {
      "epoch": 187.65151515151516,
      "grad_norm": 0.06621583551168442,
      "learning_rate": 6.174242424242425e-06,
      "loss": 0.0101,
      "step": 12385
    },
    {
      "epoch": 187.72727272727272,
      "grad_norm": 0.05213572084903717,
      "learning_rate": 6.136363636363636e-06,
      "loss": 0.0089,
      "step": 12390
    },
    {
      "epoch": 187.8030303030303,
      "grad_norm": 0.05269777402281761,
      "learning_rate": 6.098484848484849e-06,
      "loss": 0.0107,
      "step": 12395
    },
    {
      "epoch": 187.87878787878788,
      "grad_norm": 0.04080025851726532,
      "learning_rate": 6.060606060606061e-06,
      "loss": 0.0074,
      "step": 12400
    },
    {
      "epoch": 187.95454545454547,
      "grad_norm": 0.02838081121444702,
      "learning_rate": 6.022727272727273e-06,
      "loss": 0.0099,
      "step": 12405
    },
    {
      "epoch": 188.0,
      "eval_loss": 0.011175473220646381,
      "eval_mean_accuracy": 0.9568832734259378,
      "eval_mean_iou": 0.928441903809369,
      "eval_runtime": 36.0239,
      "eval_samples_per_second": 6.523,
      "eval_steps_per_second": 0.833,
      "step": 12408
    },
    {
      "epoch": 188.03030303030303,
      "grad_norm": 0.034696582704782486,
      "learning_rate": 5.984848484848485e-06,
      "loss": 0.0082,
      "step": 12410
    },
    {
      "epoch": 188.1060606060606,
      "grad_norm": 0.03696562722325325,
      "learning_rate": 5.946969696969697e-06,
      "loss": 0.0086,
      "step": 12415
    },
    {
      "epoch": 188.1818181818182,
      "grad_norm": 0.03846660256385803,
      "learning_rate": 5.909090909090909e-06,
      "loss": 0.0106,
      "step": 12420
    },
    {
      "epoch": 188.25757575757575,
      "grad_norm": 0.03934213146567345,
      "learning_rate": 5.871212121212122e-06,
      "loss": 0.0085,
      "step": 12425
    },
    {
      "epoch": 188.33333333333334,
      "grad_norm": 0.03625093027949333,
      "learning_rate": 5.833333333333334e-06,
      "loss": 0.008,
      "step": 12430
    },
    {
      "epoch": 188.4090909090909,
      "grad_norm": 0.05034266784787178,
      "learning_rate": 5.795454545454545e-06,
      "loss": 0.0087,
      "step": 12435
    },
    {
      "epoch": 188.4848484848485,
      "grad_norm": 0.03710052743554115,
      "learning_rate": 5.7575757575757586e-06,
      "loss": 0.0097,
      "step": 12440
    },
    {
      "epoch": 188.56060606060606,
      "grad_norm": 0.02078043296933174,
      "learning_rate": 5.71969696969697e-06,
      "loss": 0.0081,
      "step": 12445
    },
    {
      "epoch": 188.63636363636363,
      "grad_norm": 0.019878551363945007,
      "learning_rate": 5.681818181818182e-06,
      "loss": 0.0085,
      "step": 12450
    },
    {
      "epoch": 188.71212121212122,
      "grad_norm": 0.200034499168396,
      "learning_rate": 5.643939393939395e-06,
      "loss": 0.0097,
      "step": 12455
    },
    {
      "epoch": 188.78787878787878,
      "grad_norm": 0.032673101872205734,
      "learning_rate": 5.606060606060606e-06,
      "loss": 0.0075,
      "step": 12460
    },
    {
      "epoch": 188.86363636363637,
      "grad_norm": 0.04244418814778328,
      "learning_rate": 5.568181818181818e-06,
      "loss": 0.0064,
      "step": 12465
    },
    {
      "epoch": 188.93939393939394,
      "grad_norm": 0.04551523178815842,
      "learning_rate": 5.530303030303031e-06,
      "loss": 0.0095,
      "step": 12470
    },
    {
      "epoch": 189.0,
      "eval_loss": 0.011049928143620491,
      "eval_mean_accuracy": 0.9578599566179876,
      "eval_mean_iou": 0.9289480842667232,
      "eval_runtime": 35.1009,
      "eval_samples_per_second": 6.695,
      "eval_steps_per_second": 0.855,
      "step": 12474
    },
    {
      "epoch": 189.0151515151515,
      "grad_norm": 0.02372005395591259,
      "learning_rate": 5.492424242424243e-06,
      "loss": 0.0112,
      "step": 12475
    },
    {
      "epoch": 189.0909090909091,
      "grad_norm": 0.015009976923465729,
      "learning_rate": 5.4545454545454545e-06,
      "loss": 0.0096,
      "step": 12480
    },
    {
      "epoch": 189.16666666666666,
      "grad_norm": 0.07032933831214905,
      "learning_rate": 5.416666666666667e-06,
      "loss": 0.0082,
      "step": 12485
    },
    {
      "epoch": 189.24242424242425,
      "grad_norm": 0.04756847023963928,
      "learning_rate": 5.378787878787879e-06,
      "loss": 0.0105,
      "step": 12490
    },
    {
      "epoch": 189.3181818181818,
      "grad_norm": 0.05528568848967552,
      "learning_rate": 5.3409090909090905e-06,
      "loss": 0.0065,
      "step": 12495
    },
    {
      "epoch": 189.3939393939394,
      "grad_norm": 0.03820636123418808,
      "learning_rate": 5.303030303030304e-06,
      "loss": 0.01,
      "step": 12500
    },
    {
      "epoch": 189.46969696969697,
      "grad_norm": 0.017786502838134766,
      "learning_rate": 5.265151515151515e-06,
      "loss": 0.008,
      "step": 12505
    },
    {
      "epoch": 189.54545454545453,
      "grad_norm": 0.03382838889956474,
      "learning_rate": 5.2272727272727274e-06,
      "loss": 0.0072,
      "step": 12510
    },
    {
      "epoch": 189.62121212121212,
      "grad_norm": 0.04969978332519531,
      "learning_rate": 5.18939393939394e-06,
      "loss": 0.013,
      "step": 12515
    },
    {
      "epoch": 189.6969696969697,
      "grad_norm": 0.06430383026599884,
      "learning_rate": 5.151515151515152e-06,
      "loss": 0.008,
      "step": 12520
    },
    {
      "epoch": 189.77272727272728,
      "grad_norm": 0.06283589452505112,
      "learning_rate": 5.113636363636364e-06,
      "loss": 0.0095,
      "step": 12525
    },
    {
      "epoch": 189.84848484848484,
      "grad_norm": 0.03853165730834007,
      "learning_rate": 5.075757575757576e-06,
      "loss": 0.0067,
      "step": 12530
    },
    {
      "epoch": 189.92424242424244,
      "grad_norm": 0.03322345390915871,
      "learning_rate": 5.037878787878788e-06,
      "loss": 0.008,
      "step": 12535
    },
    {
      "epoch": 190.0,
      "grad_norm": 0.05452122166752815,
      "learning_rate": 5e-06,
      "loss": 0.0081,
      "step": 12540
    },
    {
      "epoch": 190.0,
      "eval_loss": 0.011281930841505527,
      "eval_mean_accuracy": 0.957227898785897,
      "eval_mean_iou": 0.9284021173148247,
      "eval_runtime": 34.7955,
      "eval_samples_per_second": 6.754,
      "eval_steps_per_second": 0.862,
      "step": 12540
    },
    {
      "epoch": 190.07575757575756,
      "grad_norm": 0.028724918141961098,
      "learning_rate": 4.962121212121213e-06,
      "loss": 0.0095,
      "step": 12545
    },
    {
      "epoch": 190.15151515151516,
      "grad_norm": 0.022306367754936218,
      "learning_rate": 4.924242424242424e-06,
      "loss": 0.0066,
      "step": 12550
    },
    {
      "epoch": 190.22727272727272,
      "grad_norm": 0.10543543100357056,
      "learning_rate": 4.8863636363636365e-06,
      "loss": 0.0088,
      "step": 12555
    },
    {
      "epoch": 190.3030303030303,
      "grad_norm": 0.03745311498641968,
      "learning_rate": 4.848484848484849e-06,
      "loss": 0.0084,
      "step": 12560
    },
    {
      "epoch": 190.37878787878788,
      "grad_norm": 0.03645028546452522,
      "learning_rate": 4.81060606060606e-06,
      "loss": 0.008,
      "step": 12565
    },
    {
      "epoch": 190.45454545454547,
      "grad_norm": 0.02769261971116066,
      "learning_rate": 4.772727272727273e-06,
      "loss": 0.0116,
      "step": 12570
    },
    {
      "epoch": 190.53030303030303,
      "grad_norm": 0.0208248533308506,
      "learning_rate": 4.734848484848485e-06,
      "loss": 0.0091,
      "step": 12575
    },
    {
      "epoch": 190.6060606060606,
      "grad_norm": 0.02650870382785797,
      "learning_rate": 4.696969696969697e-06,
      "loss": 0.0061,
      "step": 12580
    },
    {
      "epoch": 190.6818181818182,
      "grad_norm": 0.02445799484848976,
      "learning_rate": 4.6590909090909095e-06,
      "loss": 0.0094,
      "step": 12585
    },
    {
      "epoch": 190.75757575757575,
      "grad_norm": 0.042442042380571365,
      "learning_rate": 4.621212121212122e-06,
      "loss": 0.0118,
      "step": 12590
    },
    {
      "epoch": 190.83333333333334,
      "grad_norm": 0.03138742968440056,
      "learning_rate": 4.583333333333333e-06,
      "loss": 0.0075,
      "step": 12595
    },
    {
      "epoch": 190.9090909090909,
      "grad_norm": 0.048323728144168854,
      "learning_rate": 4.5454545454545455e-06,
      "loss": 0.0088,
      "step": 12600
    },
    {
      "epoch": 190.9848484848485,
      "grad_norm": 0.05947694182395935,
      "learning_rate": 4.507575757575758e-06,
      "loss": 0.0092,
      "step": 12605
    },
    {
      "epoch": 191.0,
      "eval_loss": 0.011221452616155148,
      "eval_mean_accuracy": 0.956500823425471,
      "eval_mean_iou": 0.928068799165354,
      "eval_runtime": 34.567,
      "eval_samples_per_second": 6.798,
      "eval_steps_per_second": 0.868,
      "step": 12606
    },
    {
      "epoch": 191.06060606060606,
      "grad_norm": 0.030747614800930023,
      "learning_rate": 4.46969696969697e-06,
      "loss": 0.0064,
      "step": 12610
    },
    {
      "epoch": 191.13636363636363,
      "grad_norm": 0.04811907559633255,
      "learning_rate": 4.4318181818181824e-06,
      "loss": 0.0093,
      "step": 12615
    },
    {
      "epoch": 191.21212121212122,
      "grad_norm": 0.02426358126103878,
      "learning_rate": 4.393939393939394e-06,
      "loss": 0.0078,
      "step": 12620
    },
    {
      "epoch": 191.28787878787878,
      "grad_norm": 0.015582934021949768,
      "learning_rate": 4.356060606060606e-06,
      "loss": 0.0087,
      "step": 12625
    },
    {
      "epoch": 191.36363636363637,
      "grad_norm": 0.03176547959446907,
      "learning_rate": 4.3181818181818185e-06,
      "loss": 0.0098,
      "step": 12630
    },
    {
      "epoch": 191.43939393939394,
      "grad_norm": 0.05080818012356758,
      "learning_rate": 4.28030303030303e-06,
      "loss": 0.0114,
      "step": 12635
    },
    {
      "epoch": 191.5151515151515,
      "grad_norm": 0.034453704953193665,
      "learning_rate": 4.242424242424243e-06,
      "loss": 0.0098,
      "step": 12640
    },
    {
      "epoch": 191.5909090909091,
      "grad_norm": 0.029454868286848068,
      "learning_rate": 4.2045454545454546e-06,
      "loss": 0.0088,
      "step": 12645
    },
    {
      "epoch": 191.66666666666666,
      "grad_norm": 0.03008393757045269,
      "learning_rate": 4.166666666666667e-06,
      "loss": 0.0087,
      "step": 12650
    },
    {
      "epoch": 191.74242424242425,
      "grad_norm": 0.027874570339918137,
      "learning_rate": 4.128787878787879e-06,
      "loss": 0.0079,
      "step": 12655
    },
    {
      "epoch": 191.8181818181818,
      "grad_norm": 0.047617558389902115,
      "learning_rate": 4.0909090909090915e-06,
      "loss": 0.0063,
      "step": 12660
    },
    {
      "epoch": 191.8939393939394,
      "grad_norm": 0.020621171221137047,
      "learning_rate": 4.053030303030303e-06,
      "loss": 0.0076,
      "step": 12665
    },
    {
      "epoch": 191.96969696969697,
      "grad_norm": 0.034963659942150116,
      "learning_rate": 4.015151515151515e-06,
      "loss": 0.0102,
      "step": 12670
    },
    {
      "epoch": 192.0,
      "eval_loss": 0.011116931214928627,
      "eval_mean_accuracy": 0.9580383831978864,
      "eval_mean_iou": 0.9290258920054608,
      "eval_runtime": 35.1182,
      "eval_samples_per_second": 6.692,
      "eval_steps_per_second": 0.854,
      "step": 12672
    },
    {
      "epoch": 192.04545454545453,
      "grad_norm": 0.01927831768989563,
      "learning_rate": 3.9772727272727275e-06,
      "loss": 0.0074,
      "step": 12675
    },
    {
      "epoch": 192.12121212121212,
      "grad_norm": 0.03503023460507393,
      "learning_rate": 3.939393939393939e-06,
      "loss": 0.0063,
      "step": 12680
    },
    {
      "epoch": 192.1969696969697,
      "grad_norm": 0.047278232872486115,
      "learning_rate": 3.901515151515152e-06,
      "loss": 0.0079,
      "step": 12685
    },
    {
      "epoch": 192.27272727272728,
      "grad_norm": 0.02908201515674591,
      "learning_rate": 3.863636363636364e-06,
      "loss": 0.0089,
      "step": 12690
    },
    {
      "epoch": 192.34848484848484,
      "grad_norm": 0.04786701127886772,
      "learning_rate": 3.825757575757576e-06,
      "loss": 0.0096,
      "step": 12695
    },
    {
      "epoch": 192.42424242424244,
      "grad_norm": 0.08360174298286438,
      "learning_rate": 3.7878787878787882e-06,
      "loss": 0.0103,
      "step": 12700
    },
    {
      "epoch": 192.5,
      "grad_norm": 0.04692220315337181,
      "learning_rate": 3.75e-06,
      "loss": 0.0088,
      "step": 12705
    },
    {
      "epoch": 192.57575757575756,
      "grad_norm": 0.0636405274271965,
      "learning_rate": 3.7121212121212124e-06,
      "loss": 0.0086,
      "step": 12710
    },
    {
      "epoch": 192.65151515151516,
      "grad_norm": 0.030773058533668518,
      "learning_rate": 3.6742424242424243e-06,
      "loss": 0.0089,
      "step": 12715
    },
    {
      "epoch": 192.72727272727272,
      "grad_norm": 0.03298947960138321,
      "learning_rate": 3.636363636363636e-06,
      "loss": 0.0083,
      "step": 12720
    },
    {
      "epoch": 192.8030303030303,
      "grad_norm": 0.027607334777712822,
      "learning_rate": 3.598484848484849e-06,
      "loss": 0.0096,
      "step": 12725
    },
    {
      "epoch": 192.87878787878788,
      "grad_norm": 0.03221719712018967,
      "learning_rate": 3.5606060606060608e-06,
      "loss": 0.0102,
      "step": 12730
    },
    {
      "epoch": 192.95454545454547,
      "grad_norm": 0.038241900503635406,
      "learning_rate": 3.5227272727272726e-06,
      "loss": 0.008,
      "step": 12735
    },
    {
      "epoch": 193.0,
      "eval_loss": 0.011079649440944195,
      "eval_mean_accuracy": 0.9572468511695315,
      "eval_mean_iou": 0.9289587990872121,
      "eval_runtime": 33.8919,
      "eval_samples_per_second": 6.934,
      "eval_steps_per_second": 0.885,
      "step": 12738
    },
    {
      "epoch": 193.03030303030303,
      "grad_norm": 0.04714489355683327,
      "learning_rate": 3.4848484848484854e-06,
      "loss": 0.0081,
      "step": 12740
    },
    {
      "epoch": 193.1060606060606,
      "grad_norm": 0.04612404853105545,
      "learning_rate": 3.4469696969696973e-06,
      "loss": 0.0088,
      "step": 12745
    },
    {
      "epoch": 193.1818181818182,
      "grad_norm": 0.07818322628736496,
      "learning_rate": 3.409090909090909e-06,
      "loss": 0.0082,
      "step": 12750
    },
    {
      "epoch": 193.25757575757575,
      "grad_norm": 0.07131578773260117,
      "learning_rate": 3.3712121212121214e-06,
      "loss": 0.0095,
      "step": 12755
    },
    {
      "epoch": 193.33333333333334,
      "grad_norm": 0.01942998357117176,
      "learning_rate": 3.3333333333333333e-06,
      "loss": 0.0096,
      "step": 12760
    },
    {
      "epoch": 193.4090909090909,
      "grad_norm": 0.02479085698723793,
      "learning_rate": 3.295454545454545e-06,
      "loss": 0.007,
      "step": 12765
    },
    {
      "epoch": 193.4848484848485,
      "grad_norm": 0.025470804423093796,
      "learning_rate": 3.257575757575758e-06,
      "loss": 0.0077,
      "step": 12770
    },
    {
      "epoch": 193.56060606060606,
      "grad_norm": 0.05264906585216522,
      "learning_rate": 3.21969696969697e-06,
      "loss": 0.0089,
      "step": 12775
    },
    {
      "epoch": 193.63636363636363,
      "grad_norm": 0.061719417572021484,
      "learning_rate": 3.1818181818181817e-06,
      "loss": 0.0081,
      "step": 12780
    },
    {
      "epoch": 193.71212121212122,
      "grad_norm": 0.08873826265335083,
      "learning_rate": 3.1439393939393944e-06,
      "loss": 0.0112,
      "step": 12785
    },
    {
      "epoch": 193.78787878787878,
      "grad_norm": 0.019469814375042915,
      "learning_rate": 3.106060606060606e-06,
      "loss": 0.0078,
      "step": 12790
    },
    {
      "epoch": 193.86363636363637,
      "grad_norm": 0.04836123064160347,
      "learning_rate": 3.068181818181818e-06,
      "loss": 0.0069,
      "step": 12795
    },
    {
      "epoch": 193.93939393939394,
      "grad_norm": 0.028848199173808098,
      "learning_rate": 3.0303030303030305e-06,
      "loss": 0.0089,
      "step": 12800
    },
    {
      "epoch": 194.0,
      "eval_loss": 0.011147638782858849,
      "eval_mean_accuracy": 0.9545456161484499,
      "eval_mean_iou": 0.9278466201613077,
      "eval_runtime": 34.8249,
      "eval_samples_per_second": 6.748,
      "eval_steps_per_second": 0.861,
      "step": 12804
    },
    {
      "epoch": 194.0151515151515,
      "grad_norm": 0.0612356923520565,
      "learning_rate": 2.9924242424242424e-06,
      "loss": 0.0107,
      "step": 12805
    },
    {
      "epoch": 194.0909090909091,
      "grad_norm": 0.044695932418107986,
      "learning_rate": 2.9545454545454547e-06,
      "loss": 0.0093,
      "step": 12810
    },
    {
      "epoch": 194.16666666666666,
      "grad_norm": 0.022538181394338608,
      "learning_rate": 2.916666666666667e-06,
      "loss": 0.0072,
      "step": 12815
    },
    {
      "epoch": 194.24242424242425,
      "grad_norm": 0.03687781095504761,
      "learning_rate": 2.8787878787878793e-06,
      "loss": 0.0089,
      "step": 12820
    },
    {
      "epoch": 194.3181818181818,
      "grad_norm": 0.06206238269805908,
      "learning_rate": 2.840909090909091e-06,
      "loss": 0.0074,
      "step": 12825
    },
    {
      "epoch": 194.3939393939394,
      "grad_norm": 0.03818834200501442,
      "learning_rate": 2.803030303030303e-06,
      "loss": 0.0111,
      "step": 12830
    },
    {
      "epoch": 194.46969696969697,
      "grad_norm": 0.02652554214000702,
      "learning_rate": 2.7651515151515153e-06,
      "loss": 0.011,
      "step": 12835
    },
    {
      "epoch": 194.54545454545453,
      "grad_norm": 0.024552030488848686,
      "learning_rate": 2.7272727272727272e-06,
      "loss": 0.0062,
      "step": 12840
    },
    {
      "epoch": 194.62121212121212,
      "grad_norm": 0.07025913894176483,
      "learning_rate": 2.6893939393939395e-06,
      "loss": 0.0067,
      "step": 12845
    },
    {
      "epoch": 194.6969696969697,
      "grad_norm": 0.02555004321038723,
      "learning_rate": 2.651515151515152e-06,
      "loss": 0.0082,
      "step": 12850
    },
    {
      "epoch": 194.77272727272728,
      "grad_norm": 0.05926690623164177,
      "learning_rate": 2.6136363636363637e-06,
      "loss": 0.0121,
      "step": 12855
    },
    {
      "epoch": 194.84848484848484,
      "grad_norm": 0.041267018765211105,
      "learning_rate": 2.575757575757576e-06,
      "loss": 0.0077,
      "step": 12860
    },
    {
      "epoch": 194.92424242424244,
      "grad_norm": 0.031060513108968735,
      "learning_rate": 2.537878787878788e-06,
      "loss": 0.0089,
      "step": 12865
    },
    {
      "epoch": 195.0,
      "grad_norm": 0.09561950713396072,
      "learning_rate": 2.5e-06,
      "loss": 0.0082,
      "step": 12870
    },
    {
      "epoch": 195.0,
      "eval_loss": 0.011095888912677765,
      "eval_mean_accuracy": 0.955877535476941,
      "eval_mean_iou": 0.9283532690593234,
      "eval_runtime": 34.2757,
      "eval_samples_per_second": 6.856,
      "eval_steps_per_second": 0.875,
      "step": 12870
    },
    {
      "epoch": 195.07575757575756,
      "grad_norm": 0.02377871796488762,
      "learning_rate": 2.462121212121212e-06,
      "loss": 0.0093,
      "step": 12875
    },
    {
      "epoch": 195.15151515151516,
      "grad_norm": 0.08005847781896591,
      "learning_rate": 2.4242424242424244e-06,
      "loss": 0.0092,
      "step": 12880
    },
    {
      "epoch": 195.22727272727272,
      "grad_norm": 0.027721857652068138,
      "learning_rate": 2.3863636363636367e-06,
      "loss": 0.0081,
      "step": 12885
    },
    {
      "epoch": 195.3030303030303,
      "grad_norm": 0.03450048342347145,
      "learning_rate": 2.3484848484848486e-06,
      "loss": 0.0094,
      "step": 12890
    },
    {
      "epoch": 195.37878787878788,
      "grad_norm": 0.024139227345585823,
      "learning_rate": 2.310606060606061e-06,
      "loss": 0.0085,
      "step": 12895
    },
    {
      "epoch": 195.45454545454547,
      "grad_norm": 0.02698620967566967,
      "learning_rate": 2.2727272727272728e-06,
      "loss": 0.0073,
      "step": 12900
    },
    {
      "epoch": 195.53030303030303,
      "grad_norm": 0.06111720949411392,
      "learning_rate": 2.234848484848485e-06,
      "loss": 0.0088,
      "step": 12905
    },
    {
      "epoch": 195.6060606060606,
      "grad_norm": 0.022991666570305824,
      "learning_rate": 2.196969696969697e-06,
      "loss": 0.0063,
      "step": 12910
    },
    {
      "epoch": 195.6818181818182,
      "grad_norm": 0.04470168426632881,
      "learning_rate": 2.1590909090909092e-06,
      "loss": 0.0103,
      "step": 12915
    },
    {
      "epoch": 195.75757575757575,
      "grad_norm": 0.02187919057905674,
      "learning_rate": 2.1212121212121216e-06,
      "loss": 0.0078,
      "step": 12920
    },
    {
      "epoch": 195.83333333333334,
      "grad_norm": 0.023166587576270103,
      "learning_rate": 2.0833333333333334e-06,
      "loss": 0.0079,
      "step": 12925
    },
    {
      "epoch": 195.9090909090909,
      "grad_norm": 0.0333612784743309,
      "learning_rate": 2.0454545454545457e-06,
      "loss": 0.0113,
      "step": 12930
    },
    {
      "epoch": 195.9848484848485,
      "grad_norm": 0.04139534384012222,
      "learning_rate": 2.0075757575757576e-06,
      "loss": 0.0073,
      "step": 12935
    },
    {
      "epoch": 196.0,
      "eval_loss": 0.011130363680422306,
      "eval_mean_accuracy": 0.9566440897083649,
      "eval_mean_iou": 0.9285688592607831,
      "eval_runtime": 34.6896,
      "eval_samples_per_second": 6.774,
      "eval_steps_per_second": 0.865,
      "step": 12936
    },
    {
      "epoch": 196.06060606060606,
      "grad_norm": 0.03473442420363426,
      "learning_rate": 1.9696969696969695e-06,
      "loss": 0.0097,
      "step": 12940
    },
    {
      "epoch": 196.13636363636363,
      "grad_norm": 0.07648752629756927,
      "learning_rate": 1.931818181818182e-06,
      "loss": 0.0093,
      "step": 12945
    },
    {
      "epoch": 196.21212121212122,
      "grad_norm": 0.02355222962796688,
      "learning_rate": 1.8939393939393941e-06,
      "loss": 0.0087,
      "step": 12950
    },
    {
      "epoch": 196.28787878787878,
      "grad_norm": 0.02421446517109871,
      "learning_rate": 1.8560606060606062e-06,
      "loss": 0.0091,
      "step": 12955
    },
    {
      "epoch": 196.36363636363637,
      "grad_norm": 0.031064894050359726,
      "learning_rate": 1.818181818181818e-06,
      "loss": 0.0075,
      "step": 12960
    },
    {
      "epoch": 196.43939393939394,
      "grad_norm": 0.03222593665122986,
      "learning_rate": 1.7803030303030304e-06,
      "loss": 0.0083,
      "step": 12965
    },
    {
      "epoch": 196.5151515151515,
      "grad_norm": 0.043864600360393524,
      "learning_rate": 1.7424242424242427e-06,
      "loss": 0.0093,
      "step": 12970
    },
    {
      "epoch": 196.5909090909091,
      "grad_norm": 0.034242983907461166,
      "learning_rate": 1.7045454545454546e-06,
      "loss": 0.0088,
      "step": 12975
    },
    {
      "epoch": 196.66666666666666,
      "grad_norm": 0.025685643777251244,
      "learning_rate": 1.6666666666666667e-06,
      "loss": 0.0073,
      "step": 12980
    },
    {
      "epoch": 196.74242424242425,
      "grad_norm": 0.03224683925509453,
      "learning_rate": 1.628787878787879e-06,
      "loss": 0.0082,
      "step": 12985
    },
    {
      "epoch": 196.8181818181818,
      "grad_norm": 0.06098167970776558,
      "learning_rate": 1.5909090909090908e-06,
      "loss": 0.0087,
      "step": 12990
    },
    {
      "epoch": 196.8939393939394,
      "grad_norm": 0.06020820513367653,
      "learning_rate": 1.553030303030303e-06,
      "loss": 0.0099,
      "step": 12995
    },
    {
      "epoch": 196.96969696969697,
      "grad_norm": 0.05158478394150734,
      "learning_rate": 1.5151515151515152e-06,
      "loss": 0.011,
      "step": 13000
    },
    {
      "epoch": 197.0,
      "eval_loss": 0.011093096807599068,
      "eval_mean_accuracy": 0.9561248933078126,
      "eval_mean_iou": 0.9285475643066524,
      "eval_runtime": 35.4431,
      "eval_samples_per_second": 6.63,
      "eval_steps_per_second": 0.846,
      "step": 13002
    },
    {
      "epoch": 197.04545454545453,
      "grad_norm": 0.025060370564460754,
      "learning_rate": 1.4772727272727273e-06,
      "loss": 0.0056,
      "step": 13005
    },
    {
      "epoch": 197.12121212121212,
      "grad_norm": 0.06854967027902603,
      "learning_rate": 1.4393939393939396e-06,
      "loss": 0.0098,
      "step": 13010
    },
    {
      "epoch": 197.1969696969697,
      "grad_norm": 0.03706763684749603,
      "learning_rate": 1.4015151515151515e-06,
      "loss": 0.0089,
      "step": 13015
    },
    {
      "epoch": 197.27272727272728,
      "grad_norm": 0.0357099212706089,
      "learning_rate": 1.3636363636363636e-06,
      "loss": 0.0087,
      "step": 13020
    },
    {
      "epoch": 197.34848484848484,
      "grad_norm": 0.01608971506357193,
      "learning_rate": 1.325757575757576e-06,
      "loss": 0.0072,
      "step": 13025
    },
    {
      "epoch": 197.42424242424244,
      "grad_norm": 0.028266845270991325,
      "learning_rate": 1.287878787878788e-06,
      "loss": 0.0069,
      "step": 13030
    },
    {
      "epoch": 197.5,
      "grad_norm": 0.034050069749355316,
      "learning_rate": 1.25e-06,
      "loss": 0.0107,
      "step": 13035
    },
    {
      "epoch": 197.57575757575756,
      "grad_norm": 0.037165168672800064,
      "learning_rate": 1.2121212121212122e-06,
      "loss": 0.0107,
      "step": 13040
    },
    {
      "epoch": 197.65151515151516,
      "grad_norm": 0.0331307128071785,
      "learning_rate": 1.1742424242424243e-06,
      "loss": 0.0069,
      "step": 13045
    },
    {
      "epoch": 197.72727272727272,
      "grad_norm": 0.056624822318553925,
      "learning_rate": 1.1363636363636364e-06,
      "loss": 0.0112,
      "step": 13050
    },
    {
      "epoch": 197.8030303030303,
      "grad_norm": 0.0371989868581295,
      "learning_rate": 1.0984848484848485e-06,
      "loss": 0.0095,
      "step": 13055
    },
    {
      "epoch": 197.87878787878788,
      "grad_norm": 0.024042312055826187,
      "learning_rate": 1.0606060606060608e-06,
      "loss": 0.0061,
      "step": 13060
    },
    {
      "epoch": 197.95454545454547,
      "grad_norm": 0.020856795832514763,
      "learning_rate": 1.0227272727272729e-06,
      "loss": 0.006,
      "step": 13065
    },
    {
      "epoch": 198.0,
      "eval_loss": 0.011225304566323757,
      "eval_mean_accuracy": 0.9538213591935866,
      "eval_mean_iou": 0.9275838510313744,
      "eval_runtime": 35.0783,
      "eval_samples_per_second": 6.699,
      "eval_steps_per_second": 0.855,
      "step": 13068
    },
    {
      "epoch": 198.03030303030303,
      "grad_norm": 0.021560063585639,
      "learning_rate": 9.848484848484847e-07,
      "loss": 0.0156,
      "step": 13070
    },
    {
      "epoch": 198.1060606060606,
      "grad_norm": 0.02193550206720829,
      "learning_rate": 9.469696969696971e-07,
      "loss": 0.0086,
      "step": 13075
    },
    {
      "epoch": 198.1818181818182,
      "grad_norm": 0.025004630908370018,
      "learning_rate": 9.09090909090909e-07,
      "loss": 0.0099,
      "step": 13080
    },
    {
      "epoch": 198.25757575757575,
      "grad_norm": 0.026762578636407852,
      "learning_rate": 8.712121212121213e-07,
      "loss": 0.0086,
      "step": 13085
    },
    {
      "epoch": 198.33333333333334,
      "grad_norm": 0.05060798302292824,
      "learning_rate": 8.333333333333333e-07,
      "loss": 0.0101,
      "step": 13090
    },
    {
      "epoch": 198.4090909090909,
      "grad_norm": 0.02744155377149582,
      "learning_rate": 7.954545454545454e-07,
      "loss": 0.0076,
      "step": 13095
    },
    {
      "epoch": 198.4848484848485,
      "grad_norm": 0.03501023352146149,
      "learning_rate": 7.575757575757576e-07,
      "loss": 0.0113,
      "step": 13100
    },
    {
      "epoch": 198.56060606060606,
      "grad_norm": 0.03702348843216896,
      "learning_rate": 7.196969696969698e-07,
      "loss": 0.0083,
      "step": 13105
    },
    {
      "epoch": 198.63636363636363,
      "grad_norm": 0.021939270198345184,
      "learning_rate": 6.818181818181818e-07,
      "loss": 0.007,
      "step": 13110
    },
    {
      "epoch": 198.71212121212122,
      "grad_norm": 0.014872685074806213,
      "learning_rate": 6.43939393939394e-07,
      "loss": 0.0065,
      "step": 13115
    },
    {
      "epoch": 198.78787878787878,
      "grad_norm": 0.06689616292715073,
      "learning_rate": 6.060606060606061e-07,
      "loss": 0.0101,
      "step": 13120
    },
    {
      "epoch": 198.86363636363637,
      "grad_norm": 0.028215624392032623,
      "learning_rate": 5.681818181818182e-07,
      "loss": 0.0085,
      "step": 13125
    },
    {
      "epoch": 198.93939393939394,
      "grad_norm": 0.03036646917462349,
      "learning_rate": 5.303030303030304e-07,
      "loss": 0.0103,
      "step": 13130
    },
    {
      "epoch": 199.0,
      "eval_loss": 0.011061455123126507,
      "eval_mean_accuracy": 0.9583028551020276,
      "eval_mean_iou": 0.929185164986373,
      "eval_runtime": 33.9717,
      "eval_samples_per_second": 6.918,
      "eval_steps_per_second": 0.883,
      "step": 13134
    }
  ],
  "logging_steps": 5,
  "max_steps": 13200,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 200,
  "save_steps": 500,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": false
      },
      "attributes": {}
    }
  },
  "total_flos": 0.0,
  "train_batch_size": 8,
  "trial_name": null,
  "trial_params": null
}