diff --git "a/trainer_state.json" "b/trainer_state.json"
new file mode 100644--- /dev/null
+++ "b/trainer_state.json"
@@ -0,0 +1,10782 @@
+{
+  "best_global_step": null,
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 0.002102738690980681,
+  "eval_steps": 500,
+  "global_step": 15000,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0,
+      "eval_loss": 10.238426208496094,
+      "eval_runtime": 111.932,
+      "eval_samples_per_second": 204.079,
+      "eval_steps_per_second": 1.599,
+      "step": 0
+    },
+    {
+      "epoch": 1.4018257939871205e-06,
+      "grad_norm": 235.0,
+      "learning_rate": 2.6999999999999996e-06,
+      "loss": 10.1619,
+      "step": 10
+    },
+    {
+      "epoch": 2.803651587974241e-06,
+      "grad_norm": 187.0,
+      "learning_rate": 5.7e-06,
+      "loss": 10.0993,
+      "step": 20
+    },
+    {
+      "epoch": 4.205477381961362e-06,
+      "grad_norm": 203.0,
+      "learning_rate": 8.7e-06,
+      "loss": 9.9129,
+      "step": 30
+    },
+    {
+      "epoch": 5.607303175948482e-06,
+      "grad_norm": 178.0,
+      "learning_rate": 1.17e-05,
+      "loss": 9.5407,
+      "step": 40
+    },
+    {
+      "epoch": 7.009128969935603e-06,
+      "grad_norm": 230.0,
+      "learning_rate": 1.47e-05,
+      "loss": 9.1468,
+      "step": 50
+    },
+    {
+      "epoch": 8.410954763922724e-06,
+      "grad_norm": 290.0,
+      "learning_rate": 1.7699999999999997e-05,
+      "loss": 8.5916,
+      "step": 60
+    },
+    {
+      "epoch": 9.812780557909844e-06,
+      "grad_norm": 398.0,
+      "learning_rate": 2.07e-05,
+      "loss": 8.1624,
+      "step": 70
+    },
+    {
+      "epoch": 1.1214606351896964e-05,
+      "grad_norm": 340.0,
+      "learning_rate": 2.3699999999999997e-05,
+      "loss": 8.0425,
+      "step": 80
+    },
+    {
+      "epoch": 1.2616432145884086e-05,
+      "grad_norm": 368.0,
+      "learning_rate": 2.6699999999999995e-05,
+      "loss": 7.2555,
+      "step": 90
+    },
+    {
+      "epoch": 1.4018257939871206e-05,
+      "grad_norm": 346.0,
+      "learning_rate": 2.97e-05,
+      "loss": 7.3853,
+      "step": 100
+    },
+    {
+      "epoch": 1.5420083733858326e-05,
+      "grad_norm": 167.0,
+      "learning_rate": 3.2699999999999995e-05,
+      "loss": 6.8838,
+      "step": 110
+    },
+    {
+      "epoch": 1.6821909527845447e-05,
+      "grad_norm": 292.0,
+      "learning_rate": 3.5699999999999994e-05,
+      "loss": 6.7212,
+      "step": 120
+    },
+    {
+      "epoch": 1.822373532183257e-05,
+      "grad_norm": 253.0,
+      "learning_rate": 3.87e-05,
+      "loss": 6.3265,
+      "step": 130
+    },
+    {
+      "epoch": 1.9625561115819688e-05,
+      "grad_norm": 133.0,
+      "learning_rate": 4.17e-05,
+      "loss": 6.2292,
+      "step": 140
+    },
+    {
+      "epoch": 2.102738690980681e-05,
+      "grad_norm": 79.5,
+      "learning_rate": 4.4699999999999996e-05,
+      "loss": 5.1204,
+      "step": 150
+    },
+    {
+      "epoch": 2.2429212703793928e-05,
+      "grad_norm": 38.5,
+      "learning_rate": 4.7699999999999994e-05,
+      "loss": 4.9199,
+      "step": 160
+    },
+    {
+      "epoch": 2.383103849778105e-05,
+      "grad_norm": 27.875,
+      "learning_rate": 5.07e-05,
+      "loss": 4.1897,
+      "step": 170
+    },
+    {
+      "epoch": 2.523286429176817e-05,
+      "grad_norm": 23.375,
+      "learning_rate": 5.369999999999999e-05,
+      "loss": 3.1582,
+      "step": 180
+    },
+    {
+      "epoch": 2.663469008575529e-05,
+      "grad_norm": 11.75,
+      "learning_rate": 5.6699999999999996e-05,
+      "loss": 2.0433,
+      "step": 190
+    },
+    {
+      "epoch": 2.803651587974241e-05,
+      "grad_norm": 4.625,
+      "learning_rate": 5.97e-05,
+      "loss": 1.3245,
+      "step": 200
+    },
+    {
+      "epoch": 2.9438341673729533e-05,
+      "grad_norm": 3.1875,
+      "learning_rate": 6.269999999999999e-05,
+      "loss": 1.1166,
+      "step": 210
+    },
+    {
+      "epoch": 3.084016746771665e-05,
+      "grad_norm": 1.40625,
+      "learning_rate": 6.57e-05,
+      "loss": 0.884,
+      "step": 220
+    },
+    {
+      "epoch": 3.224199326170377e-05,
+      "grad_norm": 2.40625,
+      "learning_rate": 6.87e-05,
+      "loss": 0.8963,
+      "step": 230
+    },
+    {
+      "epoch": 3.3643819055690895e-05,
+      "grad_norm": 1.8125,
+      "learning_rate": 7.17e-05,
+      "loss": 0.9096,
+      "step": 240
+    },
+    {
+      "epoch": 3.504564484967802e-05,
+      "grad_norm": 0.9921875,
+      "learning_rate": 7.47e-05,
+      "loss": 0.8575,
+      "step": 250
+    },
+    {
+      "epoch": 3.644747064366514e-05,
+      "grad_norm": 1.4375,
+      "learning_rate": 7.769999999999999e-05,
+      "loss": 0.791,
+      "step": 260
+    },
+    {
+      "epoch": 3.7849296437652253e-05,
+      "grad_norm": 1.046875,
+      "learning_rate": 8.07e-05,
+      "loss": 0.8676,
+      "step": 270
+    },
+    {
+      "epoch": 3.9251122231639375e-05,
+      "grad_norm": 1.5859375,
+      "learning_rate": 8.37e-05,
+      "loss": 0.8645,
+      "step": 280
+    },
+    {
+      "epoch": 4.06529480256265e-05,
+      "grad_norm": 0.84765625,
+      "learning_rate": 8.669999999999998e-05,
+      "loss": 0.7618,
+      "step": 290
+    },
+    {
+      "epoch": 4.205477381961362e-05,
+      "grad_norm": 1.1640625,
+      "learning_rate": 8.969999999999998e-05,
+      "loss": 0.8603,
+      "step": 300
+    },
+    {
+      "epoch": 4.345659961360074e-05,
+      "grad_norm": 0.64453125,
+      "learning_rate": 9.269999999999999e-05,
+      "loss": 0.7988,
+      "step": 310
+    },
+    {
+      "epoch": 4.4858425407587855e-05,
+      "grad_norm": 0.94921875,
+      "learning_rate": 9.57e-05,
+      "loss": 0.8073,
+      "step": 320
+    },
+    {
+      "epoch": 4.626025120157498e-05,
+      "grad_norm": 0.76171875,
+      "learning_rate": 9.87e-05,
+      "loss": 0.8298,
+      "step": 330
+    },
+    {
+      "epoch": 4.76620769955621e-05,
+      "grad_norm": 1.5390625,
+      "learning_rate": 0.00010169999999999999,
+      "loss": 0.6355,
+      "step": 340
+    },
+    {
+      "epoch": 4.906390278954922e-05,
+      "grad_norm": 1.7890625,
+      "learning_rate": 0.00010469999999999998,
+      "loss": 0.848,
+      "step": 350
+    },
+    {
+      "epoch": 5.046572858353634e-05,
+      "grad_norm": 0.62109375,
+      "learning_rate": 0.00010769999999999999,
+      "loss": 0.5291,
+      "step": 360
+    },
+    {
+      "epoch": 5.1867554377523464e-05,
+      "grad_norm": 1.578125,
+      "learning_rate": 0.0001107,
+      "loss": 0.6325,
+      "step": 370
+    },
+    {
+      "epoch": 5.326938017151058e-05,
+      "grad_norm": 0.8828125,
+      "learning_rate": 0.00011369999999999999,
+      "loss": 0.7049,
+      "step": 380
+    },
+    {
+      "epoch": 5.46712059654977e-05,
+      "grad_norm": 1.109375,
+      "learning_rate": 0.00011669999999999999,
+      "loss": 0.7385,
+      "step": 390
+    },
+    {
+      "epoch": 5.607303175948482e-05,
+      "grad_norm": 1.0078125,
+      "learning_rate": 0.0001197,
+      "loss": 0.7195,
+      "step": 400
+    },
+    {
+      "epoch": 5.7474857553471944e-05,
+      "grad_norm": 1.0390625,
+      "learning_rate": 0.00012269999999999997,
+      "loss": 0.7844,
+      "step": 410
+    },
+    {
+      "epoch": 5.8876683347459066e-05,
+      "grad_norm": 0.69140625,
+      "learning_rate": 0.0001257,
+      "loss": 0.677,
+      "step": 420
+    },
+    {
+      "epoch": 6.027850914144619e-05,
+      "grad_norm": 1.703125,
+      "learning_rate": 0.00012869999999999998,
+      "loss": 0.704,
+      "step": 430
+    },
+    {
+      "epoch": 6.16803349354333e-05,
+      "grad_norm": 1.21875,
+      "learning_rate": 0.00013169999999999998,
+      "loss": 0.5956,
+      "step": 440
+    },
+    {
+      "epoch": 6.308216072942043e-05,
+      "grad_norm": 1.015625,
+      "learning_rate": 0.0001347,
+      "loss": 0.769,
+      "step": 450
+    },
+    {
+      "epoch": 6.448398652340755e-05,
+      "grad_norm": 1.1484375,
+      "learning_rate": 0.00013769999999999999,
+      "loss": 0.6908,
+      "step": 460
+    },
+    {
+      "epoch": 6.588581231739466e-05,
+      "grad_norm": 0.90625,
+      "learning_rate": 0.00014069999999999998,
+      "loss": 0.6306,
+      "step": 470
+    },
+    {
+      "epoch": 6.728763811138179e-05,
+      "grad_norm": 0.9765625,
+      "learning_rate": 0.00014369999999999997,
+      "loss": 0.6421,
+      "step": 480
+    },
+    {
+      "epoch": 6.86894639053689e-05,
+      "grad_norm": 1.71875,
+      "learning_rate": 0.0001467,
+      "loss": 0.7901,
+      "step": 490
+    },
+    {
+      "epoch": 7.009128969935603e-05,
+      "grad_norm": 1.6640625,
+      "learning_rate": 0.00014969999999999998,
+      "loss": 0.6785,
+      "step": 500
+    },
+    {
+      "epoch": 7.009128969935603e-05,
+      "eval_loss": 0.27128955721855164,
+      "eval_runtime": 113.5365,
+      "eval_samples_per_second": 201.195,
+      "eval_steps_per_second": 1.577,
+      "step": 500
+    },
+    {
+      "epoch": 7.149311549334315e-05,
+      "grad_norm": 1.0859375,
+      "learning_rate": 0.0001527,
+      "loss": 0.6427,
+      "step": 510
+    },
+    {
+      "epoch": 7.289494128733028e-05,
+      "grad_norm": 0.51953125,
+      "learning_rate": 0.0001557,
+      "loss": 0.7077,
+      "step": 520
+    },
+    {
+      "epoch": 7.429676708131739e-05,
+      "grad_norm": 1.3203125,
+      "learning_rate": 0.00015869999999999998,
+      "loss": 0.6865,
+      "step": 530
+    },
+    {
+      "epoch": 7.569859287530451e-05,
+      "grad_norm": 1.140625,
+      "learning_rate": 0.0001617,
+      "loss": 0.7239,
+      "step": 540
+    },
+    {
+      "epoch": 7.710041866929164e-05,
+      "grad_norm": 0.73046875,
+      "learning_rate": 0.0001647,
+      "loss": 0.6667,
+      "step": 550
+    },
+    {
+      "epoch": 7.850224446327875e-05,
+      "grad_norm": 1.34375,
+      "learning_rate": 0.0001677,
+      "loss": 0.8085,
+      "step": 560
+    },
+    {
+      "epoch": 7.990407025726588e-05,
+      "grad_norm": 1.3046875,
+      "learning_rate": 0.00017069999999999998,
+      "loss": 0.7019,
+      "step": 570
+    },
+    {
+      "epoch": 8.1305896051253e-05,
+      "grad_norm": 1.796875,
+      "learning_rate": 0.00017369999999999997,
+      "loss": 0.7378,
+      "step": 580
+    },
+    {
+      "epoch": 8.270772184524011e-05,
+      "grad_norm": 0.98828125,
+      "learning_rate": 0.00017669999999999999,
+      "loss": 0.6776,
+      "step": 590
+    },
+    {
+      "epoch": 8.410954763922724e-05,
+      "grad_norm": 1.2578125,
+      "learning_rate": 0.00017969999999999998,
+      "loss": 0.6166,
+      "step": 600
+    },
+    {
+      "epoch": 8.551137343321435e-05,
+      "grad_norm": 0.75,
+      "learning_rate": 0.00018269999999999997,
+      "loss": 0.5741,
+      "step": 610
+    },
+    {
+      "epoch": 8.691319922720148e-05,
+      "grad_norm": 1.1484375,
+      "learning_rate": 0.0001857,
+      "loss": 0.7397,
+      "step": 620
+    },
+    {
+      "epoch": 8.83150250211886e-05,
+      "grad_norm": 1.265625,
+      "learning_rate": 0.00018869999999999998,
+      "loss": 0.6753,
+      "step": 630
+    },
+    {
+      "epoch": 8.971685081517571e-05,
+      "grad_norm": 0.9375,
+      "learning_rate": 0.0001917,
+      "loss": 0.689,
+      "step": 640
+    },
+    {
+      "epoch": 9.111867660916284e-05,
+      "grad_norm": 1.1484375,
+      "learning_rate": 0.0001947,
+      "loss": 0.5332,
+      "step": 650
+    },
+    {
+      "epoch": 9.252050240314995e-05,
+      "grad_norm": 1.234375,
+      "learning_rate": 0.00019769999999999998,
+      "loss": 0.7178,
+      "step": 660
+    },
+    {
+      "epoch": 9.392232819713708e-05,
+      "grad_norm": 0.8203125,
+      "learning_rate": 0.0002007,
+      "loss": 0.5924,
+      "step": 670
+    },
+    {
+      "epoch": 9.53241539911242e-05,
+      "grad_norm": 1.5,
+      "learning_rate": 0.0002037,
+      "loss": 0.6734,
+      "step": 680
+    },
+    {
+      "epoch": 9.672597978511133e-05,
+      "grad_norm": 0.9921875,
+      "learning_rate": 0.00020669999999999996,
+      "loss": 0.6536,
+      "step": 690
+    },
+    {
+      "epoch": 9.812780557909844e-05,
+      "grad_norm": 0.62890625,
+      "learning_rate": 0.00020969999999999997,
+      "loss": 0.5084,
+      "step": 700
+    },
+    {
+      "epoch": 9.952963137308556e-05,
+      "grad_norm": 0.890625,
+      "learning_rate": 0.00021269999999999997,
+      "loss": 0.7349,
+      "step": 710
+    },
+    {
+      "epoch": 0.00010093145716707268,
+      "grad_norm": 0.6875,
+      "learning_rate": 0.00021569999999999998,
+      "loss": 0.519,
+      "step": 720
+    },
+    {
+      "epoch": 0.0001023332829610598,
+      "grad_norm": 1.3671875,
+      "learning_rate": 0.00021869999999999998,
+      "loss": 0.6678,
+      "step": 730
+    },
+    {
+      "epoch": 0.00010373510875504693,
+      "grad_norm": 0.5625,
+      "learning_rate": 0.00022169999999999997,
+      "loss": 0.6112,
+      "step": 740
+    },
+    {
+      "epoch": 0.00010513693454903404,
+      "grad_norm": 1.09375,
+      "learning_rate": 0.0002247,
+      "loss": 0.6505,
+      "step": 750
+    },
+    {
+      "epoch": 0.00010653876034302116,
+      "grad_norm": 1.0234375,
+      "learning_rate": 0.00022769999999999998,
+      "loss": 0.6256,
+      "step": 760
+    },
+    {
+      "epoch": 0.00010794058613700829,
+      "grad_norm": 0.78125,
+      "learning_rate": 0.0002307,
+      "loss": 0.6422,
+      "step": 770
+    },
+    {
+      "epoch": 0.0001093424119309954,
+      "grad_norm": 0.72265625,
+      "learning_rate": 0.0002337,
+      "loss": 0.6357,
+      "step": 780
+    },
+    {
+      "epoch": 0.00011074423772498253,
+      "grad_norm": 1.25,
+      "learning_rate": 0.00023669999999999998,
+      "loss": 0.6263,
+      "step": 790
+    },
+    {
+      "epoch": 0.00011214606351896965,
+      "grad_norm": 1.1171875,
+      "learning_rate": 0.0002397,
+      "loss": 0.6531,
+      "step": 800
+    },
+    {
+      "epoch": 0.00011354788931295677,
+      "grad_norm": 1.1875,
+      "learning_rate": 0.0002427,
+      "loss": 0.5806,
+      "step": 810
+    },
+    {
+      "epoch": 0.00011494971510694389,
+      "grad_norm": 1.1953125,
+      "learning_rate": 0.00024569999999999995,
+      "loss": 0.5925,
+      "step": 820
+    },
+    {
+      "epoch": 0.000116351540900931,
+      "grad_norm": 1.1640625,
+      "learning_rate": 0.0002487,
+      "loss": 0.5465,
+      "step": 830
+    },
+    {
+      "epoch": 0.00011775336669491813,
+      "grad_norm": 0.66796875,
+      "learning_rate": 0.0002517,
+      "loss": 0.7182,
+      "step": 840
+    },
+    {
+      "epoch": 0.00011915519248890525,
+      "grad_norm": 1.3203125,
+      "learning_rate": 0.00025469999999999996,
+      "loss": 0.6767,
+      "step": 850
+    },
+    {
+      "epoch": 0.00012055701828289238,
+      "grad_norm": 1.2890625,
+      "learning_rate": 0.0002577,
+      "loss": 0.6315,
+      "step": 860
+    },
+    {
+      "epoch": 0.00012195884407687949,
+      "grad_norm": 0.71484375,
+      "learning_rate": 0.0002607,
+      "loss": 0.6724,
+      "step": 870
+    },
+    {
+      "epoch": 0.0001233606698708666,
+      "grad_norm": 0.9765625,
+      "learning_rate": 0.00026369999999999996,
+      "loss": 0.6528,
+      "step": 880
+    },
+    {
+      "epoch": 0.00012476249566485372,
+      "grad_norm": 1.15625,
+      "learning_rate": 0.0002667,
+      "loss": 0.5295,
+      "step": 890
+    },
+    {
+      "epoch": 0.00012616432145884086,
+      "grad_norm": 0.890625,
+      "learning_rate": 0.0002697,
+      "loss": 0.4982,
+      "step": 900
+    },
+    {
+      "epoch": 0.00012756614725282798,
+      "grad_norm": 1.2734375,
+      "learning_rate": 0.00027269999999999996,
+      "loss": 0.6562,
+      "step": 910
+    },
+    {
+      "epoch": 0.0001289679730468151,
+      "grad_norm": 0.8515625,
+      "learning_rate": 0.0002757,
+      "loss": 0.675,
+      "step": 920
+    },
+    {
+      "epoch": 0.0001303697988408022,
+      "grad_norm": 1.2890625,
+      "learning_rate": 0.0002787,
+      "loss": 0.6327,
+      "step": 930
+    },
+    {
+      "epoch": 0.00013177162463478932,
+      "grad_norm": 1.203125,
+      "learning_rate": 0.00028169999999999996,
+      "loss": 0.7426,
+      "step": 940
+    },
+    {
+      "epoch": 0.00013317345042877646,
+      "grad_norm": 0.94921875,
+      "learning_rate": 0.0002847,
+      "loss": 0.5876,
+      "step": 950
+    },
+    {
+      "epoch": 0.00013457527622276358,
+      "grad_norm": 0.625,
+      "learning_rate": 0.00028769999999999995,
+      "loss": 0.6968,
+      "step": 960
+    },
+    {
+      "epoch": 0.0001359771020167507,
+      "grad_norm": 0.54296875,
+      "learning_rate": 0.00029069999999999996,
+      "loss": 0.6377,
+      "step": 970
+    },
+    {
+      "epoch": 0.0001373789278107378,
+      "grad_norm": 0.78125,
+      "learning_rate": 0.0002937,
+      "loss": 0.515,
+      "step": 980
+    },
+    {
+      "epoch": 0.00013878075360472492,
+      "grad_norm": 0.95703125,
+      "learning_rate": 0.00029669999999999995,
+      "loss": 0.6966,
+      "step": 990
+    },
+    {
+      "epoch": 0.00014018257939871207,
+      "grad_norm": 0.376953125,
+      "learning_rate": 0.00029969999999999997,
+      "loss": 0.5426,
+      "step": 1000
+    },
+    {
+      "epoch": 0.00014018257939871207,
+      "eval_loss": 0.2417471706867218,
+      "eval_runtime": 111.9274,
+      "eval_samples_per_second": 204.088,
+      "eval_steps_per_second": 1.599,
+      "step": 1000
+    },
+    {
+      "epoch": 0.00014158440519269918,
+      "grad_norm": 1.1015625,
+      "learning_rate": 0.0002999999999988214,
+      "loss": 0.579,
+      "step": 1010
+    },
+    {
+      "epoch": 0.0001429862309866863,
+      "grad_norm": 1.0625,
+      "learning_rate": 0.0002999999999947473,
+      "loss": 0.5958,
+      "step": 1020
+    },
+    {
+      "epoch": 0.0001443880567806734,
+      "grad_norm": 1.1953125,
+      "learning_rate": 0.0002999999999877632,
+      "loss": 0.5812,
+      "step": 1030
+    },
+    {
+      "epoch": 0.00014578988257466055,
+      "grad_norm": 1.1171875,
+      "learning_rate": 0.000299999999977869,
+      "loss": 0.6251,
+      "step": 1040
+    },
+    {
+      "epoch": 0.00014719170836864767,
+      "grad_norm": 1.125,
+      "learning_rate": 0.0002999999999650648,
+      "loss": 0.6301,
+      "step": 1050
+    },
+    {
+      "epoch": 0.00014859353416263478,
+      "grad_norm": 0.71484375,
+      "learning_rate": 0.0002999999999493505,
+      "loss": 0.6041,
+      "step": 1060
+    },
+    {
+      "epoch": 0.0001499953599566219,
+      "grad_norm": 0.73828125,
+      "learning_rate": 0.0002999999999307262,
+      "loss": 0.5958,
+      "step": 1070
+    },
+    {
+      "epoch": 0.00015139718575060901,
+      "grad_norm": 0.5625,
+      "learning_rate": 0.00029999999990919175,
+      "loss": 0.5158,
+      "step": 1080
+    },
+    {
+      "epoch": 0.00015279901154459616,
+      "grad_norm": 0.93359375,
+      "learning_rate": 0.00029999999988474733,
+      "loss": 0.5815,
+      "step": 1090
+    },
+    {
+      "epoch": 0.00015420083733858327,
+      "grad_norm": 0.84765625,
+      "learning_rate": 0.00029999999985739286,
+      "loss": 0.5338,
+      "step": 1100
+    },
+    {
+      "epoch": 0.00015560266313257039,
+      "grad_norm": 0.546875,
+      "learning_rate": 0.0002999999998271283,
+      "loss": 0.3756,
+      "step": 1110
+    },
+    {
+      "epoch": 0.0001570044889265575,
+      "grad_norm": 0.86328125,
+      "learning_rate": 0.0002999999997939537,
+      "loss": 0.5302,
+      "step": 1120
+    },
+    {
+      "epoch": 0.00015840631472054462,
+      "grad_norm": 0.625,
+      "learning_rate": 0.000299999999757869,
+      "loss": 0.6399,
+      "step": 1130
+    },
+    {
+      "epoch": 0.00015980814051453176,
+      "grad_norm": 0.6640625,
+      "learning_rate": 0.0002999999997188743,
+      "loss": 0.5165,
+      "step": 1140
+    },
+    {
+      "epoch": 0.00016120996630851887,
+      "grad_norm": 1.0390625,
+      "learning_rate": 0.0002999999996769696,
+      "loss": 0.611,
+      "step": 1150
+    },
+    {
+      "epoch": 0.000162611792102506,
+      "grad_norm": 0.703125,
+      "learning_rate": 0.00029999999963215473,
+      "loss": 0.4649,
+      "step": 1160
+    },
+    {
+      "epoch": 0.0001640136178964931,
+      "grad_norm": 0.474609375,
+      "learning_rate": 0.0002999999995844299,
+      "loss": 0.5931,
+      "step": 1170
+    },
+    {
+      "epoch": 0.00016541544369048022,
+      "grad_norm": 1.515625,
+      "learning_rate": 0.000299999999533795,
+      "loss": 0.605,
+      "step": 1180
+    },
+    {
+      "epoch": 0.00016681726948446736,
+      "grad_norm": 0.8671875,
+      "learning_rate": 0.00029999999948025,
+      "loss": 0.5483,
+      "step": 1190
+    },
+    {
+      "epoch": 0.00016821909527845447,
+      "grad_norm": 0.796875,
+      "learning_rate": 0.000299999999423795,
+      "loss": 0.5798,
+      "step": 1200
+    },
+    {
+      "epoch": 0.0001696209210724416,
+      "grad_norm": 0.93359375,
+      "learning_rate": 0.0002999999993644299,
+      "loss": 0.5735,
+      "step": 1210
+    },
+    {
+      "epoch": 0.0001710227468664287,
+      "grad_norm": 0.7734375,
+      "learning_rate": 0.00029999999930215476,
+      "loss": 0.5838,
+      "step": 1220
+    },
+    {
+      "epoch": 0.00017242457266041582,
+      "grad_norm": 0.44921875,
+      "learning_rate": 0.0002999999992369696,
+      "loss": 0.495,
+      "step": 1230
+    },
+    {
+      "epoch": 0.00017382639845440296,
+      "grad_norm": 0.8984375,
+      "learning_rate": 0.00029999999916887433,
+      "loss": 0.576,
+      "step": 1240
+    },
+    {
+      "epoch": 0.00017522822424839008,
+      "grad_norm": 1.015625,
+      "learning_rate": 0.00029999999909786906,
+      "loss": 0.5682,
+      "step": 1250
+    },
+    {
+      "epoch": 0.0001766300500423772,
+      "grad_norm": 0.91796875,
+      "learning_rate": 0.00029999999902395374,
+      "loss": 0.609,
+      "step": 1260
+    },
+    {
+      "epoch": 0.0001780318758363643,
+      "grad_norm": 0.9609375,
+      "learning_rate": 0.0002999999989471283,
+      "loss": 0.5911,
+      "step": 1270
+    },
+    {
+      "epoch": 0.00017943370163035142,
+      "grad_norm": 1.078125,
+      "learning_rate": 0.0002999999988673929,
+      "loss": 0.6365,
+      "step": 1280
+    },
+    {
+      "epoch": 0.00018083552742433856,
+      "grad_norm": 0.67578125,
+      "learning_rate": 0.00029999999878474736,
+      "loss": 0.4741,
+      "step": 1290
+    },
+    {
+      "epoch": 0.00018223735321832568,
+      "grad_norm": 0.8828125,
+      "learning_rate": 0.0002999999986991918,
+      "loss": 0.6288,
+      "step": 1300
+    },
+    {
+      "epoch": 0.0001836391790123128,
+      "grad_norm": 0.5,
+      "learning_rate": 0.0002999999986107262,
+      "loss": 0.4648,
+      "step": 1310
+    },
+    {
+      "epoch": 0.0001850410048062999,
+      "grad_norm": 0.61328125,
+      "learning_rate": 0.00029999999851935055,
+      "loss": 0.5761,
+      "step": 1320
+    },
+    {
+      "epoch": 0.00018644283060028705,
+      "grad_norm": 1.4375,
+      "learning_rate": 0.00029999999842506486,
+      "loss": 0.5775,
+      "step": 1330
+    },
+    {
+      "epoch": 0.00018784465639427417,
+      "grad_norm": 0.96875,
+      "learning_rate": 0.00029999999832786906,
+      "loss": 0.5382,
+      "step": 1340
+    },
+    {
+      "epoch": 0.00018924648218826128,
+      "grad_norm": 1.078125,
+      "learning_rate": 0.00029999999822776327,
+      "loss": 0.5922,
+      "step": 1350
+    },
+    {
+      "epoch": 0.0001906483079822484,
+      "grad_norm": 0.62890625,
+      "learning_rate": 0.0002999999981247474,
+      "loss": 0.4537,
+      "step": 1360
+    },
+    {
+      "epoch": 0.0001920501337762355,
+      "grad_norm": 1.328125,
+      "learning_rate": 0.00029999999801882146,
+      "loss": 0.5119,
+      "step": 1370
+    },
+    {
+      "epoch": 0.00019345195957022265,
+      "grad_norm": 0.86328125,
+      "learning_rate": 0.0002999999979099855,
+      "loss": 0.478,
+      "step": 1380
+    },
+    {
+      "epoch": 0.00019485378536420977,
+      "grad_norm": 1.9296875,
+      "learning_rate": 0.00029999999779823944,
+      "loss": 0.4734,
+      "step": 1390
+    },
+    {
+      "epoch": 0.00019625561115819688,
+      "grad_norm": 1.2421875,
+      "learning_rate": 0.0002999999976835834,
+      "loss": 0.4924,
+      "step": 1400
+    },
+    {
+      "epoch": 0.000197657436952184,
+      "grad_norm": 2.046875,
+      "learning_rate": 0.00029999999756601727,
+      "loss": 0.5205,
+      "step": 1410
+    },
+    {
+      "epoch": 0.0001990592627461711,
+      "grad_norm": 1.3984375,
+      "learning_rate": 0.0002999999974455411,
+      "loss": 0.5122,
+      "step": 1420
+    },
+    {
+      "epoch": 0.00020046108854015825,
+      "grad_norm": 0.77734375,
+      "learning_rate": 0.00029999999732215483,
+      "loss": 0.5492,
+      "step": 1430
+    },
+    {
+      "epoch": 0.00020186291433414537,
+      "grad_norm": 0.5234375,
+      "learning_rate": 0.00029999999719585856,
+      "loss": 0.5266,
+      "step": 1440
+    },
+    {
+      "epoch": 0.00020326474012813248,
+      "grad_norm": 0.85546875,
+      "learning_rate": 0.00029999999706665224,
+      "loss": 0.5307,
+      "step": 1450
+    },
+    {
+      "epoch": 0.0002046665659221196,
+      "grad_norm": 0.53515625,
+      "learning_rate": 0.0002999999969345358,
+      "loss": 0.6882,
+      "step": 1460
+    },
+    {
+      "epoch": 0.00020606839171610671,
+      "grad_norm": 0.84375,
+      "learning_rate": 0.0002999999967995094,
+      "loss": 0.5092,
+      "step": 1470
+    },
+    {
+      "epoch": 0.00020747021751009386,
+      "grad_norm": 0.7734375,
+      "learning_rate": 0.0002999999966615729,
+      "loss": 0.4554,
+      "step": 1480
+    },
+    {
+      "epoch": 0.00020887204330408097,
+      "grad_norm": 0.84765625,
+      "learning_rate": 0.0002999999965207263,
+      "loss": 0.5944,
+      "step": 1490
+    },
+    {
+      "epoch": 0.0002102738690980681,
+      "grad_norm": 0.97265625,
+      "learning_rate": 0.0002999999963769697,
+      "loss": 0.5038,
+      "step": 1500
+    },
+    {
+      "epoch": 0.0002102738690980681,
+      "eval_loss": 0.22094833850860596,
+      "eval_runtime": 112.0927,
+      "eval_samples_per_second": 203.787,
+      "eval_steps_per_second": 1.597,
+      "step": 1500
+    },
+    {
+      "epoch": 0.0002116756948920552,
+      "grad_norm": 0.8125,
+      "learning_rate": 0.00029999999623030307,
+      "loss": 0.5026,
+      "step": 1510
+    },
+    {
+      "epoch": 0.00021307752068604232,
+      "grad_norm": 0.390625,
+      "learning_rate": 0.0002999999960807263,
+      "loss": 0.5545,
+      "step": 1520
+    },
+    {
+      "epoch": 0.00021447934648002946,
+      "grad_norm": 1.109375,
+      "learning_rate": 0.0002999999959282396,
+      "loss": 0.5133,
+      "step": 1530
+    },
+    {
+      "epoch": 0.00021588117227401657,
+      "grad_norm": 1.2578125,
+      "learning_rate": 0.0002999999957728428,
+      "loss": 0.5241,
+      "step": 1540
+    },
+    {
+      "epoch": 0.0002172829980680037,
+      "grad_norm": 0.60546875,
+      "learning_rate": 0.00029999999561453587,
+      "loss": 0.4591,
+      "step": 1550
+    },
+    {
+      "epoch": 0.0002186848238619908,
+      "grad_norm": 0.84765625,
+      "learning_rate": 0.00029999999545331896,
+      "loss": 0.5348,
+      "step": 1560
+    },
+    {
+      "epoch": 0.00022008664965597792,
+      "grad_norm": 1.171875,
+      "learning_rate": 0.00029999999528919195,
+      "loss": 0.5408,
+      "step": 1570
+    },
+    {
+      "epoch": 0.00022148847544996506,
+      "grad_norm": 1.1171875,
+      "learning_rate": 0.000299999995122155,
+      "loss": 0.616,
+      "step": 1580
+    },
+    {
+      "epoch": 0.00022289030124395218,
+      "grad_norm": 1.53125,
+      "learning_rate": 0.0002999999949522079,
+      "loss": 0.585,
+      "step": 1590
+    },
+    {
+      "epoch": 0.0002242921270379393,
+      "grad_norm": 0.373046875,
+      "learning_rate": 0.0002999999947793507,
+      "loss": 0.4513,
+      "step": 1600
+    },
+    {
+      "epoch": 0.0002256939528319264,
+      "grad_norm": 1.0078125,
+      "learning_rate": 0.0002999999946035836,
+      "loss": 0.5646,
+      "step": 1610
+    },
+    {
+      "epoch": 0.00022709577862591355,
+      "grad_norm": 0.6328125,
+      "learning_rate": 0.0002999999944249063,
+      "loss": 0.4162,
+      "step": 1620
+    },
+    {
+      "epoch": 0.00022849760441990066,
+      "grad_norm": 0.462890625,
+      "learning_rate": 0.00029999999424331903,
+      "loss": 0.5288,
+      "step": 1630
+    },
+    {
+      "epoch": 0.00022989943021388778,
+      "grad_norm": 0.26953125,
+      "learning_rate": 0.0002999999940588217,
+      "loss": 0.5398,
+      "step": 1640
+    },
+    {
+      "epoch": 0.0002313012560078749,
+      "grad_norm": 1.109375,
+      "learning_rate": 0.00029999999387141427,
+      "loss": 0.5167,
+      "step": 1650
+    },
+    {
+      "epoch": 0.000232703081801862,
+      "grad_norm": 0.765625,
+      "learning_rate": 0.00029999999368109684,
+      "loss": 0.5953,
+      "step": 1660
+    },
+    {
+      "epoch": 0.00023410490759584915,
+      "grad_norm": 0.66796875,
+      "learning_rate": 0.00029999999348786935,
+      "loss": 0.5765,
+      "step": 1670
+    },
+    {
+      "epoch": 0.00023550673338983626,
+      "grad_norm": 0.890625,
+      "learning_rate": 0.00029999999329173176,
+      "loss": 0.5683,
+      "step": 1680
+    },
+    {
+      "epoch": 0.00023690855918382338,
+      "grad_norm": 0.8203125,
+      "learning_rate": 0.00029999999309268416,
+      "loss": 0.4832,
+      "step": 1690
+    },
+    {
+      "epoch": 0.0002383103849778105,
+      "grad_norm": 0.369140625,
+      "learning_rate": 0.0002999999928907265,
+      "loss": 0.4763,
+      "step": 1700
+    },
+    {
+      "epoch": 0.0002397122107717976,
+      "grad_norm": 0.72265625,
+      "learning_rate": 0.0002999999926858588,
+      "loss": 0.5735,
+      "step": 1710
+    },
+    {
+      "epoch": 0.00024111403656578475,
+      "grad_norm": 1.078125,
+      "learning_rate": 0.000299999992478081,
+      "loss": 0.5352,
+      "step": 1720
+    },
+    {
+      "epoch": 0.00024251586235977187,
+      "grad_norm": 0.63671875,
+      "learning_rate": 0.0002999999922673932,
+      "loss": 0.5497,
+      "step": 1730
+    },
+    {
+      "epoch": 0.00024391768815375898,
+      "grad_norm": 0.310546875,
+      "learning_rate": 0.00029999999205379535,
+      "loss": 0.5335,
+      "step": 1740
+    },
+    {
+      "epoch": 0.0002453195139477461,
+      "grad_norm": 1.03125,
+      "learning_rate": 0.0002999999918372874,
+      "loss": 0.476,
+      "step": 1750
+    },
+    {
+      "epoch": 0.0002467213397417332,
+      "grad_norm": 0.86328125,
+      "learning_rate": 0.0002999999916178694,
+      "loss": 0.5105,
+      "step": 1760
+    },
+    {
+      "epoch": 0.0002481231655357203,
+      "grad_norm": 0.68359375,
+      "learning_rate": 0.0002999999913955414,
+      "loss": 0.5929,
+      "step": 1770
+    },
+    {
+      "epoch": 0.00024952499132970744,
+      "grad_norm": 0.419921875,
+      "learning_rate": 0.00029999999117030334,
+      "loss": 0.4957,
+      "step": 1780
+    },
+    {
+      "epoch": 0.0002509268171236946,
+      "grad_norm": 0.6328125,
+      "learning_rate": 0.0002999999909421552,
+      "loss": 0.5539,
+      "step": 1790
+    },
+    {
+      "epoch": 0.0002523286429176817,
+      "grad_norm": 0.96484375,
+      "learning_rate": 0.000299999990711097,
+      "loss": 0.4833,
+      "step": 1800
+    },
+    {
+      "epoch": 0.00025373046871166884,
+      "grad_norm": 0.55078125,
+      "learning_rate": 0.00029999999047712877,
+      "loss": 0.5147,
+      "step": 1810
+    },
+    {
+      "epoch": 0.00025513229450565596,
+      "grad_norm": 1.3671875,
+      "learning_rate": 0.0002999999902402505,
+      "loss": 0.4975,
+      "step": 1820
+    },
+    {
+      "epoch": 0.00025653412029964307,
+      "grad_norm": 0.6171875,
+      "learning_rate": 0.0002999999900004621,
+      "loss": 0.5936,
+      "step": 1830
+    },
+    {
+      "epoch": 0.0002579359460936302,
+      "grad_norm": 1.25,
+      "learning_rate": 0.0002999999897577637,
+      "loss": 0.5696,
+      "step": 1840
+    },
+    {
+      "epoch": 0.0002593377718876173,
+      "grad_norm": 0.8359375,
+      "learning_rate": 0.0002999999895121553,
+      "loss": 0.507,
+      "step": 1850
+    },
+    {
+      "epoch": 0.0002607395976816044,
+      "grad_norm": 0.37890625,
+      "learning_rate": 0.0002999999892636368,
+      "loss": 0.4285,
+      "step": 1860
+    },
+    {
+      "epoch": 0.00026214142347559153,
+      "grad_norm": 1.1328125,
+      "learning_rate": 0.00029999998901220824,
+      "loss": 0.5559,
+      "step": 1870
+    },
+    {
+      "epoch": 0.00026354324926957865,
+      "grad_norm": 0.57421875,
+      "learning_rate": 0.0002999999887578696,
+      "loss": 0.5271,
+      "step": 1880
+    },
+    {
+      "epoch": 0.0002649450750635658,
+      "grad_norm": 0.484375,
+      "learning_rate": 0.00029999998850062094,
+      "loss": 0.4832,
+      "step": 1890
+    },
+    {
+      "epoch": 0.00026634690085755293,
+      "grad_norm": 1.328125,
+      "learning_rate": 0.00029999998824046224,
+      "loss": 0.6106,
+      "step": 1900
+    },
+    {
+      "epoch": 0.00026774872665154004,
+      "grad_norm": 1.828125,
+      "learning_rate": 0.0002999999879773935,
+      "loss": 0.5356,
+      "step": 1910
+    },
+    {
+      "epoch": 0.00026915055244552716,
+      "grad_norm": 1.2265625,
+      "learning_rate": 0.0002999999877114147,
+      "loss": 0.5922,
+      "step": 1920
+    },
+    {
+      "epoch": 0.0002705523782395143,
+      "grad_norm": 1.0234375,
+      "learning_rate": 0.00029999998744252576,
+      "loss": 0.4454,
+      "step": 1930
+    },
+    {
+      "epoch": 0.0002719542040335014,
+      "grad_norm": 0.84765625,
+      "learning_rate": 0.00029999998717072685,
+      "loss": 0.4555,
+      "step": 1940
+    },
+    {
+      "epoch": 0.0002733560298274885,
+      "grad_norm": 0.79296875,
+      "learning_rate": 0.0002999999868960179,
+      "loss": 0.5104,
+      "step": 1950
+    },
+    {
+      "epoch": 0.0002747578556214756,
+      "grad_norm": 0.9140625,
+      "learning_rate": 0.00029999998661839886,
+      "loss": 0.5298,
+      "step": 1960
+    },
+    {
+      "epoch": 0.00027615968141546273,
+      "grad_norm": 0.3828125,
+      "learning_rate": 0.0002999999863378698,
+      "loss": 0.511,
+      "step": 1970
+    },
+    {
+      "epoch": 0.00027756150720944985,
+      "grad_norm": 0.98046875,
+      "learning_rate": 0.00029999998605443067,
+      "loss": 0.5218,
+      "step": 1980
+    },
+    {
+      "epoch": 0.000278963333003437,
+      "grad_norm": 1.1796875,
+      "learning_rate": 0.0002999999857680815,
+      "loss": 0.5483,
+      "step": 1990
+    },
+    {
+      "epoch": 0.00028036515879742413,
+      "grad_norm": 0.494140625,
+      "learning_rate": 0.00029999998547882226,
+      "loss": 0.5184,
+      "step": 2000
+    },
+    {
+      "epoch": 0.00028036515879742413,
+      "eval_loss": 0.2110401690006256,
+      "eval_runtime": 113.7235,
+      "eval_samples_per_second": 200.864,
+      "eval_steps_per_second": 1.574,
+      "step": 2000
+    },
+    {
+      "epoch": 0.00028176698459141125,
+      "grad_norm": 0.76171875,
+      "learning_rate": 0.00029999998518665297,
+      "loss": 0.5142,
+      "step": 2010
+    },
+    {
+      "epoch": 0.00028316881038539836,
+      "grad_norm": 0.64453125,
+      "learning_rate": 0.0002999999848915736,
+      "loss": 0.6032,
+      "step": 2020
+    },
+    {
+      "epoch": 0.0002845706361793855,
+      "grad_norm": 0.65625,
+      "learning_rate": 0.0002999999845935842,
+      "loss": 0.5468,
+      "step": 2030
+    },
+    {
+      "epoch": 0.0002859724619733726,
+      "grad_norm": 0.380859375,
+      "learning_rate": 0.00029999998429268475,
+      "loss": 0.4213,
+      "step": 2040
+    },
+    {
+      "epoch": 0.0002873742877673597,
+      "grad_norm": 0.78125,
+      "learning_rate": 0.00029999998398887526,
+      "loss": 0.4999,
+      "step": 2050
+    },
+    {
+      "epoch": 0.0002887761135613468,
+      "grad_norm": 1.0703125,
+      "learning_rate": 0.0002999999836821557,
+      "loss": 0.4975,
+      "step": 2060
+    },
+    {
+      "epoch": 0.00029017793935533394,
+      "grad_norm": 0.396484375,
+      "learning_rate": 0.0002999999833725261,
+      "loss": 0.4618,
+      "step": 2070
+    },
+    {
+      "epoch": 0.0002915797651493211,
+      "grad_norm": 0.7890625,
+      "learning_rate": 0.00029999998305998645,
+      "loss": 0.6292,
+      "step": 2080
+    },
+    {
+      "epoch": 0.0002929815909433082,
+      "grad_norm": 1.1015625,
+      "learning_rate": 0.0002999999827445367,
+      "loss": 0.4893,
+      "step": 2090
+    },
+    {
+      "epoch": 0.00029438341673729534,
+      "grad_norm": 0.921875,
+      "learning_rate": 0.000299999982426177,
+      "loss": 0.4749,
+      "step": 2100
+    },
+    {
+      "epoch": 0.00029578524253128245,
+      "grad_norm": 1.21875,
+      "learning_rate": 0.0002999999821049071,
+      "loss": 0.5639,
+      "step": 2110
+    },
+    {
+      "epoch": 0.00029718706832526957,
+      "grad_norm": 0.275390625,
+      "learning_rate": 0.00029999998178072726,
+      "loss": 0.4799,
+      "step": 2120
+    },
+    {
+      "epoch": 0.0002985888941192567,
+      "grad_norm": 1.1875,
+      "learning_rate": 0.00029999998145363734,
+      "loss": 0.5267,
+      "step": 2130
+    },
+    {
+      "epoch": 0.0002999907199132438,
+      "grad_norm": 0.69140625,
+      "learning_rate": 0.00029999998112363737,
+      "loss": 0.5218,
+      "step": 2140
+    },
+    {
+      "epoch": 0.0003013925457072309,
+      "grad_norm": 0.98046875,
+      "learning_rate": 0.00029999998079072735,
+      "loss": 0.5824,
+      "step": 2150
+    },
+    {
+      "epoch": 0.00030279437150121803,
+      "grad_norm": 0.3359375,
+      "learning_rate": 0.00029999998045490727,
+      "loss": 0.5229,
+      "step": 2160
+    },
+    {
+      "epoch": 0.00030419619729520514,
+      "grad_norm": 0.490234375,
+      "learning_rate": 0.00029999998011617714,
+      "loss": 0.42,
+      "step": 2170
+    },
+    {
+      "epoch": 0.0003055980230891923,
+      "grad_norm": 0.8671875,
+      "learning_rate": 0.00029999997977453696,
+      "loss": 0.5678,
+      "step": 2180
+    },
+    {
+      "epoch": 0.0003069998488831794,
+      "grad_norm": 0.8359375,
+      "learning_rate": 0.0002999999794299867,
+      "loss": 0.5489,
+      "step": 2190
+    },
+    {
+      "epoch": 0.00030840167467716654,
+      "grad_norm": 0.40625,
+      "learning_rate": 0.00029999997908252644,
+      "loss": 0.4916,
+      "step": 2200
+    },
+    {
+      "epoch": 0.00030980350047115366,
+      "grad_norm": 0.8828125,
+      "learning_rate": 0.0002999999787321561,
+      "loss": 0.4691,
+      "step": 2210
+    },
+    {
+      "epoch": 0.00031120532626514077,
+      "grad_norm": 0.71484375,
+      "learning_rate": 0.0002999999783788757,
+      "loss": 0.4792,
+      "step": 2220
+    },
+    {
+      "epoch": 0.0003126071520591279,
+      "grad_norm": 0.66796875,
+      "learning_rate": 0.00029999997802268526,
+      "loss": 0.4996,
+      "step": 2230
+    },
+    {
+      "epoch": 0.000314008977853115,
+      "grad_norm": 1.0625,
+      "learning_rate": 0.0002999999776635848,
+      "loss": 0.4695,
+      "step": 2240
+    },
+    {
+      "epoch": 0.0003154108036471021,
+      "grad_norm": 0.31640625,
+      "learning_rate": 0.00029999997730157426,
+      "loss": 0.4476,
+      "step": 2250
+    },
+    {
+      "epoch": 0.00031681262944108923,
+      "grad_norm": 0.80859375,
+      "learning_rate": 0.0002999999769366536,
+      "loss": 0.5829,
+      "step": 2260
+    },
+    {
+      "epoch": 0.00031821445523507635,
+      "grad_norm": 0.98046875,
+      "learning_rate": 0.000299999976568823,
+      "loss": 0.521,
+      "step": 2270
+    },
+    {
+      "epoch": 0.0003196162810290635,
+      "grad_norm": 0.6640625,
+      "learning_rate": 0.0002999999761980823,
+      "loss": 0.5269,
+      "step": 2280
+    },
+    {
+      "epoch": 0.00032101810682305063,
+      "grad_norm": 0.80859375,
+      "learning_rate": 0.00029999997582443147,
+      "loss": 0.5492,
+      "step": 2290
+    },
+    {
+      "epoch": 0.00032241993261703775,
+      "grad_norm": 0.466796875,
+      "learning_rate": 0.00029999997544787066,
+      "loss": 0.5568,
+      "step": 2300
+    },
+    {
+      "epoch": 0.00032382175841102486,
+      "grad_norm": 0.2734375,
+      "learning_rate": 0.00029999997506839984,
+      "loss": 0.5229,
+      "step": 2310
+    },
+    {
+      "epoch": 0.000325223584205012,
+      "grad_norm": 1.0390625,
+      "learning_rate": 0.0002999999746860189,
+      "loss": 0.4528,
+      "step": 2320
+    },
+    {
+      "epoch": 0.0003266254099989991,
+      "grad_norm": 0.84765625,
+      "learning_rate": 0.00029999997430072794,
+      "loss": 0.5159,
+      "step": 2330
+    },
+    {
+      "epoch": 0.0003280272357929862,
+      "grad_norm": 0.640625,
+      "learning_rate": 0.0002999999739125269,
+      "loss": 0.4858,
+      "step": 2340
+    },
+    {
+      "epoch": 0.0003294290615869733,
+      "grad_norm": 1.109375,
+      "learning_rate": 0.00029999997352141584,
+      "loss": 0.5768,
+      "step": 2350
+    },
+    {
+      "epoch": 0.00033083088738096044,
+      "grad_norm": 0.859375,
+      "learning_rate": 0.0002999999731273947,
+      "loss": 0.5316,
+      "step": 2360
+    },
+    {
+      "epoch": 0.0003322327131749476,
+      "grad_norm": 0.5625,
+      "learning_rate": 0.0002999999727304635,
+      "loss": 0.4754,
+      "step": 2370
+    },
+    {
+      "epoch": 0.0003336345389689347,
+      "grad_norm": 0.86328125,
+      "learning_rate": 0.00029999997233062233,
+      "loss": 0.5883,
+      "step": 2380
+    },
+    {
+      "epoch": 0.00033503636476292183,
+      "grad_norm": 0.341796875,
+      "learning_rate": 0.00029999997192787104,
+      "loss": 0.5788,
+      "step": 2390
+    },
+    {
+      "epoch": 0.00033643819055690895,
+      "grad_norm": 0.376953125,
+      "learning_rate": 0.0002999999715222097,
+      "loss": 0.4608,
+      "step": 2400
+    },
+    {
+      "epoch": 0.00033784001635089606,
+      "grad_norm": 0.427734375,
+      "learning_rate": 0.00029999997111363836,
+      "loss": 0.4271,
+      "step": 2410
+    },
+    {
+      "epoch": 0.0003392418421448832,
+      "grad_norm": 0.984375,
+      "learning_rate": 0.0002999999707021569,
+      "loss": 0.5311,
+      "step": 2420
+    },
+    {
+      "epoch": 0.0003406436679388703,
+      "grad_norm": 0.71484375,
+      "learning_rate": 0.00029999997028776535,
+      "loss": 0.528,
+      "step": 2430
+    },
+    {
+      "epoch": 0.0003420454937328574,
+      "grad_norm": 0.7890625,
+      "learning_rate": 0.00029999996987046385,
+      "loss": 0.5641,
+      "step": 2440
+    },
+    {
+      "epoch": 0.0003434473195268445,
+      "grad_norm": 1.0078125,
+      "learning_rate": 0.00029999996945025224,
+      "loss": 0.597,
+      "step": 2450
+    },
+    {
+      "epoch": 0.00034484914532083164,
+      "grad_norm": 1.1015625,
+      "learning_rate": 0.00029999996902713063,
+      "loss": 0.4643,
+      "step": 2460
+    },
+    {
+      "epoch": 0.0003462509711148188,
+      "grad_norm": 0.60546875,
+      "learning_rate": 0.0002999999686010989,
+      "loss": 0.3736,
+      "step": 2470
+    },
+    {
+      "epoch": 0.0003476527969088059,
+      "grad_norm": 0.439453125,
+      "learning_rate": 0.00029999996817215715,
+      "loss": 0.5049,
+      "step": 2480
+    },
+    {
+      "epoch": 0.00034905462270279304,
+      "grad_norm": 0.74609375,
+      "learning_rate": 0.00029999996774030533,
+      "loss": 0.3526,
+      "step": 2490
+    },
+    {
+      "epoch": 0.00035045644849678015,
+      "grad_norm": 0.91796875,
+      "learning_rate": 0.0002999999673055435,
+      "loss": 0.5538,
+      "step": 2500
+    },
+    {
+      "epoch": 0.00035045644849678015,
+      "eval_loss": 0.20391806960105896,
+      "eval_runtime": 112.0697,
+      "eval_samples_per_second": 203.829,
+      "eval_steps_per_second": 1.597,
+      "step": 2500
+    },
+    {
+      "epoch": 0.00035185827429076727,
+      "grad_norm": 0.353515625,
+      "learning_rate": 0.0002999999668678716,
+      "loss": 0.4682,
+      "step": 2510
+    },
+    {
+      "epoch": 0.0003532601000847544,
+      "grad_norm": 0.421875,
+      "learning_rate": 0.0002999999664272896,
+      "loss": 0.5603,
+      "step": 2520
+    },
+    {
+      "epoch": 0.0003546619258787415,
+      "grad_norm": 0.52734375,
+      "learning_rate": 0.0002999999659837976,
+      "loss": 0.4438,
+      "step": 2530
+    },
+    {
+      "epoch": 0.0003560637516727286,
+      "grad_norm": 0.328125,
+      "learning_rate": 0.00029999996553739555,
+      "loss": 0.4908,
+      "step": 2540
+    },
+    {
+      "epoch": 0.00035746557746671573,
+      "grad_norm": 1.046875,
+      "learning_rate": 0.0002999999650880834,
+      "loss": 0.4753,
+      "step": 2550
+    },
+    {
+      "epoch": 0.00035886740326070284,
+      "grad_norm": 0.99609375,
+      "learning_rate": 0.0002999999646358613,
+      "loss": 0.4611,
+      "step": 2560
+    },
+    {
+      "epoch": 0.00036026922905469,
+      "grad_norm": 0.609375,
+      "learning_rate": 0.00029999996418072903,
+      "loss": 0.3998,
+      "step": 2570
+    },
+    {
+      "epoch": 0.00036167105484867713,
+      "grad_norm": 0.359375,
+      "learning_rate": 0.00029999996372268673,
+      "loss": 0.4206,
+      "step": 2580
+    },
+    {
+      "epoch": 0.00036307288064266424,
+      "grad_norm": 1.1484375,
+      "learning_rate": 0.00029999996326173444,
+      "loss": 0.5298,
+      "step": 2590
+    },
+    {
+      "epoch": 0.00036447470643665136,
+      "grad_norm": 0.255859375,
+      "learning_rate": 0.00029999996279787204,
+      "loss": 0.4437,
+      "step": 2600
+    },
+    {
+      "epoch": 0.00036587653223063847,
+      "grad_norm": 0.40625,
+      "learning_rate": 0.00029999996233109964,
+      "loss": 0.4519,
+      "step": 2610
+    },
+    {
+      "epoch": 0.0003672783580246256,
+      "grad_norm": 0.427734375,
+      "learning_rate": 0.00029999996186141713,
+      "loss": 0.3585,
+      "step": 2620
+    },
+    {
+      "epoch": 0.0003686801838186127,
+      "grad_norm": 0.486328125,
+      "learning_rate": 0.00029999996138882463,
+      "loss": 0.4805,
+      "step": 2630
+    },
+    {
+      "epoch": 0.0003700820096125998,
+      "grad_norm": 0.94140625,
+      "learning_rate": 0.000299999960913322,
+      "loss": 0.549,
+      "step": 2640
+    },
+    {
+      "epoch": 0.00037148383540658693,
+      "grad_norm": 1.1796875,
+      "learning_rate": 0.0002999999604349094,
+      "loss": 0.4753,
+      "step": 2650
+    },
+    {
+      "epoch": 0.0003728856612005741,
+      "grad_norm": 0.4609375,
+      "learning_rate": 0.00029999995995358674,
+      "loss": 0.5911,
+      "step": 2660
+    },
+    {
+      "epoch": 0.0003742874869945612,
+      "grad_norm": 0.83203125,
+      "learning_rate": 0.00029999995946935397,
+      "loss": 0.5762,
+      "step": 2670
+    },
+    {
+      "epoch": 0.00037568931278854833,
+      "grad_norm": 0.404296875,
+      "learning_rate": 0.0002999999589822112,
+      "loss": 0.5504,
+      "step": 2680
+    },
+    {
+      "epoch": 0.00037709113858253545,
+      "grad_norm": 0.67578125,
+      "learning_rate": 0.0002999999584921583,
+      "loss": 0.435,
+      "step": 2690
+    },
+    {
+      "epoch": 0.00037849296437652256,
+      "grad_norm": 1.3984375,
+      "learning_rate": 0.00029999995799919545,
+      "loss": 0.4766,
+      "step": 2700
+    },
+    {
+      "epoch": 0.0003798947901705097,
+      "grad_norm": 0.69921875,
+      "learning_rate": 0.00029999995750332246,
+      "loss": 0.5923,
+      "step": 2710
+    },
+    {
+      "epoch": 0.0003812966159644968,
+      "grad_norm": 0.48828125,
+      "learning_rate": 0.0002999999570045395,
+      "loss": 0.5589,
+      "step": 2720
+    },
+    {
+      "epoch": 0.0003826984417584839,
+      "grad_norm": 0.365234375,
+      "learning_rate": 0.00029999995650284645,
+      "loss": 0.5023,
+      "step": 2730
+    },
+    {
+      "epoch": 0.000384100267552471,
+      "grad_norm": 0.69140625,
+      "learning_rate": 0.0002999999559982433,
+      "loss": 0.5224,
+      "step": 2740
+    },
+    {
+      "epoch": 0.00038550209334645814,
+      "grad_norm": 0.54296875,
+      "learning_rate": 0.00029999995549073017,
+      "loss": 0.4,
+      "step": 2750
+    },
+    {
+      "epoch": 0.0003869039191404453,
+      "grad_norm": 0.396484375,
+      "learning_rate": 0.0002999999549803069,
+      "loss": 0.4416,
+      "step": 2760
+    },
+    {
+      "epoch": 0.0003883057449344324,
+      "grad_norm": 0.68359375,
+      "learning_rate": 0.0002999999544669737,
+      "loss": 0.4588,
+      "step": 2770
+    },
+    {
+      "epoch": 0.00038970757072841954,
+      "grad_norm": 0.6015625,
+      "learning_rate": 0.0002999999539507304,
+      "loss": 0.5128,
+      "step": 2780
+    },
+    {
+      "epoch": 0.00039110939652240665,
+      "grad_norm": 0.640625,
+      "learning_rate": 0.00029999995343157703,
+      "loss": 0.4519,
+      "step": 2790
+    },
+    {
+      "epoch": 0.00039251122231639377,
+      "grad_norm": 0.75,
+      "learning_rate": 0.00029999995290951357,
+      "loss": 0.553,
+      "step": 2800
+    },
+    {
+      "epoch": 0.0003939130481103809,
+      "grad_norm": 0.58203125,
+      "learning_rate": 0.0002999999523845401,
+      "loss": 0.4382,
+      "step": 2810
+    },
+    {
+      "epoch": 0.000395314873904368,
+      "grad_norm": 0.8984375,
+      "learning_rate": 0.0002999999518566566,
+      "loss": 0.6044,
+      "step": 2820
+    },
+    {
+      "epoch": 0.0003967166996983551,
+      "grad_norm": 0.65625,
+      "learning_rate": 0.00029999995132586304,
+      "loss": 0.5411,
+      "step": 2830
+    },
+    {
+      "epoch": 0.0003981185254923422,
+      "grad_norm": 0.5859375,
+      "learning_rate": 0.0002999999507921594,
+      "loss": 0.48,
+      "step": 2840
+    },
+    {
+      "epoch": 0.00039952035128632934,
+      "grad_norm": 0.96484375,
+      "learning_rate": 0.00029999995025554576,
+      "loss": 0.5314,
+      "step": 2850
+    },
+    {
+      "epoch": 0.0004009221770803165,
+      "grad_norm": 1.0859375,
+      "learning_rate": 0.000299999949716022,
+      "loss": 0.5695,
+      "step": 2860
+    },
+    {
+      "epoch": 0.0004023240028743036,
+      "grad_norm": 1.0078125,
+      "learning_rate": 0.0002999999491735882,
+      "loss": 0.4795,
+      "step": 2870
+    },
+    {
+      "epoch": 0.00040372582866829074,
+      "grad_norm": 0.46484375,
+      "learning_rate": 0.0002999999486282444,
+      "loss": 0.3841,
+      "step": 2880
+    },
+    {
+      "epoch": 0.00040512765446227785,
+      "grad_norm": 0.67578125,
+      "learning_rate": 0.0002999999480799905,
+      "loss": 0.4925,
+      "step": 2890
+    },
+    {
+      "epoch": 0.00040652948025626497,
+      "grad_norm": 0.68359375,
+      "learning_rate": 0.00029999994752882657,
+      "loss": 0.4798,
+      "step": 2900
+    },
+    {
+      "epoch": 0.0004079313060502521,
+      "grad_norm": 0.9453125,
+      "learning_rate": 0.0002999999469747526,
+      "loss": 0.5165,
+      "step": 2910
+    },
+    {
+      "epoch": 0.0004093331318442392,
+      "grad_norm": 0.70703125,
+      "learning_rate": 0.00029999994641776855,
+      "loss": 0.5801,
+      "step": 2920
+    },
+    {
+      "epoch": 0.0004107349576382263,
+      "grad_norm": 0.3359375,
+      "learning_rate": 0.00029999994585787446,
+      "loss": 0.5142,
+      "step": 2930
+    },
+    {
+      "epoch": 0.00041213678343221343,
+      "grad_norm": 1.0234375,
+      "learning_rate": 0.0002999999452950703,
+      "loss": 0.4823,
+      "step": 2940
+    },
+    {
+      "epoch": 0.0004135386092262006,
+      "grad_norm": 0.294921875,
+      "learning_rate": 0.0002999999447293561,
+      "loss": 0.4351,
+      "step": 2950
+    },
+    {
+      "epoch": 0.0004149404350201877,
+      "grad_norm": 0.396484375,
+      "learning_rate": 0.0002999999441607319,
+      "loss": 0.5292,
+      "step": 2960
+    },
+    {
+      "epoch": 0.00041634226081417483,
+      "grad_norm": 0.69921875,
+      "learning_rate": 0.0002999999435891976,
+      "loss": 0.458,
+      "step": 2970
+    },
+    {
+      "epoch": 0.00041774408660816194,
+      "grad_norm": 0.294921875,
+      "learning_rate": 0.00029999994301475327,
+      "loss": 0.4933,
+      "step": 2980
+    },
+    {
+      "epoch": 0.00041914591240214906,
+      "grad_norm": 1.140625,
+      "learning_rate": 0.0002999999424373988,
+      "loss": 0.4954,
+      "step": 2990
+    },
+    {
+      "epoch": 0.0004205477381961362,
+      "grad_norm": 1.5234375,
+      "learning_rate": 0.0002999999418571344,
+      "loss": 0.4878,
+      "step": 3000
+    },
+    {
+      "epoch": 0.0004205477381961362,
+      "eval_loss": 0.20102550089359283,
+      "eval_runtime": 112.0673,
+      "eval_samples_per_second": 203.833,
+      "eval_steps_per_second": 1.597,
+      "step": 3000
+    },
+    {
+      "epoch": 0.0004219495639901233,
+      "grad_norm": 0.62890625,
+      "learning_rate": 0.0002999999412739599,
+      "loss": 0.5205,
+      "step": 3010
+    },
+    {
+      "epoch": 0.0004233513897841104,
+      "grad_norm": 0.40625,
+      "learning_rate": 0.0002999999406878753,
+      "loss": 0.4147,
+      "step": 3020
+    },
+    {
+      "epoch": 0.0004247532155780975,
+      "grad_norm": 0.4921875,
+      "learning_rate": 0.0002999999400988807,
+      "loss": 0.424,
+      "step": 3030
+    },
+    {
+      "epoch": 0.00042615504137208463,
+      "grad_norm": 0.93359375,
+      "learning_rate": 0.000299999939506976,
+      "loss": 0.4885,
+      "step": 3040
+    },
+    {
+      "epoch": 0.0004275568671660718,
+      "grad_norm": 1.2265625,
+      "learning_rate": 0.00029999993891216136,
+      "loss": 0.517,
+      "step": 3050
+    },
+    {
+      "epoch": 0.0004289586929600589,
+      "grad_norm": 0.6796875,
+      "learning_rate": 0.0002999999383144366,
+      "loss": 0.5413,
+      "step": 3060
+    },
+    {
+      "epoch": 0.00043036051875404603,
+      "grad_norm": 0.5390625,
+      "learning_rate": 0.00029999993771380175,
+      "loss": 0.4722,
+      "step": 3070
+    },
+    {
+      "epoch": 0.00043176234454803315,
+      "grad_norm": 0.58203125,
+      "learning_rate": 0.0002999999371102569,
+      "loss": 0.5205,
+      "step": 3080
+    },
+    {
+      "epoch": 0.00043316417034202026,
+      "grad_norm": 0.314453125,
+      "learning_rate": 0.000299999936503802,
+      "loss": 0.407,
+      "step": 3090
+    },
+    {
+      "epoch": 0.0004345659961360074,
+      "grad_norm": 1.2109375,
+      "learning_rate": 0.00029999993589443705,
+      "loss": 0.495,
+      "step": 3100
+    },
+    {
+      "epoch": 0.0004359678219299945,
+      "grad_norm": 0.81640625,
+      "learning_rate": 0.000299999935282162,
+      "loss": 0.4925,
+      "step": 3110
+    },
+    {
+      "epoch": 0.0004373696477239816,
+      "grad_norm": 0.9921875,
+      "learning_rate": 0.0002999999346669769,
+      "loss": 0.42,
+      "step": 3120
+    },
+    {
+      "epoch": 0.0004387714735179687,
+      "grad_norm": 0.3984375,
+      "learning_rate": 0.0002999999340488818,
+      "loss": 0.3723,
+      "step": 3130
+    },
+    {
+      "epoch": 0.00044017329931195584,
+      "grad_norm": 0.88671875,
+      "learning_rate": 0.0002999999334278766,
+      "loss": 0.5267,
+      "step": 3140
+    },
+    {
+      "epoch": 0.000441575125105943,
+      "grad_norm": 0.59765625,
+      "learning_rate": 0.0002999999328039614,
+      "loss": 0.4224,
+      "step": 3150
+    },
+    {
+      "epoch": 0.0004429769508999301,
+      "grad_norm": 1.3203125,
+      "learning_rate": 0.0002999999321771361,
+      "loss": 0.5182,
+      "step": 3160
+    },
+    {
+      "epoch": 0.00044437877669391724,
+      "grad_norm": 0.8359375,
+      "learning_rate": 0.0002999999315474008,
+      "loss": 0.5608,
+      "step": 3170
+    },
+    {
+      "epoch": 0.00044578060248790435,
+      "grad_norm": 0.482421875,
+      "learning_rate": 0.00029999993091475545,
+      "loss": 0.3993,
+      "step": 3180
+    },
+    {
+      "epoch": 0.00044718242828189147,
+      "grad_norm": 0.97265625,
+      "learning_rate": 0.0002999999302792,
+      "loss": 0.4892,
+      "step": 3190
+    },
+    {
+      "epoch": 0.0004485842540758786,
+      "grad_norm": 0.92578125,
+      "learning_rate": 0.00029999992964073447,
+      "loss": 0.5218,
+      "step": 3200
+    },
+    {
+      "epoch": 0.0004499860798698657,
+      "grad_norm": 0.69921875,
+      "learning_rate": 0.00029999992899935895,
+      "loss": 0.3582,
+      "step": 3210
+    },
+    {
+      "epoch": 0.0004513879056638528,
+      "grad_norm": 0.53515625,
+      "learning_rate": 0.0002999999283550734,
+      "loss": 0.5051,
+      "step": 3220
+    },
+    {
+      "epoch": 0.0004527897314578399,
+      "grad_norm": 0.90625,
+      "learning_rate": 0.00029999992770787776,
+      "loss": 0.4788,
+      "step": 3230
+    },
+    {
+      "epoch": 0.0004541915572518271,
+      "grad_norm": 0.58203125,
+      "learning_rate": 0.00029999992705777203,
+      "loss": 0.43,
+      "step": 3240
+    },
+    {
+      "epoch": 0.0004555933830458142,
+      "grad_norm": 0.9921875,
+      "learning_rate": 0.00029999992640475636,
+      "loss": 0.5673,
+      "step": 3250
+    },
+    {
+      "epoch": 0.0004569952088398013,
+      "grad_norm": 0.6171875,
+      "learning_rate": 0.0002999999257488305,
+      "loss": 0.4597,
+      "step": 3260
+    },
+    {
+      "epoch": 0.00045839703463378844,
+      "grad_norm": 0.94921875,
+      "learning_rate": 0.0002999999250899947,
+      "loss": 0.5661,
+      "step": 3270
+    },
+    {
+      "epoch": 0.00045979886042777556,
+      "grad_norm": 0.427734375,
+      "learning_rate": 0.0002999999244282488,
+      "loss": 0.5277,
+      "step": 3280
+    },
+    {
+      "epoch": 0.00046120068622176267,
+      "grad_norm": 0.71875,
+      "learning_rate": 0.00029999992376359287,
+      "loss": 0.4895,
+      "step": 3290
+    },
+    {
+      "epoch": 0.0004626025120157498,
+      "grad_norm": 0.96875,
+      "learning_rate": 0.0002999999230960269,
+      "loss": 0.4452,
+      "step": 3300
+    },
+    {
+      "epoch": 0.0004640043378097369,
+      "grad_norm": 0.55078125,
+      "learning_rate": 0.00029999992242555083,
+      "loss": 0.4958,
+      "step": 3310
+    },
+    {
+      "epoch": 0.000465406163603724,
+      "grad_norm": 0.8671875,
+      "learning_rate": 0.0002999999217521647,
+      "loss": 0.4742,
+      "step": 3320
+    },
+    {
+      "epoch": 0.00046680798939771113,
+      "grad_norm": 0.62890625,
+      "learning_rate": 0.00029999992107586853,
+      "loss": 0.4959,
+      "step": 3330
+    },
+    {
+      "epoch": 0.0004682098151916983,
+      "grad_norm": 0.640625,
+      "learning_rate": 0.0002999999203966623,
+      "loss": 0.5154,
+      "step": 3340
+    },
+    {
+      "epoch": 0.0004696116409856854,
+      "grad_norm": 0.486328125,
+      "learning_rate": 0.00029999991971454607,
+      "loss": 0.5168,
+      "step": 3350
+    },
+    {
+      "epoch": 0.00047101346677967253,
+      "grad_norm": 0.455078125,
+      "learning_rate": 0.00029999991902951976,
+      "loss": 0.5385,
+      "step": 3360
+    },
+    {
+      "epoch": 0.00047241529257365964,
+      "grad_norm": 0.47265625,
+      "learning_rate": 0.00029999991834158345,
+      "loss": 0.5004,
+      "step": 3370
+    },
+    {
+      "epoch": 0.00047381711836764676,
+      "grad_norm": 0.2421875,
+      "learning_rate": 0.00029999991765073704,
+      "loss": 0.4502,
+      "step": 3380
+    },
+    {
+      "epoch": 0.0004752189441616339,
+      "grad_norm": 0.4765625,
+      "learning_rate": 0.00029999991695698057,
+      "loss": 0.4549,
+      "step": 3390
+    },
+    {
+      "epoch": 0.000476620769955621,
+      "grad_norm": 0.4453125,
+      "learning_rate": 0.00029999991626031405,
+      "loss": 0.4557,
+      "step": 3400
+    },
+    {
+      "epoch": 0.0004780225957496081,
+      "grad_norm": 0.302734375,
+      "learning_rate": 0.00029999991556073753,
+      "loss": 0.4765,
+      "step": 3410
+    },
+    {
+      "epoch": 0.0004794244215435952,
+      "grad_norm": 0.66796875,
+      "learning_rate": 0.0002999999148582509,
+      "loss": 0.4911,
+      "step": 3420
+    },
+    {
+      "epoch": 0.00048082624733758233,
+      "grad_norm": 0.796875,
+      "learning_rate": 0.00029999991415285423,
+      "loss": 0.5774,
+      "step": 3430
+    },
+    {
+      "epoch": 0.0004822280731315695,
+      "grad_norm": 1.0234375,
+      "learning_rate": 0.0002999999134445475,
+      "loss": 0.4328,
+      "step": 3440
+    },
+    {
+      "epoch": 0.0004836298989255566,
+      "grad_norm": 0.37109375,
+      "learning_rate": 0.00029999991273333077,
+      "loss": 0.4692,
+      "step": 3450
+    },
+    {
+      "epoch": 0.00048503172471954373,
+      "grad_norm": 0.8515625,
+      "learning_rate": 0.00029999991201920393,
+      "loss": 0.4571,
+      "step": 3460
+    },
+    {
+      "epoch": 0.00048643355051353085,
+      "grad_norm": 0.62890625,
+      "learning_rate": 0.00029999991130216704,
+      "loss": 0.5374,
+      "step": 3470
+    },
+    {
+      "epoch": 0.00048783537630751796,
+      "grad_norm": 0.98046875,
+      "learning_rate": 0.00029999991058222015,
+      "loss": 0.5196,
+      "step": 3480
+    },
+    {
+      "epoch": 0.0004892372021015051,
+      "grad_norm": 0.44921875,
+      "learning_rate": 0.0002999999098593632,
+      "loss": 0.4814,
+      "step": 3490
+    },
+    {
+      "epoch": 0.0004906390278954922,
+      "grad_norm": 1.21875,
+      "learning_rate": 0.00029999990913359616,
+      "loss": 0.5114,
+      "step": 3500
+    },
+    {
+      "epoch": 0.0004906390278954922,
+      "eval_loss": 0.19615845382213593,
+      "eval_runtime": 113.7184,
+      "eval_samples_per_second": 200.873,
+      "eval_steps_per_second": 1.574,
+      "step": 3500
+    },
+    {
+      "epoch": 0.0004920408536894794,
+      "grad_norm": 0.6875,
+      "learning_rate": 0.00029999990840491906,
+      "loss": 0.5248,
+      "step": 3510
+    },
+    {
+      "epoch": 0.0004934426794834664,
+      "grad_norm": 0.97265625,
+      "learning_rate": 0.00029999990767333196,
+      "loss": 0.4625,
+      "step": 3520
+    },
+    {
+      "epoch": 0.0004948445052774536,
+      "grad_norm": 0.69921875,
+      "learning_rate": 0.0002999999069388348,
+      "loss": 0.4906,
+      "step": 3530
+    },
+    {
+      "epoch": 0.0004962463310714407,
+      "grad_norm": 0.703125,
+      "learning_rate": 0.00029999990620142755,
+      "loss": 0.3924,
+      "step": 3540
+    },
+    {
+      "epoch": 0.0004976481568654278,
+      "grad_norm": 0.97265625,
+      "learning_rate": 0.0002999999054611103,
+      "loss": 0.5016,
+      "step": 3550
+    },
+    {
+      "epoch": 0.0004990499826594149,
+      "grad_norm": 0.87890625,
+      "learning_rate": 0.000299999904717883,
+      "loss": 0.4974,
+      "step": 3560
+    },
+    {
+      "epoch": 0.000500451808453402,
+      "grad_norm": 0.84375,
+      "learning_rate": 0.0002999999039717456,
+      "loss": 0.5082,
+      "step": 3570
+    },
+    {
+      "epoch": 0.0005018536342473892,
+      "grad_norm": 0.279296875,
+      "learning_rate": 0.00029999990322269814,
+      "loss": 0.5716,
+      "step": 3580
+    },
+    {
+      "epoch": 0.0005032554600413763,
+      "grad_norm": 0.734375,
+      "learning_rate": 0.00029999990247074073,
+      "loss": 0.548,
+      "step": 3590
+    },
+    {
+      "epoch": 0.0005046572858353635,
+      "grad_norm": 0.58203125,
+      "learning_rate": 0.00029999990171587315,
+      "loss": 0.4553,
+      "step": 3600
+    },
+    {
+      "epoch": 0.0005060591116293505,
+      "grad_norm": 0.7734375,
+      "learning_rate": 0.0002999999009580956,
+      "loss": 0.5029,
+      "step": 3610
+    },
+    {
+      "epoch": 0.0005074609374233377,
+      "grad_norm": 0.46484375,
+      "learning_rate": 0.0002999999001974079,
+      "loss": 0.4318,
+      "step": 3620
+    },
+    {
+      "epoch": 0.0005088627632173247,
+      "grad_norm": 0.61328125,
+      "learning_rate": 0.00029999989943381027,
+      "loss": 0.5442,
+      "step": 3630
+    },
+    {
+      "epoch": 0.0005102645890113119,
+      "grad_norm": 0.72265625,
+      "learning_rate": 0.00029999989866730253,
+      "loss": 0.4907,
+      "step": 3640
+    },
+    {
+      "epoch": 0.000511666414805299,
+      "grad_norm": 0.30078125,
+      "learning_rate": 0.00029999989789788475,
+      "loss": 0.4758,
+      "step": 3650
+    },
+    {
+      "epoch": 0.0005130682405992861,
+      "grad_norm": 0.96484375,
+      "learning_rate": 0.0002999998971255569,
+      "loss": 0.4656,
+      "step": 3660
+    },
+    {
+      "epoch": 0.0005144700663932732,
+      "grad_norm": 0.431640625,
+      "learning_rate": 0.000299999896350319,
+      "loss": 0.5217,
+      "step": 3670
+    },
+    {
+      "epoch": 0.0005158718921872604,
+      "grad_norm": 0.44921875,
+      "learning_rate": 0.00029999989557217107,
+      "loss": 0.5127,
+      "step": 3680
+    },
+    {
+      "epoch": 0.0005172737179812475,
+      "grad_norm": 0.796875,
+      "learning_rate": 0.0002999998947911131,
+      "loss": 0.5713,
+      "step": 3690
+    },
+    {
+      "epoch": 0.0005186755437752346,
+      "grad_norm": 0.66796875,
+      "learning_rate": 0.00029999989400714507,
+      "loss": 0.4652,
+      "step": 3700
+    },
+    {
+      "epoch": 0.0005200773695692218,
+      "grad_norm": 0.66796875,
+      "learning_rate": 0.00029999989322026697,
+      "loss": 0.5029,
+      "step": 3710
+    },
+    {
+      "epoch": 0.0005214791953632088,
+      "grad_norm": 0.21484375,
+      "learning_rate": 0.0002999998924304788,
+      "loss": 0.4575,
+      "step": 3720
+    },
+    {
+      "epoch": 0.000522881021157196,
+      "grad_norm": 0.84375,
+      "learning_rate": 0.00029999989163778066,
+      "loss": 0.4947,
+      "step": 3730
+    },
+    {
+      "epoch": 0.0005242828469511831,
+      "grad_norm": 0.6640625,
+      "learning_rate": 0.0002999998908421724,
+      "loss": 0.4046,
+      "step": 3740
+    },
+    {
+      "epoch": 0.0005256846727451702,
+      "grad_norm": 0.9609375,
+      "learning_rate": 0.0002999998900436541,
+      "loss": 0.5298,
+      "step": 3750
+    },
+    {
+      "epoch": 0.0005270864985391573,
+      "grad_norm": 0.77734375,
+      "learning_rate": 0.00029999988924222577,
+      "loss": 0.4179,
+      "step": 3760
+    },
+    {
+      "epoch": 0.0005284883243331445,
+      "grad_norm": 0.291015625,
+      "learning_rate": 0.00029999988843788734,
+      "loss": 0.5142,
+      "step": 3770
+    },
+    {
+      "epoch": 0.0005298901501271316,
+      "grad_norm": 0.90625,
+      "learning_rate": 0.0002999998876306389,
+      "loss": 0.4722,
+      "step": 3780
+    },
+    {
+      "epoch": 0.0005312919759211187,
+      "grad_norm": 0.92578125,
+      "learning_rate": 0.0002999998868204804,
+      "loss": 0.5202,
+      "step": 3790
+    },
+    {
+      "epoch": 0.0005326938017151059,
+      "grad_norm": 1.1328125,
+      "learning_rate": 0.00029999988600741187,
+      "loss": 0.5537,
+      "step": 3800
+    },
+    {
+      "epoch": 0.0005340956275090929,
+      "grad_norm": 0.33984375,
+      "learning_rate": 0.0002999998851914333,
+      "loss": 0.5072,
+      "step": 3810
+    },
+    {
+      "epoch": 0.0005354974533030801,
+      "grad_norm": 1.140625,
+      "learning_rate": 0.0002999998843725446,
+      "loss": 0.4999,
+      "step": 3820
+    },
+    {
+      "epoch": 0.0005368992790970672,
+      "grad_norm": 0.7265625,
+      "learning_rate": 0.0002999998835507459,
+      "loss": 0.4916,
+      "step": 3830
+    },
+    {
+      "epoch": 0.0005383011048910543,
+      "grad_norm": 0.82421875,
+      "learning_rate": 0.0002999998827260372,
+      "loss": 0.6081,
+      "step": 3840
+    },
+    {
+      "epoch": 0.0005397029306850414,
+      "grad_norm": 0.671875,
+      "learning_rate": 0.0002999998818984184,
+      "loss": 0.5145,
+      "step": 3850
+    },
+    {
+      "epoch": 0.0005411047564790285,
+      "grad_norm": 1.4765625,
+      "learning_rate": 0.00029999988106788956,
+      "loss": 0.4729,
+      "step": 3860
+    },
+    {
+      "epoch": 0.0005425065822730157,
+      "grad_norm": 0.81640625,
+      "learning_rate": 0.0002999998802344506,
+      "loss": 0.5762,
+      "step": 3870
+    },
+    {
+      "epoch": 0.0005439084080670028,
+      "grad_norm": 0.90234375,
+      "learning_rate": 0.0002999998793981017,
+      "loss": 0.5645,
+      "step": 3880
+    },
+    {
+      "epoch": 0.00054531023386099,
+      "grad_norm": 0.3203125,
+      "learning_rate": 0.0002999998785588427,
+      "loss": 0.5388,
+      "step": 3890
+    },
+    {
+      "epoch": 0.000546712059654977,
+      "grad_norm": 0.80859375,
+      "learning_rate": 0.00029999987771667366,
+      "loss": 0.4449,
+      "step": 3900
+    },
+    {
+      "epoch": 0.0005481138854489642,
+      "grad_norm": 0.65625,
+      "learning_rate": 0.00029999987687159455,
+      "loss": 0.5345,
+      "step": 3910
+    },
+    {
+      "epoch": 0.0005495157112429512,
+      "grad_norm": 0.53125,
+      "learning_rate": 0.00029999987602360545,
+      "loss": 0.319,
+      "step": 3920
+    },
+    {
+      "epoch": 0.0005509175370369384,
+      "grad_norm": 0.427734375,
+      "learning_rate": 0.0002999998751727062,
+      "loss": 0.3845,
+      "step": 3930
+    },
+    {
+      "epoch": 0.0005523193628309255,
+      "grad_norm": 0.87890625,
+      "learning_rate": 0.00029999987431889697,
+      "loss": 0.4358,
+      "step": 3940
+    },
+    {
+      "epoch": 0.0005537211886249126,
+      "grad_norm": 0.76171875,
+      "learning_rate": 0.00029999987346217765,
+      "loss": 0.424,
+      "step": 3950
+    },
+    {
+      "epoch": 0.0005551230144188997,
+      "grad_norm": 0.59765625,
+      "learning_rate": 0.00029999987260254834,
+      "loss": 0.5856,
+      "step": 3960
+    },
+    {
+      "epoch": 0.0005565248402128869,
+      "grad_norm": 0.69140625,
+      "learning_rate": 0.00029999987174000886,
+      "loss": 0.4395,
+      "step": 3970
+    },
+    {
+      "epoch": 0.000557926666006874,
+      "grad_norm": 0.74609375,
+      "learning_rate": 0.00029999987087455944,
+      "loss": 0.4667,
+      "step": 3980
+    },
+    {
+      "epoch": 0.0005593284918008611,
+      "grad_norm": 0.62109375,
+      "learning_rate": 0.0002999998700061999,
+      "loss": 0.5008,
+      "step": 3990
+    },
+    {
+      "epoch": 0.0005607303175948483,
+      "grad_norm": 0.69140625,
+      "learning_rate": 0.0002999998691349304,
+      "loss": 0.5067,
+      "step": 4000
+    },
+    {
+      "epoch": 0.0005607303175948483,
+      "eval_loss": 0.1947367787361145,
+      "eval_runtime": 112.0665,
+      "eval_samples_per_second": 203.834,
+      "eval_steps_per_second": 1.597,
+      "step": 4000
+    },
+    {
+      "epoch": 0.0005621321433888353,
+      "grad_norm": 0.8984375,
+      "learning_rate": 0.0002999998682607508,
+      "loss": 0.4297,
+      "step": 4010
+    },
+    {
+      "epoch": 0.0005635339691828225,
+      "grad_norm": 0.22265625,
+      "learning_rate": 0.0002999998673836611,
+      "loss": 0.3309,
+      "step": 4020
+    },
+    {
+      "epoch": 0.0005649357949768096,
+      "grad_norm": 0.61328125,
+      "learning_rate": 0.0002999998665036614,
+      "loss": 0.3924,
+      "step": 4030
+    },
+    {
+      "epoch": 0.0005663376207707967,
+      "grad_norm": 0.75390625,
+      "learning_rate": 0.00029999986562075163,
+      "loss": 0.4433,
+      "step": 4040
+    },
+    {
+      "epoch": 0.0005677394465647838,
+      "grad_norm": 0.458984375,
+      "learning_rate": 0.00029999986473493184,
+      "loss": 0.4218,
+      "step": 4050
+    },
+    {
+      "epoch": 0.000569141272358771,
+      "grad_norm": 0.39453125,
+      "learning_rate": 0.000299999863846202,
+      "loss": 0.5819,
+      "step": 4060
+    },
+    {
+      "epoch": 0.0005705430981527581,
+      "grad_norm": 0.71484375,
+      "learning_rate": 0.0002999998629545621,
+      "loss": 0.5292,
+      "step": 4070
+    },
+    {
+      "epoch": 0.0005719449239467452,
+      "grad_norm": 0.55078125,
+      "learning_rate": 0.00029999986206001215,
+      "loss": 0.4745,
+      "step": 4080
+    },
+    {
+      "epoch": 0.0005733467497407324,
+      "grad_norm": 0.5234375,
+      "learning_rate": 0.00029999986116255214,
+      "loss": 0.5194,
+      "step": 4090
+    },
+    {
+      "epoch": 0.0005747485755347194,
+      "grad_norm": 0.5546875,
+      "learning_rate": 0.0002999998602621821,
+      "loss": 0.5094,
+      "step": 4100
+    },
+    {
+      "epoch": 0.0005761504013287066,
+      "grad_norm": 0.416015625,
+      "learning_rate": 0.00029999985935890203,
+      "loss": 0.4252,
+      "step": 4110
+    },
+    {
+      "epoch": 0.0005775522271226936,
+      "grad_norm": 0.53125,
+      "learning_rate": 0.0002999998584527118,
+      "loss": 0.439,
+      "step": 4120
+    },
+    {
+      "epoch": 0.0005789540529166808,
+      "grad_norm": 0.6953125,
+      "learning_rate": 0.0002999998575436116,
+      "loss": 0.5719,
+      "step": 4130
+    },
+    {
+      "epoch": 0.0005803558787106679,
+      "grad_norm": 0.6640625,
+      "learning_rate": 0.00029999985663160133,
+      "loss": 0.4448,
+      "step": 4140
+    },
+    {
+      "epoch": 0.000581757704504655,
+      "grad_norm": 0.484375,
+      "learning_rate": 0.00029999985571668107,
+      "loss": 0.4758,
+      "step": 4150
+    },
+    {
+      "epoch": 0.0005831595302986422,
+      "grad_norm": 0.91796875,
+      "learning_rate": 0.0002999998547988507,
+      "loss": 0.4819,
+      "step": 4160
+    },
+    {
+      "epoch": 0.0005845613560926293,
+      "grad_norm": 0.45703125,
+      "learning_rate": 0.0002999998538781103,
+      "loss": 0.4092,
+      "step": 4170
+    },
+    {
+      "epoch": 0.0005859631818866164,
+      "grad_norm": 0.29296875,
+      "learning_rate": 0.00029999985295445984,
+      "loss": 0.4887,
+      "step": 4180
+    },
+    {
+      "epoch": 0.0005873650076806035,
+      "grad_norm": 0.890625,
+      "learning_rate": 0.0002999998520278993,
+      "loss": 0.5011,
+      "step": 4190
+    },
+    {
+      "epoch": 0.0005887668334745907,
+      "grad_norm": 1.140625,
+      "learning_rate": 0.0002999998510984288,
+      "loss": 0.4195,
+      "step": 4200
+    },
+    {
+      "epoch": 0.0005901686592685777,
+      "grad_norm": 0.8984375,
+      "learning_rate": 0.0002999998501660482,
+      "loss": 0.4402,
+      "step": 4210
+    },
+    {
+      "epoch": 0.0005915704850625649,
+      "grad_norm": 0.75390625,
+      "learning_rate": 0.0002999998492307575,
+      "loss": 0.4796,
+      "step": 4220
+    },
+    {
+      "epoch": 0.000592972310856552,
+      "grad_norm": 0.578125,
+      "learning_rate": 0.0002999998482925568,
+      "loss": 0.4227,
+      "step": 4230
+    },
+    {
+      "epoch": 0.0005943741366505391,
+      "grad_norm": 0.87109375,
+      "learning_rate": 0.000299999847351446,
+      "loss": 0.4143,
+      "step": 4240
+    },
+    {
+      "epoch": 0.0005957759624445262,
+      "grad_norm": 0.46875,
+      "learning_rate": 0.0002999998464074252,
+      "loss": 0.4414,
+      "step": 4250
+    },
+    {
+      "epoch": 0.0005971777882385134,
+      "grad_norm": 0.361328125,
+      "learning_rate": 0.0002999998454604944,
+      "loss": 0.4819,
+      "step": 4260
+    },
+    {
+      "epoch": 0.0005985796140325005,
+      "grad_norm": 0.64453125,
+      "learning_rate": 0.00029999984451065353,
+      "loss": 0.3205,
+      "step": 4270
+    },
+    {
+      "epoch": 0.0005999814398264876,
+      "grad_norm": 0.80859375,
+      "learning_rate": 0.0002999998435579026,
+      "loss": 0.443,
+      "step": 4280
+    },
+    {
+      "epoch": 0.0006013832656204748,
+      "grad_norm": 0.76953125,
+      "learning_rate": 0.00029999984260224157,
+      "loss": 0.4969,
+      "step": 4290
+    },
+    {
+      "epoch": 0.0006027850914144618,
+      "grad_norm": 0.703125,
+      "learning_rate": 0.0002999998416436705,
+      "loss": 0.5088,
+      "step": 4300
+    },
+    {
+      "epoch": 0.000604186917208449,
+      "grad_norm": 0.7890625,
+      "learning_rate": 0.00029999984068218935,
+      "loss": 0.544,
+      "step": 4310
+    },
+    {
+      "epoch": 0.0006055887430024361,
+      "grad_norm": 0.482421875,
+      "learning_rate": 0.00029999983971779824,
+      "loss": 0.4043,
+      "step": 4320
+    },
+    {
+      "epoch": 0.0006069905687964232,
+      "grad_norm": 0.94921875,
+      "learning_rate": 0.000299999838750497,
+      "loss": 0.4532,
+      "step": 4330
+    },
+    {
+      "epoch": 0.0006083923945904103,
+      "grad_norm": 0.61328125,
+      "learning_rate": 0.0002999998377802858,
+      "loss": 0.4442,
+      "step": 4340
+    },
+    {
+      "epoch": 0.0006097942203843975,
+      "grad_norm": 0.52734375,
+      "learning_rate": 0.0002999998368071645,
+      "loss": 0.4439,
+      "step": 4350
+    },
+    {
+      "epoch": 0.0006111960461783846,
+      "grad_norm": 1.0390625,
+      "learning_rate": 0.00029999983583113316,
+      "loss": 0.429,
+      "step": 4360
+    },
+    {
+      "epoch": 0.0006125978719723717,
+      "grad_norm": 0.30078125,
+      "learning_rate": 0.00029999983485219173,
+      "loss": 0.4306,
+      "step": 4370
+    },
+    {
+      "epoch": 0.0006139996977663589,
+      "grad_norm": 0.2890625,
+      "learning_rate": 0.0002999998338703403,
+      "loss": 0.4207,
+      "step": 4380
+    },
+    {
+      "epoch": 0.0006154015235603459,
+      "grad_norm": 1.1953125,
+      "learning_rate": 0.00029999983288557877,
+      "loss": 0.4525,
+      "step": 4390
+    },
+    {
+      "epoch": 0.0006168033493543331,
+      "grad_norm": 0.7265625,
+      "learning_rate": 0.00029999983189790724,
+      "loss": 0.3968,
+      "step": 4400
+    },
+    {
+      "epoch": 0.0006182051751483201,
+      "grad_norm": 0.3203125,
+      "learning_rate": 0.00029999983090732565,
+      "loss": 0.3892,
+      "step": 4410
+    },
+    {
+      "epoch": 0.0006196070009423073,
+      "grad_norm": 0.255859375,
+      "learning_rate": 0.00029999982991383396,
+      "loss": 0.3471,
+      "step": 4420
+    },
+    {
+      "epoch": 0.0006210088267362944,
+      "grad_norm": 0.63671875,
+      "learning_rate": 0.00029999982891743226,
+      "loss": 0.5504,
+      "step": 4430
+    },
+    {
+      "epoch": 0.0006224106525302815,
+      "grad_norm": 0.330078125,
+      "learning_rate": 0.0002999998279181205,
+      "loss": 0.434,
+      "step": 4440
+    },
+    {
+      "epoch": 0.0006238124783242687,
+      "grad_norm": 1.09375,
+      "learning_rate": 0.0002999998269158987,
+      "loss": 0.5546,
+      "step": 4450
+    },
+    {
+      "epoch": 0.0006252143041182558,
+      "grad_norm": 0.78125,
+      "learning_rate": 0.00029999982591076687,
+      "loss": 0.4573,
+      "step": 4460
+    },
+    {
+      "epoch": 0.0006266161299122429,
+      "grad_norm": 1.0078125,
+      "learning_rate": 0.00029999982490272497,
+      "loss": 0.4395,
+      "step": 4470
+    },
+    {
+      "epoch": 0.00062801795570623,
+      "grad_norm": 0.48828125,
+      "learning_rate": 0.000299999823891773,
+      "loss": 0.4691,
+      "step": 4480
+    },
+    {
+      "epoch": 0.0006294197815002172,
+      "grad_norm": 0.38671875,
+      "learning_rate": 0.00029999982287791106,
+      "loss": 0.4948,
+      "step": 4490
+    },
+    {
+      "epoch": 0.0006308216072942042,
+      "grad_norm": 0.59765625,
+      "learning_rate": 0.000299999821861139,
+      "loss": 0.3723,
+      "step": 4500
+    },
+    {
+      "epoch": 0.0006308216072942042,
+      "eval_loss": 0.19076591730117798,
+      "eval_runtime": 112.1557,
+      "eval_samples_per_second": 203.672,
+      "eval_steps_per_second": 1.596,
+      "step": 4500
+    },
+    {
+      "epoch": 0.0006322234330881914,
+      "grad_norm": 0.61328125,
+      "learning_rate": 0.0002999998208414569,
+      "loss": 0.4265,
+      "step": 4510
+    },
+    {
+      "epoch": 0.0006336252588821785,
+      "grad_norm": 1.078125,
+      "learning_rate": 0.00029999981981886477,
+      "loss": 0.506,
+      "step": 4520
+    },
+    {
+      "epoch": 0.0006350270846761656,
+      "grad_norm": 0.59375,
+      "learning_rate": 0.0002999998187933626,
+      "loss": 0.4459,
+      "step": 4530
+    },
+    {
+      "epoch": 0.0006364289104701527,
+      "grad_norm": 0.84375,
+      "learning_rate": 0.00029999981776495033,
+      "loss": 0.506,
+      "step": 4540
+    },
+    {
+      "epoch": 0.0006378307362641399,
+      "grad_norm": 0.275390625,
+      "learning_rate": 0.000299999816733628,
+      "loss": 0.4536,
+      "step": 4550
+    },
+    {
+      "epoch": 0.000639232562058127,
+      "grad_norm": 0.86328125,
+      "learning_rate": 0.0002999998156993957,
+      "loss": 0.554,
+      "step": 4560
+    },
+    {
+      "epoch": 0.0006406343878521141,
+      "grad_norm": 0.84765625,
+      "learning_rate": 0.0002999998146622533,
+      "loss": 0.5602,
+      "step": 4570
+    },
+    {
+      "epoch": 0.0006420362136461013,
+      "grad_norm": 0.578125,
+      "learning_rate": 0.0002999998136222009,
+      "loss": 0.471,
+      "step": 4580
+    },
+    {
+      "epoch": 0.0006434380394400883,
+      "grad_norm": 0.7578125,
+      "learning_rate": 0.0002999998125792384,
+      "loss": 0.49,
+      "step": 4590
+    },
+    {
+      "epoch": 0.0006448398652340755,
+      "grad_norm": 0.86328125,
+      "learning_rate": 0.00029999981153336586,
+      "loss": 0.4289,
+      "step": 4600
+    },
+    {
+      "epoch": 0.0006462416910280626,
+      "grad_norm": 0.82421875,
+      "learning_rate": 0.00029999981048458327,
+      "loss": 0.4727,
+      "step": 4610
+    },
+    {
+      "epoch": 0.0006476435168220497,
+      "grad_norm": 0.40625,
+      "learning_rate": 0.0002999998094328906,
+      "loss": 0.4323,
+      "step": 4620
+    },
+    {
+      "epoch": 0.0006490453426160368,
+      "grad_norm": 0.921875,
+      "learning_rate": 0.00029999980837828793,
+      "loss": 0.4516,
+      "step": 4630
+    },
+    {
+      "epoch": 0.000650447168410024,
+      "grad_norm": 0.828125,
+      "learning_rate": 0.0002999998073207752,
+      "loss": 0.4945,
+      "step": 4640
+    },
+    {
+      "epoch": 0.0006518489942040111,
+      "grad_norm": 0.87109375,
+      "learning_rate": 0.00029999980626035244,
+      "loss": 0.4593,
+      "step": 4650
+    },
+    {
+      "epoch": 0.0006532508199979982,
+      "grad_norm": 0.6953125,
+      "learning_rate": 0.0002999998051970196,
+      "loss": 0.3685,
+      "step": 4660
+    },
+    {
+      "epoch": 0.0006546526457919854,
+      "grad_norm": 0.466796875,
+      "learning_rate": 0.00029999980413077674,
+      "loss": 0.4393,
+      "step": 4670
+    },
+    {
+      "epoch": 0.0006560544715859724,
+      "grad_norm": 0.80859375,
+      "learning_rate": 0.0002999998030616238,
+      "loss": 0.4165,
+      "step": 4680
+    },
+    {
+      "epoch": 0.0006574562973799596,
+      "grad_norm": 0.52734375,
+      "learning_rate": 0.0002999998019895608,
+      "loss": 0.4665,
+      "step": 4690
+    },
+    {
+      "epoch": 0.0006588581231739466,
+      "grad_norm": 0.765625,
+      "learning_rate": 0.0002999998009145878,
+      "loss": 0.544,
+      "step": 4700
+    },
+    {
+      "epoch": 0.0006602599489679338,
+      "grad_norm": 0.7578125,
+      "learning_rate": 0.0002999997998367047,
+      "loss": 0.5659,
+      "step": 4710
+    },
+    {
+      "epoch": 0.0006616617747619209,
+      "grad_norm": 0.7578125,
+      "learning_rate": 0.0002999997987559116,
+      "loss": 0.4516,
+      "step": 4720
+    },
+    {
+      "epoch": 0.000663063600555908,
+      "grad_norm": 0.69921875,
+      "learning_rate": 0.0002999997976722084,
+      "loss": 0.4434,
+      "step": 4730
+    },
+    {
+      "epoch": 0.0006644654263498952,
+      "grad_norm": 0.51953125,
+      "learning_rate": 0.0002999997965855952,
+      "loss": 0.4576,
+      "step": 4740
+    },
+    {
+      "epoch": 0.0006658672521438823,
+      "grad_norm": 0.396484375,
+      "learning_rate": 0.0002999997954960719,
+      "loss": 0.4261,
+      "step": 4750
+    },
+    {
+      "epoch": 0.0006672690779378694,
+      "grad_norm": 0.55078125,
+      "learning_rate": 0.0002999997944036386,
+      "loss": 0.477,
+      "step": 4760
+    },
+    {
+      "epoch": 0.0006686709037318565,
+      "grad_norm": 0.671875,
+      "learning_rate": 0.0002999997933082952,
+      "loss": 0.4698,
+      "step": 4770
+    },
+    {
+      "epoch": 0.0006700727295258437,
+      "grad_norm": 0.61328125,
+      "learning_rate": 0.0002999997922100418,
+      "loss": 0.4822,
+      "step": 4780
+    },
+    {
+      "epoch": 0.0006714745553198307,
+      "grad_norm": 0.5390625,
+      "learning_rate": 0.00029999979110887834,
+      "loss": 0.4478,
+      "step": 4790
+    },
+    {
+      "epoch": 0.0006728763811138179,
+      "grad_norm": 0.373046875,
+      "learning_rate": 0.0002999997900048048,
+      "loss": 0.4292,
+      "step": 4800
+    },
+    {
+      "epoch": 0.000674278206907805,
+      "grad_norm": 0.4921875,
+      "learning_rate": 0.00029999978889782127,
+      "loss": 0.5094,
+      "step": 4810
+    },
+    {
+      "epoch": 0.0006756800327017921,
+      "grad_norm": 0.337890625,
+      "learning_rate": 0.0002999997877879277,
+      "loss": 0.4333,
+      "step": 4820
+    },
+    {
+      "epoch": 0.0006770818584957792,
+      "grad_norm": 0.474609375,
+      "learning_rate": 0.00029999978667512403,
+      "loss": 0.4495,
+      "step": 4830
+    },
+    {
+      "epoch": 0.0006784836842897664,
+      "grad_norm": 0.6171875,
+      "learning_rate": 0.0002999997855594103,
+      "loss": 0.4684,
+      "step": 4840
+    },
+    {
+      "epoch": 0.0006798855100837535,
+      "grad_norm": 0.263671875,
+      "learning_rate": 0.0002999997844407865,
+      "loss": 0.3566,
+      "step": 4850
+    },
+    {
+      "epoch": 0.0006812873358777406,
+      "grad_norm": 0.69140625,
+      "learning_rate": 0.0002999997833192527,
+      "loss": 0.4387,
+      "step": 4860
+    },
+    {
+      "epoch": 0.0006826891616717278,
+      "grad_norm": 1.2265625,
+      "learning_rate": 0.0002999997821948089,
+      "loss": 0.4308,
+      "step": 4870
+    },
+    {
+      "epoch": 0.0006840909874657148,
+      "grad_norm": 0.5859375,
+      "learning_rate": 0.0002999997810674549,
+      "loss": 0.5234,
+      "step": 4880
+    },
+    {
+      "epoch": 0.000685492813259702,
+      "grad_norm": 0.234375,
+      "learning_rate": 0.000299999779937191,
+      "loss": 0.5019,
+      "step": 4890
+    },
+    {
+      "epoch": 0.000686894639053689,
+      "grad_norm": 0.51953125,
+      "learning_rate": 0.000299999778804017,
+      "loss": 0.4867,
+      "step": 4900
+    },
+    {
+      "epoch": 0.0006882964648476762,
+      "grad_norm": 0.9921875,
+      "learning_rate": 0.00029999977766793294,
+      "loss": 0.5128,
+      "step": 4910
+    },
+    {
+      "epoch": 0.0006896982906416633,
+      "grad_norm": 1.921875,
+      "learning_rate": 0.0002999997765289389,
+      "loss": 0.5317,
+      "step": 4920
+    },
+    {
+      "epoch": 0.0006911001164356504,
+      "grad_norm": 0.78125,
+      "learning_rate": 0.0002999997753870347,
+      "loss": 0.5652,
+      "step": 4930
+    },
+    {
+      "epoch": 0.0006925019422296376,
+      "grad_norm": 0.90625,
+      "learning_rate": 0.0002999997742422205,
+      "loss": 0.4115,
+      "step": 4940
+    },
+    {
+      "epoch": 0.0006939037680236247,
+      "grad_norm": 0.8359375,
+      "learning_rate": 0.00029999977309449626,
+      "loss": 0.5186,
+      "step": 4950
+    },
+    {
+      "epoch": 0.0006953055938176118,
+      "grad_norm": 0.93359375,
+      "learning_rate": 0.000299999771943862,
+      "loss": 0.5305,
+      "step": 4960
+    },
+    {
+      "epoch": 0.0006967074196115989,
+      "grad_norm": 0.9140625,
+      "learning_rate": 0.0002999997707903176,
+      "loss": 0.4705,
+      "step": 4970
+    },
+    {
+      "epoch": 0.0006981092454055861,
+      "grad_norm": 0.5859375,
+      "learning_rate": 0.0002999997696338632,
+      "loss": 0.4838,
+      "step": 4980
+    },
+    {
+      "epoch": 0.0006995110711995731,
+      "grad_norm": 0.40625,
+      "learning_rate": 0.0002999997684744988,
+      "loss": 0.5191,
+      "step": 4990
+    },
+    {
+      "epoch": 0.0007009128969935603,
+      "grad_norm": 0.478515625,
+      "learning_rate": 0.00029999976731222435,
+      "loss": 0.4744,
+      "step": 5000
+    },
+    {
+      "epoch": 0.0007009128969935603,
+      "eval_loss": 0.1905738264322281,
+      "eval_runtime": 113.7667,
+      "eval_samples_per_second": 200.788,
+      "eval_steps_per_second": 1.573,
+      "step": 5000
+    },
+    {
+      "epoch": 0.0007023147227875474,
+      "grad_norm": 0.6640625,
+      "learning_rate": 0.00029999976614703976,
+      "loss": 0.5915,
+      "step": 5010
+    },
+    {
+      "epoch": 0.0007037165485815345,
+      "grad_norm": 0.62109375,
+      "learning_rate": 0.00029999976497894516,
+      "loss": 0.5237,
+      "step": 5020
+    },
+    {
+      "epoch": 0.0007051183743755217,
+      "grad_norm": 0.6328125,
+      "learning_rate": 0.0002999997638079406,
+      "loss": 0.5251,
+      "step": 5030
+    },
+    {
+      "epoch": 0.0007065202001695088,
+      "grad_norm": 0.439453125,
+      "learning_rate": 0.0002999997626340259,
+      "loss": 0.4624,
+      "step": 5040
+    },
+    {
+      "epoch": 0.0007079220259634959,
+      "grad_norm": 0.953125,
+      "learning_rate": 0.0002999997614572012,
+      "loss": 0.4527,
+      "step": 5050
+    },
+    {
+      "epoch": 0.000709323851757483,
+      "grad_norm": 0.83203125,
+      "learning_rate": 0.0002999997602774664,
+      "loss": 0.4864,
+      "step": 5060
+    },
+    {
+      "epoch": 0.0007107256775514702,
+      "grad_norm": 1.046875,
+      "learning_rate": 0.00029999975909482157,
+      "loss": 0.4791,
+      "step": 5070
+    },
+    {
+      "epoch": 0.0007121275033454572,
+      "grad_norm": 0.2109375,
+      "learning_rate": 0.0002999997579092667,
+      "loss": 0.4468,
+      "step": 5080
+    },
+    {
+      "epoch": 0.0007135293291394444,
+      "grad_norm": 0.318359375,
+      "learning_rate": 0.0002999997567208018,
+      "loss": 0.4285,
+      "step": 5090
+    },
+    {
+      "epoch": 0.0007149311549334315,
+      "grad_norm": 0.30078125,
+      "learning_rate": 0.0002999997555294268,
+      "loss": 0.4075,
+      "step": 5100
+    },
+    {
+      "epoch": 0.0007163329807274186,
+      "grad_norm": 0.8125,
+      "learning_rate": 0.0002999997543351418,
+      "loss": 0.4852,
+      "step": 5110
+    },
+    {
+      "epoch": 0.0007177348065214057,
+      "grad_norm": 0.94140625,
+      "learning_rate": 0.0002999997531379467,
+      "loss": 0.4541,
+      "step": 5120
+    },
+    {
+      "epoch": 0.0007191366323153929,
+      "grad_norm": 0.9375,
+      "learning_rate": 0.00029999975193784165,
+      "loss": 0.5359,
+      "step": 5130
+    },
+    {
+      "epoch": 0.00072053845810938,
+      "grad_norm": 0.8125,
+      "learning_rate": 0.0002999997507348265,
+      "loss": 0.507,
+      "step": 5140
+    },
+    {
+      "epoch": 0.0007219402839033671,
+      "grad_norm": 0.796875,
+      "learning_rate": 0.00029999974952890125,
+      "loss": 0.5468,
+      "step": 5150
+    },
+    {
+      "epoch": 0.0007233421096973543,
+      "grad_norm": 0.8515625,
+      "learning_rate": 0.00029999974832006603,
+      "loss": 0.3978,
+      "step": 5160
+    },
+    {
+      "epoch": 0.0007247439354913413,
+      "grad_norm": 0.2109375,
+      "learning_rate": 0.0002999997471083207,
+      "loss": 0.3386,
+      "step": 5170
+    },
+    {
+      "epoch": 0.0007261457612853285,
+      "grad_norm": 0.396484375,
+      "learning_rate": 0.0002999997458936653,
+      "loss": 0.4484,
+      "step": 5180
+    },
+    {
+      "epoch": 0.0007275475870793155,
+      "grad_norm": 0.69921875,
+      "learning_rate": 0.00029999974467609993,
+      "loss": 0.4583,
+      "step": 5190
+    },
+    {
+      "epoch": 0.0007289494128733027,
+      "grad_norm": 0.33203125,
+      "learning_rate": 0.0002999997434556245,
+      "loss": 0.4929,
+      "step": 5200
+    },
+    {
+      "epoch": 0.0007303512386672898,
+      "grad_norm": 0.92578125,
+      "learning_rate": 0.00029999974223223895,
+      "loss": 0.4117,
+      "step": 5210
+    },
+    {
+      "epoch": 0.0007317530644612769,
+      "grad_norm": 0.51953125,
+      "learning_rate": 0.0002999997410059434,
+      "loss": 0.4758,
+      "step": 5220
+    },
+    {
+      "epoch": 0.0007331548902552641,
+      "grad_norm": 0.5546875,
+      "learning_rate": 0.00029999973977673787,
+      "loss": 0.38,
+      "step": 5230
+    },
+    {
+      "epoch": 0.0007345567160492512,
+      "grad_norm": 0.88671875,
+      "learning_rate": 0.0002999997385446222,
+      "loss": 0.4684,
+      "step": 5240
+    },
+    {
+      "epoch": 0.0007359585418432383,
+      "grad_norm": 0.7109375,
+      "learning_rate": 0.0002999997373095965,
+      "loss": 0.4076,
+      "step": 5250
+    },
+    {
+      "epoch": 0.0007373603676372254,
+      "grad_norm": 0.439453125,
+      "learning_rate": 0.0002999997360716608,
+      "loss": 0.5375,
+      "step": 5260
+    },
+    {
+      "epoch": 0.0007387621934312126,
+      "grad_norm": 0.546875,
+      "learning_rate": 0.000299999734830815,
+      "loss": 0.4483,
+      "step": 5270
+    },
+    {
+      "epoch": 0.0007401640192251996,
+      "grad_norm": 1.1328125,
+      "learning_rate": 0.0002999997335870592,
+      "loss": 0.4744,
+      "step": 5280
+    },
+    {
+      "epoch": 0.0007415658450191868,
+      "grad_norm": 0.6953125,
+      "learning_rate": 0.0002999997323403933,
+      "loss": 0.4671,
+      "step": 5290
+    },
+    {
+      "epoch": 0.0007429676708131739,
+      "grad_norm": 0.609375,
+      "learning_rate": 0.0002999997310908174,
+      "loss": 0.4767,
+      "step": 5300
+    },
+    {
+      "epoch": 0.000744369496607161,
+      "grad_norm": 1.015625,
+      "learning_rate": 0.0002999997298383314,
+      "loss": 0.5018,
+      "step": 5310
+    },
+    {
+      "epoch": 0.0007457713224011482,
+      "grad_norm": 0.79296875,
+      "learning_rate": 0.0002999997285829354,
+      "loss": 0.4238,
+      "step": 5320
+    },
+    {
+      "epoch": 0.0007471731481951353,
+      "grad_norm": 0.78125,
+      "learning_rate": 0.00029999972732462936,
+      "loss": 0.4321,
+      "step": 5330
+    },
+    {
+      "epoch": 0.0007485749739891224,
+      "grad_norm": 0.271484375,
+      "learning_rate": 0.00029999972606341323,
+      "loss": 0.4522,
+      "step": 5340
+    },
+    {
+      "epoch": 0.0007499767997831095,
+      "grad_norm": 0.5078125,
+      "learning_rate": 0.00029999972479928706,
+      "loss": 0.4748,
+      "step": 5350
+    },
+    {
+      "epoch": 0.0007513786255770967,
+      "grad_norm": 0.357421875,
+      "learning_rate": 0.00029999972353225083,
+      "loss": 0.4502,
+      "step": 5360
+    },
+    {
+      "epoch": 0.0007527804513710837,
+      "grad_norm": 0.80859375,
+      "learning_rate": 0.0002999997222623046,
+      "loss": 0.5169,
+      "step": 5370
+    },
+    {
+      "epoch": 0.0007541822771650709,
+      "grad_norm": 0.68359375,
+      "learning_rate": 0.00029999972098944833,
+      "loss": 0.4289,
+      "step": 5380
+    },
+    {
+      "epoch": 0.000755584102959058,
+      "grad_norm": 0.67578125,
+      "learning_rate": 0.00029999971971368195,
+      "loss": 0.4885,
+      "step": 5390
+    },
+    {
+      "epoch": 0.0007569859287530451,
+      "grad_norm": 0.3203125,
+      "learning_rate": 0.0002999997184350055,
+      "loss": 0.4621,
+      "step": 5400
+    },
+    {
+      "epoch": 0.0007583877545470322,
+      "grad_norm": 0.314453125,
+      "learning_rate": 0.0002999997171534191,
+      "loss": 0.3875,
+      "step": 5410
+    },
+    {
+      "epoch": 0.0007597895803410194,
+      "grad_norm": 0.54296875,
+      "learning_rate": 0.0002999997158689226,
+      "loss": 0.3891,
+      "step": 5420
+    },
+    {
+      "epoch": 0.0007611914061350065,
+      "grad_norm": 0.408203125,
+      "learning_rate": 0.00029999971458151604,
+      "loss": 0.4734,
+      "step": 5430
+    },
+    {
+      "epoch": 0.0007625932319289936,
+      "grad_norm": 1.046875,
+      "learning_rate": 0.00029999971329119945,
+      "loss": 0.5127,
+      "step": 5440
+    },
+    {
+      "epoch": 0.0007639950577229808,
+      "grad_norm": 0.63671875,
+      "learning_rate": 0.00029999971199797285,
+      "loss": 0.5229,
+      "step": 5450
+    },
+    {
+      "epoch": 0.0007653968835169678,
+      "grad_norm": 0.73046875,
+      "learning_rate": 0.00029999971070183615,
+      "loss": 0.5307,
+      "step": 5460
+    },
+    {
+      "epoch": 0.000766798709310955,
+      "grad_norm": 0.77734375,
+      "learning_rate": 0.0002999997094027894,
+      "loss": 0.4848,
+      "step": 5470
+    },
+    {
+      "epoch": 0.000768200535104942,
+      "grad_norm": 0.80859375,
+      "learning_rate": 0.00029999970810083265,
+      "loss": 0.4901,
+      "step": 5480
+    },
+    {
+      "epoch": 0.0007696023608989292,
+      "grad_norm": 0.369140625,
+      "learning_rate": 0.00029999970679596584,
+      "loss": 0.4517,
+      "step": 5490
+    },
+    {
+      "epoch": 0.0007710041866929163,
+      "grad_norm": 0.263671875,
+      "learning_rate": 0.000299999705488189,
+      "loss": 0.4509,
+      "step": 5500
+    },
+    {
+      "epoch": 0.0007710041866929163,
+      "eval_loss": 0.18971732258796692,
+      "eval_runtime": 112.0563,
+      "eval_samples_per_second": 203.853,
+      "eval_steps_per_second": 1.597,
+      "step": 5500
+    },
+    {
+      "epoch": 0.0007724060124869034,
+      "grad_norm": 0.72265625,
+      "learning_rate": 0.000299999704177502,
+      "loss": 0.4598,
+      "step": 5510
+    },
+    {
+      "epoch": 0.0007738078382808906,
+      "grad_norm": 0.365234375,
+      "learning_rate": 0.00029999970286390506,
+      "loss": 0.4504,
+      "step": 5520
+    },
+    {
+      "epoch": 0.0007752096640748777,
+      "grad_norm": 0.81640625,
+      "learning_rate": 0.0002999997015473981,
+      "loss": 0.4226,
+      "step": 5530
+    },
+    {
+      "epoch": 0.0007766114898688648,
+      "grad_norm": 0.78125,
+      "learning_rate": 0.000299999700227981,
+      "loss": 0.4202,
+      "step": 5540
+    },
+    {
+      "epoch": 0.0007780133156628519,
+      "grad_norm": 0.59765625,
+      "learning_rate": 0.0002999996989056539,
+      "loss": 0.4346,
+      "step": 5550
+    },
+    {
+      "epoch": 0.0007794151414568391,
+      "grad_norm": 0.392578125,
+      "learning_rate": 0.00029999969758041673,
+      "loss": 0.3287,
+      "step": 5560
+    },
+    {
+      "epoch": 0.0007808169672508261,
+      "grad_norm": 0.6640625,
+      "learning_rate": 0.00029999969625226956,
+      "loss": 0.5162,
+      "step": 5570
+    },
+    {
+      "epoch": 0.0007822187930448133,
+      "grad_norm": 0.70703125,
+      "learning_rate": 0.0002999996949212123,
+      "loss": 0.4915,
+      "step": 5580
+    },
+    {
+      "epoch": 0.0007836206188388004,
+      "grad_norm": 0.28125,
+      "learning_rate": 0.000299999693587245,
+      "loss": 0.4213,
+      "step": 5590
+    },
+    {
+      "epoch": 0.0007850224446327875,
+      "grad_norm": 0.94140625,
+      "learning_rate": 0.00029999969225036766,
+      "loss": 0.3892,
+      "step": 5600
+    },
+    {
+      "epoch": 0.0007864242704267747,
+      "grad_norm": 0.921875,
+      "learning_rate": 0.0002999996909105803,
+      "loss": 0.4821,
+      "step": 5610
+    },
+    {
+      "epoch": 0.0007878260962207618,
+      "grad_norm": 0.423828125,
+      "learning_rate": 0.00029999968956788284,
+      "loss": 0.3946,
+      "step": 5620
+    },
+    {
+      "epoch": 0.0007892279220147489,
+      "grad_norm": 0.349609375,
+      "learning_rate": 0.00029999968822227535,
+      "loss": 0.4474,
+      "step": 5630
+    },
+    {
+      "epoch": 0.000790629747808736,
+      "grad_norm": 0.54296875,
+      "learning_rate": 0.00029999968687375786,
+      "loss": 0.4803,
+      "step": 5640
+    },
+    {
+      "epoch": 0.0007920315736027232,
+      "grad_norm": 0.515625,
+      "learning_rate": 0.00029999968552233027,
+      "loss": 0.4669,
+      "step": 5650
+    },
+    {
+      "epoch": 0.0007934333993967102,
+      "grad_norm": 0.287109375,
+      "learning_rate": 0.0002999996841679926,
+      "loss": 0.3951,
+      "step": 5660
+    },
+    {
+      "epoch": 0.0007948352251906974,
+      "grad_norm": 0.275390625,
+      "learning_rate": 0.00029999968281074497,
+      "loss": 0.3465,
+      "step": 5670
+    },
+    {
+      "epoch": 0.0007962370509846845,
+      "grad_norm": 0.37109375,
+      "learning_rate": 0.00029999968145058727,
+      "loss": 0.4077,
+      "step": 5680
+    },
+    {
+      "epoch": 0.0007976388767786716,
+      "grad_norm": 0.37890625,
+      "learning_rate": 0.0002999996800875195,
+      "loss": 0.4023,
+      "step": 5690
+    },
+    {
+      "epoch": 0.0007990407025726587,
+      "grad_norm": 0.5234375,
+      "learning_rate": 0.00029999967872154166,
+      "loss": 0.4003,
+      "step": 5700
+    },
+    {
+      "epoch": 0.0008004425283666459,
+      "grad_norm": 0.8046875,
+      "learning_rate": 0.00029999967735265385,
+      "loss": 0.494,
+      "step": 5710
+    },
+    {
+      "epoch": 0.000801844354160633,
+      "grad_norm": 0.7734375,
+      "learning_rate": 0.00029999967598085594,
+      "loss": 0.4937,
+      "step": 5720
+    },
+    {
+      "epoch": 0.0008032461799546201,
+      "grad_norm": 0.65625,
+      "learning_rate": 0.000299999674606148,
+      "loss": 0.4274,
+      "step": 5730
+    },
+    {
+      "epoch": 0.0008046480057486072,
+      "grad_norm": 0.71875,
+      "learning_rate": 0.00029999967322853,
+      "loss": 0.4245,
+      "step": 5740
+    },
+    {
+      "epoch": 0.0008060498315425943,
+      "grad_norm": 0.8125,
+      "learning_rate": 0.00029999967184800194,
+      "loss": 0.5633,
+      "step": 5750
+    },
+    {
+      "epoch": 0.0008074516573365815,
+      "grad_norm": 0.70703125,
+      "learning_rate": 0.00029999967046456387,
+      "loss": 0.4766,
+      "step": 5760
+    },
+    {
+      "epoch": 0.0008088534831305685,
+      "grad_norm": 0.5625,
+      "learning_rate": 0.00029999966907821575,
+      "loss": 0.4312,
+      "step": 5770
+    },
+    {
+      "epoch": 0.0008102553089245557,
+      "grad_norm": 0.45703125,
+      "learning_rate": 0.0002999996676889576,
+      "loss": 0.3684,
+      "step": 5780
+    },
+    {
+      "epoch": 0.0008116571347185428,
+      "grad_norm": 0.80859375,
+      "learning_rate": 0.0002999996662967893,
+      "loss": 0.3765,
+      "step": 5790
+    },
+    {
+      "epoch": 0.0008130589605125299,
+      "grad_norm": 0.46484375,
+      "learning_rate": 0.00029999966490171107,
+      "loss": 0.4691,
+      "step": 5800
+    },
+    {
+      "epoch": 0.0008144607863065171,
+      "grad_norm": 0.75390625,
+      "learning_rate": 0.00029999966350372273,
+      "loss": 0.4451,
+      "step": 5810
+    },
+    {
+      "epoch": 0.0008158626121005042,
+      "grad_norm": 0.302734375,
+      "learning_rate": 0.00029999966210282435,
+      "loss": 0.4453,
+      "step": 5820
+    },
+    {
+      "epoch": 0.0008172644378944913,
+      "grad_norm": 0.423828125,
+      "learning_rate": 0.00029999966069901596,
+      "loss": 0.3543,
+      "step": 5830
+    },
+    {
+      "epoch": 0.0008186662636884784,
+      "grad_norm": 0.86328125,
+      "learning_rate": 0.0002999996592922975,
+      "loss": 0.4092,
+      "step": 5840
+    },
+    {
+      "epoch": 0.0008200680894824656,
+      "grad_norm": 0.71875,
+      "learning_rate": 0.00029999965788266904,
+      "loss": 0.4338,
+      "step": 5850
+    },
+    {
+      "epoch": 0.0008214699152764526,
+      "grad_norm": 0.287109375,
+      "learning_rate": 0.00029999965647013044,
+      "loss": 0.4996,
+      "step": 5860
+    },
+    {
+      "epoch": 0.0008228717410704398,
+      "grad_norm": 0.66015625,
+      "learning_rate": 0.00029999965505468184,
+      "loss": 0.4116,
+      "step": 5870
+    },
+    {
+      "epoch": 0.0008242735668644269,
+      "grad_norm": 0.267578125,
+      "learning_rate": 0.00029999965363632325,
+      "loss": 0.4168,
+      "step": 5880
+    },
+    {
+      "epoch": 0.000825675392658414,
+      "grad_norm": 0.68359375,
+      "learning_rate": 0.00029999965221505455,
+      "loss": 0.4213,
+      "step": 5890
+    },
+    {
+      "epoch": 0.0008270772184524012,
+      "grad_norm": 0.9296875,
+      "learning_rate": 0.0002999996507908758,
+      "loss": 0.4801,
+      "step": 5900
+    },
+    {
+      "epoch": 0.0008284790442463883,
+      "grad_norm": 1.0703125,
+      "learning_rate": 0.000299999649363787,
+      "loss": 0.4743,
+      "step": 5910
+    },
+    {
+      "epoch": 0.0008298808700403754,
+      "grad_norm": 0.22265625,
+      "learning_rate": 0.0002999996479337882,
+      "loss": 0.2799,
+      "step": 5920
+    },
+    {
+      "epoch": 0.0008312826958343625,
+      "grad_norm": 0.4609375,
+      "learning_rate": 0.0002999996465008793,
+      "loss": 0.5026,
+      "step": 5930
+    },
+    {
+      "epoch": 0.0008326845216283497,
+      "grad_norm": 1.03125,
+      "learning_rate": 0.0002999996450650604,
+      "loss": 0.4738,
+      "step": 5940
+    },
+    {
+      "epoch": 0.0008340863474223367,
+      "grad_norm": 0.60546875,
+      "learning_rate": 0.00029999964362633144,
+      "loss": 0.4826,
+      "step": 5950
+    },
+    {
+      "epoch": 0.0008354881732163239,
+      "grad_norm": 0.765625,
+      "learning_rate": 0.0002999996421846924,
+      "loss": 0.4934,
+      "step": 5960
+    },
+    {
+      "epoch": 0.000836889999010311,
+      "grad_norm": 0.38671875,
+      "learning_rate": 0.00029999964074014336,
+      "loss": 0.4993,
+      "step": 5970
+    },
+    {
+      "epoch": 0.0008382918248042981,
+      "grad_norm": 0.625,
+      "learning_rate": 0.0002999996392926843,
+      "loss": 0.4764,
+      "step": 5980
+    },
+    {
+      "epoch": 0.0008396936505982852,
+      "grad_norm": 0.58984375,
+      "learning_rate": 0.0002999996378423151,
+      "loss": 0.4652,
+      "step": 5990
+    },
+    {
+      "epoch": 0.0008410954763922723,
+      "grad_norm": 0.51953125,
+      "learning_rate": 0.00029999963638903594,
+      "loss": 0.465,
+      "step": 6000
+    },
+    {
+      "epoch": 0.0008410954763922723,
+      "eval_loss": 0.18518111109733582,
+      "eval_runtime": 112.0637,
+      "eval_samples_per_second": 203.839,
+      "eval_steps_per_second": 1.597,
+      "step": 6000
+    },
+    {
+      "epoch": 0.0008424973021862595,
+      "grad_norm": 1.0,
+      "learning_rate": 0.0002999996349328467,
+      "loss": 0.5146,
+      "step": 6010
+    },
+    {
+      "epoch": 0.0008438991279802466,
+      "grad_norm": 0.369140625,
+      "learning_rate": 0.00029999963347374743,
+      "loss": 0.4153,
+      "step": 6020
+    },
+    {
+      "epoch": 0.0008453009537742337,
+      "grad_norm": 0.84375,
+      "learning_rate": 0.0002999996320117381,
+      "loss": 0.4145,
+      "step": 6030
+    },
+    {
+      "epoch": 0.0008467027795682208,
+      "grad_norm": 0.298828125,
+      "learning_rate": 0.0002999996305468187,
+      "loss": 0.4206,
+      "step": 6040
+    },
+    {
+      "epoch": 0.000848104605362208,
+      "grad_norm": 0.6875,
+      "learning_rate": 0.00029999962907898927,
+      "loss": 0.4394,
+      "step": 6050
+    },
+    {
+      "epoch": 0.000849506431156195,
+      "grad_norm": 0.416015625,
+      "learning_rate": 0.0002999996276082498,
+      "loss": 0.3766,
+      "step": 6060
+    },
+    {
+      "epoch": 0.0008509082569501822,
+      "grad_norm": 0.42578125,
+      "learning_rate": 0.0002999996261346003,
+      "loss": 0.4398,
+      "step": 6070
+    },
+    {
+      "epoch": 0.0008523100827441693,
+      "grad_norm": 0.58203125,
+      "learning_rate": 0.00029999962465804075,
+      "loss": 0.4544,
+      "step": 6080
+    },
+    {
+      "epoch": 0.0008537119085381564,
+      "grad_norm": 0.609375,
+      "learning_rate": 0.00029999962317857115,
+      "loss": 0.4677,
+      "step": 6090
+    },
+    {
+      "epoch": 0.0008551137343321436,
+      "grad_norm": 0.84375,
+      "learning_rate": 0.0002999996216961915,
+      "loss": 0.4984,
+      "step": 6100
+    },
+    {
+      "epoch": 0.0008565155601261307,
+      "grad_norm": 0.53515625,
+      "learning_rate": 0.0002999996202109018,
+      "loss": 0.4111,
+      "step": 6110
+    },
+    {
+      "epoch": 0.0008579173859201178,
+      "grad_norm": 0.68359375,
+      "learning_rate": 0.00029999961872270205,
+      "loss": 0.4226,
+      "step": 6120
+    },
+    {
+      "epoch": 0.0008593192117141049,
+      "grad_norm": 0.70703125,
+      "learning_rate": 0.0002999996172315923,
+      "loss": 0.491,
+      "step": 6130
+    },
+    {
+      "epoch": 0.0008607210375080921,
+      "grad_norm": 0.5625,
+      "learning_rate": 0.00029999961573757243,
+      "loss": 0.4407,
+      "step": 6140
+    },
+    {
+      "epoch": 0.0008621228633020791,
+      "grad_norm": 0.63671875,
+      "learning_rate": 0.0002999996142406426,
+      "loss": 0.4048,
+      "step": 6150
+    },
+    {
+      "epoch": 0.0008635246890960663,
+      "grad_norm": 0.75390625,
+      "learning_rate": 0.00029999961274080266,
+      "loss": 0.3733,
+      "step": 6160
+    },
+    {
+      "epoch": 0.0008649265148900534,
+      "grad_norm": 0.58984375,
+      "learning_rate": 0.0002999996112380527,
+      "loss": 0.4596,
+      "step": 6170
+    },
+    {
+      "epoch": 0.0008663283406840405,
+      "grad_norm": 0.57421875,
+      "learning_rate": 0.0002999996097323927,
+      "loss": 0.4819,
+      "step": 6180
+    },
+    {
+      "epoch": 0.0008677301664780277,
+      "grad_norm": 0.30859375,
+      "learning_rate": 0.0002999996082238226,
+      "loss": 0.4029,
+      "step": 6190
+    },
+    {
+      "epoch": 0.0008691319922720148,
+      "grad_norm": 0.65234375,
+      "learning_rate": 0.00029999960671234254,
+      "loss": 0.5001,
+      "step": 6200
+    },
+    {
+      "epoch": 0.0008705338180660019,
+      "grad_norm": 0.88671875,
+      "learning_rate": 0.00029999960519795237,
+      "loss": 0.4356,
+      "step": 6210
+    },
+    {
+      "epoch": 0.000871935643859989,
+      "grad_norm": 0.97265625,
+      "learning_rate": 0.0002999996036806522,
+      "loss": 0.4399,
+      "step": 6220
+    },
+    {
+      "epoch": 0.0008733374696539762,
+      "grad_norm": 0.76171875,
+      "learning_rate": 0.00029999960216044196,
+      "loss": 0.5213,
+      "step": 6230
+    },
+    {
+      "epoch": 0.0008747392954479632,
+      "grad_norm": 1.1640625,
+      "learning_rate": 0.00029999960063732163,
+      "loss": 0.4467,
+      "step": 6240
+    },
+    {
+      "epoch": 0.0008761411212419504,
+      "grad_norm": 1.1015625,
+      "learning_rate": 0.00029999959911129135,
+      "loss": 0.5033,
+      "step": 6250
+    },
+    {
+      "epoch": 0.0008775429470359374,
+      "grad_norm": 0.330078125,
+      "learning_rate": 0.00029999959758235097,
+      "loss": 0.3733,
+      "step": 6260
+    },
+    {
+      "epoch": 0.0008789447728299246,
+      "grad_norm": 0.19921875,
+      "learning_rate": 0.0002999995960505005,
+      "loss": 0.3974,
+      "step": 6270
+    },
+    {
+      "epoch": 0.0008803465986239117,
+      "grad_norm": 0.33984375,
+      "learning_rate": 0.0002999995945157401,
+      "loss": 0.4224,
+      "step": 6280
+    },
+    {
+      "epoch": 0.0008817484244178988,
+      "grad_norm": 0.53515625,
+      "learning_rate": 0.00029999959297806955,
+      "loss": 0.4442,
+      "step": 6290
+    },
+    {
+      "epoch": 0.000883150250211886,
+      "grad_norm": 0.46484375,
+      "learning_rate": 0.000299999591437489,
+      "loss": 0.4996,
+      "step": 6300
+    },
+    {
+      "epoch": 0.0008845520760058731,
+      "grad_norm": 0.9375,
+      "learning_rate": 0.0002999995898939984,
+      "loss": 0.4069,
+      "step": 6310
+    },
+    {
+      "epoch": 0.0008859539017998602,
+      "grad_norm": 0.64453125,
+      "learning_rate": 0.00029999958834759776,
+      "loss": 0.4179,
+      "step": 6320
+    },
+    {
+      "epoch": 0.0008873557275938473,
+      "grad_norm": 0.79296875,
+      "learning_rate": 0.0002999995867982871,
+      "loss": 0.5006,
+      "step": 6330
+    },
+    {
+      "epoch": 0.0008887575533878345,
+      "grad_norm": 0.74609375,
+      "learning_rate": 0.00029999958524606636,
+      "loss": 0.46,
+      "step": 6340
+    },
+    {
+      "epoch": 0.0008901593791818215,
+      "grad_norm": 0.232421875,
+      "learning_rate": 0.0002999995836909356,
+      "loss": 0.4165,
+      "step": 6350
+    },
+    {
+      "epoch": 0.0008915612049758087,
+      "grad_norm": 0.98046875,
+      "learning_rate": 0.00029999958213289474,
+      "loss": 0.3791,
+      "step": 6360
+    },
+    {
+      "epoch": 0.0008929630307697958,
+      "grad_norm": 0.77734375,
+      "learning_rate": 0.0002999995805719439,
+      "loss": 0.3416,
+      "step": 6370
+    },
+    {
+      "epoch": 0.0008943648565637829,
+      "grad_norm": 0.6328125,
+      "learning_rate": 0.000299999579008083,
+      "loss": 0.4323,
+      "step": 6380
+    },
+    {
+      "epoch": 0.0008957666823577701,
+      "grad_norm": 0.4375,
+      "learning_rate": 0.000299999577441312,
+      "loss": 0.4569,
+      "step": 6390
+    },
+    {
+      "epoch": 0.0008971685081517572,
+      "grad_norm": 0.546875,
+      "learning_rate": 0.000299999575871631,
+      "loss": 0.4431,
+      "step": 6400
+    },
+    {
+      "epoch": 0.0008985703339457443,
+      "grad_norm": 1.1171875,
+      "learning_rate": 0.00029999957429904,
+      "loss": 0.5003,
+      "step": 6410
+    },
+    {
+      "epoch": 0.0008999721597397314,
+      "grad_norm": 0.85546875,
+      "learning_rate": 0.0002999995727235389,
+      "loss": 0.4411,
+      "step": 6420
+    },
+    {
+      "epoch": 0.0009013739855337186,
+      "grad_norm": 0.69921875,
+      "learning_rate": 0.0002999995711451278,
+      "loss": 0.4253,
+      "step": 6430
+    },
+    {
+      "epoch": 0.0009027758113277056,
+      "grad_norm": 0.98828125,
+      "learning_rate": 0.0002999995695638066,
+      "loss": 0.4617,
+      "step": 6440
+    },
+    {
+      "epoch": 0.0009041776371216928,
+      "grad_norm": 0.4375,
+      "learning_rate": 0.0002999995679795754,
+      "loss": 0.4213,
+      "step": 6450
+    },
+    {
+      "epoch": 0.0009055794629156799,
+      "grad_norm": 1.3828125,
+      "learning_rate": 0.0002999995663924341,
+      "loss": 0.471,
+      "step": 6460
+    },
+    {
+      "epoch": 0.000906981288709667,
+      "grad_norm": 0.6015625,
+      "learning_rate": 0.00029999956480238277,
+      "loss": 0.4532,
+      "step": 6470
+    },
+    {
+      "epoch": 0.0009083831145036542,
+      "grad_norm": 0.9609375,
+      "learning_rate": 0.00029999956320942144,
+      "loss": 0.4206,
+      "step": 6480
+    },
+    {
+      "epoch": 0.0009097849402976413,
+      "grad_norm": 0.94921875,
+      "learning_rate": 0.00029999956161355006,
+      "loss": 0.5678,
+      "step": 6490
+    },
+    {
+      "epoch": 0.0009111867660916284,
+      "grad_norm": 0.8359375,
+      "learning_rate": 0.00029999956001476857,
+      "loss": 0.3739,
+      "step": 6500
+    },
+    {
+      "epoch": 0.0009111867660916284,
+      "eval_loss": 0.18671706318855286,
+      "eval_runtime": 113.805,
+      "eval_samples_per_second": 200.721,
+      "eval_steps_per_second": 1.573,
+      "step": 6500
+    },
+    {
+      "epoch": 0.0009125885918856155,
+      "grad_norm": 0.3515625,
+      "learning_rate": 0.00029999955841307713,
+      "loss": 0.4547,
+      "step": 6510
+    },
+    {
+      "epoch": 0.0009139904176796027,
+      "grad_norm": 0.8359375,
+      "learning_rate": 0.0002999995568084756,
+      "loss": 0.4481,
+      "step": 6520
+    },
+    {
+      "epoch": 0.0009153922434735897,
+      "grad_norm": 0.58203125,
+      "learning_rate": 0.000299999555200964,
+      "loss": 0.4814,
+      "step": 6530
+    },
+    {
+      "epoch": 0.0009167940692675769,
+      "grad_norm": 0.40625,
+      "learning_rate": 0.0002999995535905424,
+      "loss": 0.4352,
+      "step": 6540
+    },
+    {
+      "epoch": 0.0009181958950615639,
+      "grad_norm": 0.88671875,
+      "learning_rate": 0.0002999995519772107,
+      "loss": 0.4045,
+      "step": 6550
+    },
+    {
+      "epoch": 0.0009195977208555511,
+      "grad_norm": 0.6875,
+      "learning_rate": 0.000299999550360969,
+      "loss": 0.4183,
+      "step": 6560
+    },
+    {
+      "epoch": 0.0009209995466495382,
+      "grad_norm": 0.58984375,
+      "learning_rate": 0.00029999954874181724,
+      "loss": 0.3972,
+      "step": 6570
+    },
+    {
+      "epoch": 0.0009224013724435253,
+      "grad_norm": 0.78515625,
+      "learning_rate": 0.0002999995471197555,
+      "loss": 0.488,
+      "step": 6580
+    },
+    {
+      "epoch": 0.0009238031982375125,
+      "grad_norm": 0.83203125,
+      "learning_rate": 0.00029999954549478363,
+      "loss": 0.4826,
+      "step": 6590
+    },
+    {
+      "epoch": 0.0009252050240314996,
+      "grad_norm": 0.3671875,
+      "learning_rate": 0.0002999995438669017,
+      "loss": 0.3605,
+      "step": 6600
+    },
+    {
+      "epoch": 0.0009266068498254867,
+      "grad_norm": 0.255859375,
+      "learning_rate": 0.00029999954223610976,
+      "loss": 0.3911,
+      "step": 6610
+    },
+    {
+      "epoch": 0.0009280086756194738,
+      "grad_norm": 0.4140625,
+      "learning_rate": 0.0002999995406024078,
+      "loss": 0.5245,
+      "step": 6620
+    },
+    {
+      "epoch": 0.000929410501413461,
+      "grad_norm": 0.390625,
+      "learning_rate": 0.00029999953896579583,
+      "loss": 0.3988,
+      "step": 6630
+    },
+    {
+      "epoch": 0.000930812327207448,
+      "grad_norm": 0.50390625,
+      "learning_rate": 0.00029999953732627377,
+      "loss": 0.522,
+      "step": 6640
+    },
+    {
+      "epoch": 0.0009322141530014352,
+      "grad_norm": 0.66015625,
+      "learning_rate": 0.00029999953568384164,
+      "loss": 0.462,
+      "step": 6650
+    },
+    {
+      "epoch": 0.0009336159787954223,
+      "grad_norm": 0.6015625,
+      "learning_rate": 0.00029999953403849947,
+      "loss": 0.4395,
+      "step": 6660
+    },
+    {
+      "epoch": 0.0009350178045894094,
+      "grad_norm": 0.69921875,
+      "learning_rate": 0.0002999995323902473,
+      "loss": 0.5009,
+      "step": 6670
+    },
+    {
+      "epoch": 0.0009364196303833966,
+      "grad_norm": 0.890625,
+      "learning_rate": 0.00029999953073908507,
+      "loss": 0.4336,
+      "step": 6680
+    },
+    {
+      "epoch": 0.0009378214561773837,
+      "grad_norm": 0.609375,
+      "learning_rate": 0.0002999995290850128,
+      "loss": 0.4126,
+      "step": 6690
+    },
+    {
+      "epoch": 0.0009392232819713708,
+      "grad_norm": 0.359375,
+      "learning_rate": 0.00029999952742803046,
+      "loss": 0.4957,
+      "step": 6700
+    },
+    {
+      "epoch": 0.0009406251077653579,
+      "grad_norm": 0.8359375,
+      "learning_rate": 0.0002999995257681381,
+      "loss": 0.4758,
+      "step": 6710
+    },
+    {
+      "epoch": 0.0009420269335593451,
+      "grad_norm": 0.314453125,
+      "learning_rate": 0.0002999995241053357,
+      "loss": 0.3698,
+      "step": 6720
+    },
+    {
+      "epoch": 0.0009434287593533321,
+      "grad_norm": 0.462890625,
+      "learning_rate": 0.0002999995224396232,
+      "loss": 0.4371,
+      "step": 6730
+    },
+    {
+      "epoch": 0.0009448305851473193,
+      "grad_norm": 0.390625,
+      "learning_rate": 0.00029999952077100077,
+      "loss": 0.4538,
+      "step": 6740
+    },
+    {
+      "epoch": 0.0009462324109413064,
+      "grad_norm": 0.61328125,
+      "learning_rate": 0.0002999995190994682,
+      "loss": 0.4969,
+      "step": 6750
+    },
+    {
+      "epoch": 0.0009476342367352935,
+      "grad_norm": 1.2109375,
+      "learning_rate": 0.0002999995174250256,
+      "loss": 0.4484,
+      "step": 6760
+    },
+    {
+      "epoch": 0.0009490360625292807,
+      "grad_norm": 0.5859375,
+      "learning_rate": 0.000299999515747673,
+      "loss": 0.4748,
+      "step": 6770
+    },
+    {
+      "epoch": 0.0009504378883232677,
+      "grad_norm": 0.84765625,
+      "learning_rate": 0.00029999951406741034,
+      "loss": 0.5366,
+      "step": 6780
+    },
+    {
+      "epoch": 0.0009518397141172549,
+      "grad_norm": 0.26171875,
+      "learning_rate": 0.0002999995123842376,
+      "loss": 0.4718,
+      "step": 6790
+    },
+    {
+      "epoch": 0.000953241539911242,
+      "grad_norm": 0.6796875,
+      "learning_rate": 0.00029999951069815484,
+      "loss": 0.3908,
+      "step": 6800
+    },
+    {
+      "epoch": 0.0009546433657052291,
+      "grad_norm": 0.796875,
+      "learning_rate": 0.00029999950900916203,
+      "loss": 0.4005,
+      "step": 6810
+    },
+    {
+      "epoch": 0.0009560451914992162,
+      "grad_norm": 0.515625,
+      "learning_rate": 0.0002999995073172592,
+      "loss": 0.4265,
+      "step": 6820
+    },
+    {
+      "epoch": 0.0009574470172932034,
+      "grad_norm": 0.62890625,
+      "learning_rate": 0.0002999995056224463,
+      "loss": 0.4673,
+      "step": 6830
+    },
+    {
+      "epoch": 0.0009588488430871904,
+      "grad_norm": 0.28515625,
+      "learning_rate": 0.00029999950392472336,
+      "loss": 0.3959,
+      "step": 6840
+    },
+    {
+      "epoch": 0.0009602506688811776,
+      "grad_norm": 0.8828125,
+      "learning_rate": 0.0002999995022240904,
+      "loss": 0.5393,
+      "step": 6850
+    },
+    {
+      "epoch": 0.0009616524946751647,
+      "grad_norm": 0.828125,
+      "learning_rate": 0.0002999995005205474,
+      "loss": 0.4281,
+      "step": 6860
+    },
+    {
+      "epoch": 0.0009630543204691518,
+      "grad_norm": 0.6796875,
+      "learning_rate": 0.0002999994988140943,
+      "loss": 0.378,
+      "step": 6870
+    },
+    {
+      "epoch": 0.000964456146263139,
+      "grad_norm": 0.7734375,
+      "learning_rate": 0.0002999994971047312,
+      "loss": 0.445,
+      "step": 6880
+    },
+    {
+      "epoch": 0.0009658579720571261,
+      "grad_norm": 0.57421875,
+      "learning_rate": 0.0002999994953924581,
+      "loss": 0.4868,
+      "step": 6890
+    },
+    {
+      "epoch": 0.0009672597978511132,
+      "grad_norm": 0.93359375,
+      "learning_rate": 0.0002999994936772749,
+      "loss": 0.479,
+      "step": 6900
+    },
+    {
+      "epoch": 0.0009686616236451003,
+      "grad_norm": 0.921875,
+      "learning_rate": 0.0002999994919591817,
+      "loss": 0.5115,
+      "step": 6910
+    },
+    {
+      "epoch": 0.0009700634494390875,
+      "grad_norm": 0.890625,
+      "learning_rate": 0.0002999994902381784,
+      "loss": 0.4955,
+      "step": 6920
+    },
+    {
+      "epoch": 0.0009714652752330745,
+      "grad_norm": 1.34375,
+      "learning_rate": 0.0002999994885142651,
+      "loss": 0.5283,
+      "step": 6930
+    },
+    {
+      "epoch": 0.0009728671010270617,
+      "grad_norm": 0.458984375,
+      "learning_rate": 0.00029999948678744175,
+      "loss": 0.5511,
+      "step": 6940
+    },
+    {
+      "epoch": 0.0009742689268210488,
+      "grad_norm": 0.7109375,
+      "learning_rate": 0.0002999994850577083,
+      "loss": 0.4193,
+      "step": 6950
+    },
+    {
+      "epoch": 0.0009756707526150359,
+      "grad_norm": 0.7734375,
+      "learning_rate": 0.0002999994833250649,
+      "loss": 0.4947,
+      "step": 6960
+    },
+    {
+      "epoch": 0.000977072578409023,
+      "grad_norm": 0.4140625,
+      "learning_rate": 0.0002999994815895114,
+      "loss": 0.3735,
+      "step": 6970
+    },
+    {
+      "epoch": 0.0009784744042030103,
+      "grad_norm": 0.34765625,
+      "learning_rate": 0.00029999947985104785,
+      "loss": 0.4227,
+      "step": 6980
+    },
+    {
+      "epoch": 0.0009798762299969973,
+      "grad_norm": 0.65625,
+      "learning_rate": 0.00029999947810967426,
+      "loss": 0.3318,
+      "step": 6990
+    },
+    {
+      "epoch": 0.0009812780557909844,
+      "grad_norm": 0.80078125,
+      "learning_rate": 0.0002999994763653906,
+      "loss": 0.4498,
+      "step": 7000
+    },
+    {
+      "epoch": 0.0009812780557909844,
+      "eval_loss": 0.1828116476535797,
+      "eval_runtime": 112.0643,
+      "eval_samples_per_second": 203.838,
+      "eval_steps_per_second": 1.597,
+      "step": 7000
+    },
+    {
+      "epoch": 0.0009826798815849714,
+      "grad_norm": 0.62890625,
+      "learning_rate": 0.00029999947461819697,
+      "loss": 0.4011,
+      "step": 7010
+    },
+    {
+      "epoch": 0.0009840817073789587,
+      "grad_norm": 1.2109375,
+      "learning_rate": 0.00029999947286809327,
+      "loss": 0.5584,
+      "step": 7020
+    },
+    {
+      "epoch": 0.0009854835331729458,
+      "grad_norm": 0.4296875,
+      "learning_rate": 0.0002999994711150795,
+      "loss": 0.4164,
+      "step": 7030
+    },
+    {
+      "epoch": 0.0009868853589669328,
+      "grad_norm": 0.58984375,
+      "learning_rate": 0.0002999994693591557,
+      "loss": 0.4351,
+      "step": 7040
+    },
+    {
+      "epoch": 0.00098828718476092,
+      "grad_norm": 0.5703125,
+      "learning_rate": 0.0002999994676003219,
+      "loss": 0.3533,
+      "step": 7050
+    },
+    {
+      "epoch": 0.0009896890105549072,
+      "grad_norm": 0.423828125,
+      "learning_rate": 0.00029999946583857806,
+      "loss": 0.4176,
+      "step": 7060
+    },
+    {
+      "epoch": 0.0009910908363488942,
+      "grad_norm": 0.412109375,
+      "learning_rate": 0.0002999994640739241,
+      "loss": 0.4999,
+      "step": 7070
+    },
+    {
+      "epoch": 0.0009924926621428813,
+      "grad_norm": 0.68359375,
+      "learning_rate": 0.00029999946230636014,
+      "loss": 0.4877,
+      "step": 7080
+    },
+    {
+      "epoch": 0.0009938944879368686,
+      "grad_norm": 0.62890625,
+      "learning_rate": 0.0002999994605358862,
+      "loss": 0.473,
+      "step": 7090
+    },
+    {
+      "epoch": 0.0009952963137308556,
+      "grad_norm": 1.0,
+      "learning_rate": 0.0002999994587625022,
+      "loss": 0.4986,
+      "step": 7100
+    },
+    {
+      "epoch": 0.0009966981395248427,
+      "grad_norm": 0.51953125,
+      "learning_rate": 0.00029999945698620806,
+      "loss": 0.3922,
+      "step": 7110
+    },
+    {
+      "epoch": 0.0009980999653188298,
+      "grad_norm": 1.0234375,
+      "learning_rate": 0.00029999945520700394,
+      "loss": 0.4771,
+      "step": 7120
+    },
+    {
+      "epoch": 0.000999501791112817,
+      "grad_norm": 0.8515625,
+      "learning_rate": 0.0002999994534248897,
+      "loss": 0.441,
+      "step": 7130
+    },
+    {
+      "epoch": 0.001000903616906804,
+      "grad_norm": 0.72265625,
+      "learning_rate": 0.00029999945163986555,
+      "loss": 0.5059,
+      "step": 7140
+    },
+    {
+      "epoch": 0.0010023054427007912,
+      "grad_norm": 0.80859375,
+      "learning_rate": 0.00029999944985193127,
+      "loss": 0.465,
+      "step": 7150
+    },
+    {
+      "epoch": 0.0010037072684947784,
+      "grad_norm": 0.78515625,
+      "learning_rate": 0.000299999448061087,
+      "loss": 0.4113,
+      "step": 7160
+    },
+    {
+      "epoch": 0.0010051090942887655,
+      "grad_norm": 0.953125,
+      "learning_rate": 0.00029999944626733267,
+      "loss": 0.5342,
+      "step": 7170
+    },
+    {
+      "epoch": 0.0010065109200827526,
+      "grad_norm": 0.56640625,
+      "learning_rate": 0.0002999994444706683,
+      "loss": 0.531,
+      "step": 7180
+    },
+    {
+      "epoch": 0.0010079127458767396,
+      "grad_norm": 0.51953125,
+      "learning_rate": 0.00029999944267109386,
+      "loss": 0.4291,
+      "step": 7190
+    },
+    {
+      "epoch": 0.001009314571670727,
+      "grad_norm": 0.91796875,
+      "learning_rate": 0.00029999944086860937,
+      "loss": 0.56,
+      "step": 7200
+    },
+    {
+      "epoch": 0.001010716397464714,
+      "grad_norm": 1.0859375,
+      "learning_rate": 0.0002999994390632149,
+      "loss": 0.4209,
+      "step": 7210
+    },
+    {
+      "epoch": 0.001012118223258701,
+      "grad_norm": 0.83203125,
+      "learning_rate": 0.0002999994372549103,
+      "loss": 0.5099,
+      "step": 7220
+    },
+    {
+      "epoch": 0.001013520049052688,
+      "grad_norm": 1.0625,
+      "learning_rate": 0.00029999943544369576,
+      "loss": 0.567,
+      "step": 7230
+    },
+    {
+      "epoch": 0.0010149218748466754,
+      "grad_norm": 0.7890625,
+      "learning_rate": 0.0002999994336295711,
+      "loss": 0.474,
+      "step": 7240
+    },
+    {
+      "epoch": 0.0010163237006406624,
+      "grad_norm": 0.68359375,
+      "learning_rate": 0.0002999994318125364,
+      "loss": 0.4313,
+      "step": 7250
+    },
+    {
+      "epoch": 0.0010177255264346495,
+      "grad_norm": 0.30078125,
+      "learning_rate": 0.00029999942999259173,
+      "loss": 0.4166,
+      "step": 7260
+    },
+    {
+      "epoch": 0.0010191273522286368,
+      "grad_norm": 0.9296875,
+      "learning_rate": 0.000299999428169737,
+      "loss": 0.3713,
+      "step": 7270
+    },
+    {
+      "epoch": 0.0010205291780226238,
+      "grad_norm": 0.78515625,
+      "learning_rate": 0.0002999994263439722,
+      "loss": 0.4229,
+      "step": 7280
+    },
+    {
+      "epoch": 0.0010219310038166109,
+      "grad_norm": 0.68359375,
+      "learning_rate": 0.00029999942451529733,
+      "loss": 0.4572,
+      "step": 7290
+    },
+    {
+      "epoch": 0.001023332829610598,
+      "grad_norm": 0.6015625,
+      "learning_rate": 0.0002999994226837125,
+      "loss": 0.4676,
+      "step": 7300
+    },
+    {
+      "epoch": 0.0010247346554045852,
+      "grad_norm": 0.85546875,
+      "learning_rate": 0.0002999994208492175,
+      "loss": 0.4612,
+      "step": 7310
+    },
+    {
+      "epoch": 0.0010261364811985723,
+      "grad_norm": 0.95703125,
+      "learning_rate": 0.00029999941901181256,
+      "loss": 0.4747,
+      "step": 7320
+    },
+    {
+      "epoch": 0.0010275383069925593,
+      "grad_norm": 0.61328125,
+      "learning_rate": 0.00029999941717149755,
+      "loss": 0.4067,
+      "step": 7330
+    },
+    {
+      "epoch": 0.0010289401327865464,
+      "grad_norm": 0.5078125,
+      "learning_rate": 0.0002999994153282725,
+      "loss": 0.433,
+      "step": 7340
+    },
+    {
+      "epoch": 0.0010303419585805337,
+      "grad_norm": 0.8828125,
+      "learning_rate": 0.00029999941348213743,
+      "loss": 0.4342,
+      "step": 7350
+    },
+    {
+      "epoch": 0.0010317437843745207,
+      "grad_norm": 0.52734375,
+      "learning_rate": 0.0002999994116330923,
+      "loss": 0.5533,
+      "step": 7360
+    },
+    {
+      "epoch": 0.0010331456101685078,
+      "grad_norm": 1.140625,
+      "learning_rate": 0.00029999940978113715,
+      "loss": 0.5318,
+      "step": 7370
+    },
+    {
+      "epoch": 0.001034547435962495,
+      "grad_norm": 0.84375,
+      "learning_rate": 0.00029999940792627193,
+      "loss": 0.4702,
+      "step": 7380
+    },
+    {
+      "epoch": 0.0010359492617564821,
+      "grad_norm": 0.83203125,
+      "learning_rate": 0.00029999940606849666,
+      "loss": 0.4446,
+      "step": 7390
+    },
+    {
+      "epoch": 0.0010373510875504692,
+      "grad_norm": 0.58984375,
+      "learning_rate": 0.00029999940420781133,
+      "loss": 0.4316,
+      "step": 7400
+    },
+    {
+      "epoch": 0.0010387529133444563,
+      "grad_norm": 0.498046875,
+      "learning_rate": 0.000299999402344216,
+      "loss": 0.378,
+      "step": 7410
+    },
+    {
+      "epoch": 0.0010401547391384435,
+      "grad_norm": 0.56640625,
+      "learning_rate": 0.00029999940047771063,
+      "loss": 0.4685,
+      "step": 7420
+    },
+    {
+      "epoch": 0.0010415565649324306,
+      "grad_norm": 0.2255859375,
+      "learning_rate": 0.0002999993986082952,
+      "loss": 0.4783,
+      "step": 7430
+    },
+    {
+      "epoch": 0.0010429583907264177,
+      "grad_norm": 0.58203125,
+      "learning_rate": 0.00029999939673596977,
+      "loss": 0.4002,
+      "step": 7440
+    },
+    {
+      "epoch": 0.001044360216520405,
+      "grad_norm": 0.8671875,
+      "learning_rate": 0.0002999993948607343,
+      "loss": 0.473,
+      "step": 7450
+    },
+    {
+      "epoch": 0.001045762042314392,
+      "grad_norm": 0.703125,
+      "learning_rate": 0.0002999993929825887,
+      "loss": 0.4573,
+      "step": 7460
+    },
+    {
+      "epoch": 0.001047163868108379,
+      "grad_norm": 0.59765625,
+      "learning_rate": 0.0002999993911015331,
+      "loss": 0.4299,
+      "step": 7470
+    },
+    {
+      "epoch": 0.0010485656939023661,
+      "grad_norm": 0.546875,
+      "learning_rate": 0.00029999938921756753,
+      "loss": 0.4994,
+      "step": 7480
+    },
+    {
+      "epoch": 0.0010499675196963534,
+      "grad_norm": 0.443359375,
+      "learning_rate": 0.00029999938733069184,
+      "loss": 0.4764,
+      "step": 7490
+    },
+    {
+      "epoch": 0.0010513693454903405,
+      "grad_norm": 0.1767578125,
+      "learning_rate": 0.00029999938544090615,
+      "loss": 0.3844,
+      "step": 7500
+    },
+    {
+      "epoch": 0.0010513693454903405,
+      "eval_loss": 0.18380461633205414,
+      "eval_runtime": 114.1658,
+      "eval_samples_per_second": 200.086,
+      "eval_steps_per_second": 1.568,
+      "step": 7500
+    },
+    {
+      "epoch": 0.0010527711712843275,
+      "grad_norm": 0.4921875,
+      "learning_rate": 0.0002999993835482104,
+      "loss": 0.4668,
+      "step": 7510
+    },
+    {
+      "epoch": 0.0010541729970783146,
+      "grad_norm": 0.41796875,
+      "learning_rate": 0.0002999993816526046,
+      "loss": 0.4184,
+      "step": 7520
+    },
+    {
+      "epoch": 0.0010555748228723019,
+      "grad_norm": 0.9296875,
+      "learning_rate": 0.00029999937975408876,
+      "loss": 0.4809,
+      "step": 7530
+    },
+    {
+      "epoch": 0.001056976648666289,
+      "grad_norm": 0.80078125,
+      "learning_rate": 0.0002999993778526629,
+      "loss": 0.5232,
+      "step": 7540
+    },
+    {
+      "epoch": 0.001058378474460276,
+      "grad_norm": 0.59765625,
+      "learning_rate": 0.00029999937594832695,
+      "loss": 0.4522,
+      "step": 7550
+    },
+    {
+      "epoch": 0.0010597803002542633,
+      "grad_norm": 0.68359375,
+      "learning_rate": 0.00029999937404108105,
+      "loss": 0.4973,
+      "step": 7560
+    },
+    {
+      "epoch": 0.0010611821260482503,
+      "grad_norm": 0.435546875,
+      "learning_rate": 0.00029999937213092505,
+      "loss": 0.4475,
+      "step": 7570
+    },
+    {
+      "epoch": 0.0010625839518422374,
+      "grad_norm": 0.7578125,
+      "learning_rate": 0.000299999370217859,
+      "loss": 0.5236,
+      "step": 7580
+    },
+    {
+      "epoch": 0.0010639857776362244,
+      "grad_norm": 0.9921875,
+      "learning_rate": 0.000299999368301883,
+      "loss": 0.4271,
+      "step": 7590
+    },
+    {
+      "epoch": 0.0010653876034302117,
+      "grad_norm": 0.93359375,
+      "learning_rate": 0.0002999993663829968,
+      "loss": 0.4317,
+      "step": 7600
+    },
+    {
+      "epoch": 0.0010667894292241988,
+      "grad_norm": 0.4375,
+      "learning_rate": 0.00029999936446120065,
+      "loss": 0.3837,
+      "step": 7610
+    },
+    {
+      "epoch": 0.0010681912550181858,
+      "grad_norm": 0.64453125,
+      "learning_rate": 0.00029999936253649444,
+      "loss": 0.4802,
+      "step": 7620
+    },
+    {
+      "epoch": 0.001069593080812173,
+      "grad_norm": 0.24609375,
+      "learning_rate": 0.0002999993606088783,
+      "loss": 0.4218,
+      "step": 7630
+    },
+    {
+      "epoch": 0.0010709949066061602,
+      "grad_norm": 0.419921875,
+      "learning_rate": 0.000299999358678352,
+      "loss": 0.3988,
+      "step": 7640
+    },
+    {
+      "epoch": 0.0010723967324001472,
+      "grad_norm": 0.67578125,
+      "learning_rate": 0.00029999935674491563,
+      "loss": 0.4898,
+      "step": 7650
+    },
+    {
+      "epoch": 0.0010737985581941343,
+      "grad_norm": 0.734375,
+      "learning_rate": 0.0002999993548085693,
+      "loss": 0.5254,
+      "step": 7660
+    },
+    {
+      "epoch": 0.0010752003839881216,
+      "grad_norm": 0.365234375,
+      "learning_rate": 0.0002999993528693129,
+      "loss": 0.3633,
+      "step": 7670
+    },
+    {
+      "epoch": 0.0010766022097821086,
+      "grad_norm": 0.6796875,
+      "learning_rate": 0.00029999935092714646,
+      "loss": 0.4217,
+      "step": 7680
+    },
+    {
+      "epoch": 0.0010780040355760957,
+      "grad_norm": 0.45703125,
+      "learning_rate": 0.00029999934898207003,
+      "loss": 0.4005,
+      "step": 7690
+    },
+    {
+      "epoch": 0.0010794058613700828,
+      "grad_norm": 0.765625,
+      "learning_rate": 0.00029999934703408345,
+      "loss": 0.5546,
+      "step": 7700
+    },
+    {
+      "epoch": 0.00108080768716407,
+      "grad_norm": 0.90625,
+      "learning_rate": 0.0002999993450831869,
+      "loss": 0.5362,
+      "step": 7710
+    },
+    {
+      "epoch": 0.001082209512958057,
+      "grad_norm": 0.53125,
+      "learning_rate": 0.00029999934312938034,
+      "loss": 0.4965,
+      "step": 7720
+    },
+    {
+      "epoch": 0.0010836113387520442,
+      "grad_norm": 0.8671875,
+      "learning_rate": 0.0002999993411726637,
+      "loss": 0.4596,
+      "step": 7730
+    },
+    {
+      "epoch": 0.0010850131645460314,
+      "grad_norm": 0.486328125,
+      "learning_rate": 0.000299999339213037,
+      "loss": 0.4036,
+      "step": 7740
+    },
+    {
+      "epoch": 0.0010864149903400185,
+      "grad_norm": 0.310546875,
+      "learning_rate": 0.00029999933725050027,
+      "loss": 0.3324,
+      "step": 7750
+    },
+    {
+      "epoch": 0.0010878168161340056,
+      "grad_norm": 0.88671875,
+      "learning_rate": 0.00029999933528505353,
+      "loss": 0.3971,
+      "step": 7760
+    },
+    {
+      "epoch": 0.0010892186419279926,
+      "grad_norm": 0.5703125,
+      "learning_rate": 0.00029999933331669674,
+      "loss": 0.4291,
+      "step": 7770
+    },
+    {
+      "epoch": 0.00109062046772198,
+      "grad_norm": 0.609375,
+      "learning_rate": 0.0002999993313454299,
+      "loss": 0.5129,
+      "step": 7780
+    },
+    {
+      "epoch": 0.001092022293515967,
+      "grad_norm": 0.828125,
+      "learning_rate": 0.00029999932937125305,
+      "loss": 0.4775,
+      "step": 7790
+    },
+    {
+      "epoch": 0.001093424119309954,
+      "grad_norm": 0.41796875,
+      "learning_rate": 0.0002999993273941661,
+      "loss": 0.3819,
+      "step": 7800
+    },
+    {
+      "epoch": 0.001094825945103941,
+      "grad_norm": 0.84375,
+      "learning_rate": 0.00029999932541416915,
+      "loss": 0.421,
+      "step": 7810
+    },
+    {
+      "epoch": 0.0010962277708979284,
+      "grad_norm": 0.255859375,
+      "learning_rate": 0.00029999932343126214,
+      "loss": 0.3671,
+      "step": 7820
+    },
+    {
+      "epoch": 0.0010976295966919154,
+      "grad_norm": 0.375,
+      "learning_rate": 0.0002999993214454451,
+      "loss": 0.3949,
+      "step": 7830
+    },
+    {
+      "epoch": 0.0010990314224859025,
+      "grad_norm": 0.3125,
+      "learning_rate": 0.00029999931945671804,
+      "loss": 0.4246,
+      "step": 7840
+    },
+    {
+      "epoch": 0.0011004332482798898,
+      "grad_norm": 0.353515625,
+      "learning_rate": 0.00029999931746508093,
+      "loss": 0.3755,
+      "step": 7850
+    },
+    {
+      "epoch": 0.0011018350740738768,
+      "grad_norm": 0.328125,
+      "learning_rate": 0.0002999993154705338,
+      "loss": 0.3755,
+      "step": 7860
+    },
+    {
+      "epoch": 0.0011032368998678639,
+      "grad_norm": 0.60546875,
+      "learning_rate": 0.0002999993134730766,
+      "loss": 0.382,
+      "step": 7870
+    },
+    {
+      "epoch": 0.001104638725661851,
+      "grad_norm": 0.515625,
+      "learning_rate": 0.00029999931147270934,
+      "loss": 0.4672,
+      "step": 7880
+    },
+    {
+      "epoch": 0.0011060405514558382,
+      "grad_norm": 0.72265625,
+      "learning_rate": 0.0002999993094694321,
+      "loss": 0.3952,
+      "step": 7890
+    },
+    {
+      "epoch": 0.0011074423772498253,
+      "grad_norm": 0.95703125,
+      "learning_rate": 0.00029999930746324476,
+      "loss": 0.4959,
+      "step": 7900
+    },
+    {
+      "epoch": 0.0011088442030438123,
+      "grad_norm": 0.28125,
+      "learning_rate": 0.0002999993054541474,
+      "loss": 0.3773,
+      "step": 7910
+    },
+    {
+      "epoch": 0.0011102460288377994,
+      "grad_norm": 0.46875,
+      "learning_rate": 0.00029999930344214,
+      "loss": 0.4186,
+      "step": 7920
+    },
+    {
+      "epoch": 0.0011116478546317867,
+      "grad_norm": 0.6953125,
+      "learning_rate": 0.0002999993014272226,
+      "loss": 0.4753,
+      "step": 7930
+    },
+    {
+      "epoch": 0.0011130496804257737,
+      "grad_norm": 0.66796875,
+      "learning_rate": 0.00029999929940939513,
+      "loss": 0.5497,
+      "step": 7940
+    },
+    {
+      "epoch": 0.0011144515062197608,
+      "grad_norm": 0.6328125,
+      "learning_rate": 0.0002999992973886576,
+      "loss": 0.4491,
+      "step": 7950
+    },
+    {
+      "epoch": 0.001115853332013748,
+      "grad_norm": 0.80859375,
+      "learning_rate": 0.0002999992953650101,
+      "loss": 0.3838,
+      "step": 7960
+    },
+    {
+      "epoch": 0.0011172551578077351,
+      "grad_norm": 0.5546875,
+      "learning_rate": 0.0002999992933384525,
+      "loss": 0.4518,
+      "step": 7970
+    },
+    {
+      "epoch": 0.0011186569836017222,
+      "grad_norm": 0.8515625,
+      "learning_rate": 0.0002999992913089848,
+      "loss": 0.4592,
+      "step": 7980
+    },
+    {
+      "epoch": 0.0011200588093957093,
+      "grad_norm": 0.333984375,
+      "learning_rate": 0.0002999992892766072,
+      "loss": 0.438,
+      "step": 7990
+    },
+    {
+      "epoch": 0.0011214606351896965,
+      "grad_norm": 0.2314453125,
+      "learning_rate": 0.0002999992872413195,
+      "loss": 0.4245,
+      "step": 8000
+    },
+    {
+      "epoch": 0.0011214606351896965,
+      "eval_loss": 0.18062791228294373,
+      "eval_runtime": 112.1678,
+      "eval_samples_per_second": 203.65,
+      "eval_steps_per_second": 1.596,
+      "step": 8000
+    },
+    {
+      "epoch": 0.0011228624609836836,
+      "grad_norm": 0.85546875,
+      "learning_rate": 0.00029999928520312177,
+      "loss": 0.3692,
+      "step": 8010
+    },
+    {
+      "epoch": 0.0011242642867776707,
+      "grad_norm": 0.189453125,
+      "learning_rate": 0.0002999992831620139,
+      "loss": 0.4953,
+      "step": 8020
+    },
+    {
+      "epoch": 0.001125666112571658,
+      "grad_norm": 0.66015625,
+      "learning_rate": 0.00029999928111799614,
+      "loss": 0.4422,
+      "step": 8030
+    },
+    {
+      "epoch": 0.001127067938365645,
+      "grad_norm": 0.443359375,
+      "learning_rate": 0.0002999992790710683,
+      "loss": 0.428,
+      "step": 8040
+    },
+    {
+      "epoch": 0.001128469764159632,
+      "grad_norm": 0.96875,
+      "learning_rate": 0.0002999992770212304,
+      "loss": 0.4894,
+      "step": 8050
+    },
+    {
+      "epoch": 0.0011298715899536191,
+      "grad_norm": 0.421875,
+      "learning_rate": 0.00029999927496848246,
+      "loss": 0.4164,
+      "step": 8060
+    },
+    {
+      "epoch": 0.0011312734157476064,
+      "grad_norm": 0.384765625,
+      "learning_rate": 0.00029999927291282446,
+      "loss": 0.4902,
+      "step": 8070
+    },
+    {
+      "epoch": 0.0011326752415415935,
+      "grad_norm": 1.0390625,
+      "learning_rate": 0.00029999927085425646,
+      "loss": 0.4481,
+      "step": 8080
+    },
+    {
+      "epoch": 0.0011340770673355805,
+      "grad_norm": 1.140625,
+      "learning_rate": 0.0002999992687927784,
+      "loss": 0.3717,
+      "step": 8090
+    },
+    {
+      "epoch": 0.0011354788931295676,
+      "grad_norm": 0.345703125,
+      "learning_rate": 0.0002999992667283903,
+      "loss": 0.4978,
+      "step": 8100
+    },
+    {
+      "epoch": 0.0011368807189235549,
+      "grad_norm": 0.97265625,
+      "learning_rate": 0.00029999926466109215,
+      "loss": 0.4372,
+      "step": 8110
+    },
+    {
+      "epoch": 0.001138282544717542,
+      "grad_norm": 0.73828125,
+      "learning_rate": 0.000299999262590884,
+      "loss": 0.4048,
+      "step": 8120
+    },
+    {
+      "epoch": 0.001139684370511529,
+      "grad_norm": 0.73828125,
+      "learning_rate": 0.0002999992605177658,
+      "loss": 0.4937,
+      "step": 8130
+    },
+    {
+      "epoch": 0.0011410861963055163,
+      "grad_norm": 0.65625,
+      "learning_rate": 0.0002999992584417376,
+      "loss": 0.4527,
+      "step": 8140
+    },
+    {
+      "epoch": 0.0011424880220995033,
+      "grad_norm": 0.609375,
+      "learning_rate": 0.0002999992563627993,
+      "loss": 0.502,
+      "step": 8150
+    },
+    {
+      "epoch": 0.0011438898478934904,
+      "grad_norm": 0.50390625,
+      "learning_rate": 0.0002999992542809509,
+      "loss": 0.4664,
+      "step": 8160
+    },
+    {
+      "epoch": 0.0011452916736874774,
+      "grad_norm": 0.7734375,
+      "learning_rate": 0.0002999992521961926,
+      "loss": 0.5214,
+      "step": 8170
+    },
+    {
+      "epoch": 0.0011466934994814647,
+      "grad_norm": 0.5546875,
+      "learning_rate": 0.0002999992501085242,
+      "loss": 0.4809,
+      "step": 8180
+    },
+    {
+      "epoch": 0.0011480953252754518,
+      "grad_norm": 0.54296875,
+      "learning_rate": 0.00029999924801794577,
+      "loss": 0.4875,
+      "step": 8190
+    },
+    {
+      "epoch": 0.0011494971510694388,
+      "grad_norm": 0.73046875,
+      "learning_rate": 0.0002999992459244573,
+      "loss": 0.4272,
+      "step": 8200
+    },
+    {
+      "epoch": 0.001150898976863426,
+      "grad_norm": 0.6796875,
+      "learning_rate": 0.0002999992438280588,
+      "loss": 0.5558,
+      "step": 8210
+    },
+    {
+      "epoch": 0.0011523008026574132,
+      "grad_norm": 1.0234375,
+      "learning_rate": 0.0002999992417287502,
+      "loss": 0.3699,
+      "step": 8220
+    },
+    {
+      "epoch": 0.0011537026284514002,
+      "grad_norm": 0.90625,
+      "learning_rate": 0.00029999923962653163,
+      "loss": 0.4134,
+      "step": 8230
+    },
+    {
+      "epoch": 0.0011551044542453873,
+      "grad_norm": 0.78125,
+      "learning_rate": 0.000299999237521403,
+      "loss": 0.3791,
+      "step": 8240
+    },
+    {
+      "epoch": 0.0011565062800393746,
+      "grad_norm": 0.57421875,
+      "learning_rate": 0.0002999992354133643,
+      "loss": 0.3503,
+      "step": 8250
+    },
+    {
+      "epoch": 0.0011579081058333616,
+      "grad_norm": 0.62109375,
+      "learning_rate": 0.00029999923330241565,
+      "loss": 0.4115,
+      "step": 8260
+    },
+    {
+      "epoch": 0.0011593099316273487,
+      "grad_norm": 0.240234375,
+      "learning_rate": 0.00029999923118855686,
+      "loss": 0.3909,
+      "step": 8270
+    },
+    {
+      "epoch": 0.0011607117574213358,
+      "grad_norm": 0.177734375,
+      "learning_rate": 0.0002999992290717881,
+      "loss": 0.4405,
+      "step": 8280
+    },
+    {
+      "epoch": 0.001162113583215323,
+      "grad_norm": 1.046875,
+      "learning_rate": 0.0002999992269521093,
+      "loss": 0.4961,
+      "step": 8290
+    },
+    {
+      "epoch": 0.00116351540900931,
+      "grad_norm": 0.72265625,
+      "learning_rate": 0.0002999992248295204,
+      "loss": 0.3559,
+      "step": 8300
+    },
+    {
+      "epoch": 0.0011649172348032972,
+      "grad_norm": 0.494140625,
+      "learning_rate": 0.0002999992227040215,
+      "loss": 0.4841,
+      "step": 8310
+    },
+    {
+      "epoch": 0.0011663190605972844,
+      "grad_norm": 0.34375,
+      "learning_rate": 0.0002999992205756126,
+      "loss": 0.4471,
+      "step": 8320
+    },
+    {
+      "epoch": 0.0011677208863912715,
+      "grad_norm": 0.671875,
+      "learning_rate": 0.00029999921844429364,
+      "loss": 0.4465,
+      "step": 8330
+    },
+    {
+      "epoch": 0.0011691227121852586,
+      "grad_norm": 0.97265625,
+      "learning_rate": 0.00029999921631006465,
+      "loss": 0.4335,
+      "step": 8340
+    },
+    {
+      "epoch": 0.0011705245379792456,
+      "grad_norm": 0.478515625,
+      "learning_rate": 0.0002999992141729256,
+      "loss": 0.4393,
+      "step": 8350
+    },
+    {
+      "epoch": 0.0011719263637732329,
+      "grad_norm": 1.0859375,
+      "learning_rate": 0.0002999992120328765,
+      "loss": 0.451,
+      "step": 8360
+    },
+    {
+      "epoch": 0.00117332818956722,
+      "grad_norm": 0.62890625,
+      "learning_rate": 0.0002999992098899174,
+      "loss": 0.432,
+      "step": 8370
+    },
+    {
+      "epoch": 0.001174730015361207,
+      "grad_norm": 0.609375,
+      "learning_rate": 0.0002999992077440482,
+      "loss": 0.4606,
+      "step": 8380
+    },
+    {
+      "epoch": 0.001176131841155194,
+      "grad_norm": 0.578125,
+      "learning_rate": 0.00029999920559526907,
+      "loss": 0.5024,
+      "step": 8390
+    },
+    {
+      "epoch": 0.0011775336669491814,
+      "grad_norm": 0.2314453125,
+      "learning_rate": 0.0002999992034435798,
+      "loss": 0.3843,
+      "step": 8400
+    },
+    {
+      "epoch": 0.0011789354927431684,
+      "grad_norm": 0.361328125,
+      "learning_rate": 0.0002999992012889805,
+      "loss": 0.4505,
+      "step": 8410
+    },
+    {
+      "epoch": 0.0011803373185371555,
+      "grad_norm": 1.0546875,
+      "learning_rate": 0.0002999991991314712,
+      "loss": 0.5677,
+      "step": 8420
+    },
+    {
+      "epoch": 0.0011817391443311428,
+      "grad_norm": 0.78515625,
+      "learning_rate": 0.0002999991969710519,
+      "loss": 0.4781,
+      "step": 8430
+    },
+    {
+      "epoch": 0.0011831409701251298,
+      "grad_norm": 0.75390625,
+      "learning_rate": 0.00029999919480772254,
+      "loss": 0.4487,
+      "step": 8440
+    },
+    {
+      "epoch": 0.0011845427959191169,
+      "grad_norm": 0.478515625,
+      "learning_rate": 0.00029999919264148307,
+      "loss": 0.4665,
+      "step": 8450
+    },
+    {
+      "epoch": 0.001185944621713104,
+      "grad_norm": 0.984375,
+      "learning_rate": 0.00029999919047233366,
+      "loss": 0.5278,
+      "step": 8460
+    },
+    {
+      "epoch": 0.0011873464475070912,
+      "grad_norm": 0.953125,
+      "learning_rate": 0.00029999918830027415,
+      "loss": 0.4261,
+      "step": 8470
+    },
+    {
+      "epoch": 0.0011887482733010783,
+      "grad_norm": 0.26171875,
+      "learning_rate": 0.0002999991861253046,
+      "loss": 0.4043,
+      "step": 8480
+    },
+    {
+      "epoch": 0.0011901500990950653,
+      "grad_norm": 0.91796875,
+      "learning_rate": 0.00029999918394742507,
+      "loss": 0.4822,
+      "step": 8490
+    },
+    {
+      "epoch": 0.0011915519248890524,
+      "grad_norm": 0.75,
+      "learning_rate": 0.00029999918176663545,
+      "loss": 0.5023,
+      "step": 8500
+    },
+    {
+      "epoch": 0.0011915519248890524,
+      "eval_loss": 0.18091367185115814,
+      "eval_runtime": 112.0263,
+      "eval_samples_per_second": 203.907,
+      "eval_steps_per_second": 1.598,
+      "step": 8500
+    },
+    {
+      "epoch": 0.0011929537506830397,
+      "grad_norm": 0.2490234375,
+      "learning_rate": 0.00029999917958293583,
+      "loss": 0.4,
+      "step": 8510
+    },
+    {
+      "epoch": 0.0011943555764770267,
+      "grad_norm": 0.32421875,
+      "learning_rate": 0.0002999991773963261,
+      "loss": 0.3551,
+      "step": 8520
+    },
+    {
+      "epoch": 0.0011957574022710138,
+      "grad_norm": 0.84375,
+      "learning_rate": 0.00029999917520680643,
+      "loss": 0.3701,
+      "step": 8530
+    },
+    {
+      "epoch": 0.001197159228065001,
+      "grad_norm": 0.31640625,
+      "learning_rate": 0.00029999917301437665,
+      "loss": 0.4104,
+      "step": 8540
+    },
+    {
+      "epoch": 0.0011985610538589881,
+      "grad_norm": 0.2578125,
+      "learning_rate": 0.0002999991708190369,
+      "loss": 0.4813,
+      "step": 8550
+    },
+    {
+      "epoch": 0.0011999628796529752,
+      "grad_norm": 0.310546875,
+      "learning_rate": 0.00029999916862078705,
+      "loss": 0.3847,
+      "step": 8560
+    },
+    {
+      "epoch": 0.0012013647054469623,
+      "grad_norm": 0.6640625,
+      "learning_rate": 0.0002999991664196272,
+      "loss": 0.4201,
+      "step": 8570
+    },
+    {
+      "epoch": 0.0012027665312409495,
+      "grad_norm": 0.59375,
+      "learning_rate": 0.00029999916421555734,
+      "loss": 0.4336,
+      "step": 8580
+    },
+    {
+      "epoch": 0.0012041683570349366,
+      "grad_norm": 0.451171875,
+      "learning_rate": 0.0002999991620085774,
+      "loss": 0.5009,
+      "step": 8590
+    },
+    {
+      "epoch": 0.0012055701828289237,
+      "grad_norm": 0.443359375,
+      "learning_rate": 0.00029999915979868737,
+      "loss": 0.4635,
+      "step": 8600
+    },
+    {
+      "epoch": 0.001206972008622911,
+      "grad_norm": 0.28515625,
+      "learning_rate": 0.0002999991575858874,
+      "loss": 0.3898,
+      "step": 8610
+    },
+    {
+      "epoch": 0.001208373834416898,
+      "grad_norm": 0.26953125,
+      "learning_rate": 0.00029999915537017735,
+      "loss": 0.35,
+      "step": 8620
+    },
+    {
+      "epoch": 0.001209775660210885,
+      "grad_norm": 0.46484375,
+      "learning_rate": 0.00029999915315155726,
+      "loss": 0.4431,
+      "step": 8630
+    },
+    {
+      "epoch": 0.0012111774860048721,
+      "grad_norm": 0.7890625,
+      "learning_rate": 0.00029999915093002717,
+      "loss": 0.4535,
+      "step": 8640
+    },
+    {
+      "epoch": 0.0012125793117988594,
+      "grad_norm": 0.455078125,
+      "learning_rate": 0.00029999914870558703,
+      "loss": 0.3979,
+      "step": 8650
+    },
+    {
+      "epoch": 0.0012139811375928464,
+      "grad_norm": 0.83203125,
+      "learning_rate": 0.00029999914647823683,
+      "loss": 0.4376,
+      "step": 8660
+    },
+    {
+      "epoch": 0.0012153829633868335,
+      "grad_norm": 0.73046875,
+      "learning_rate": 0.0002999991442479766,
+      "loss": 0.482,
+      "step": 8670
+    },
+    {
+      "epoch": 0.0012167847891808206,
+      "grad_norm": 0.78515625,
+      "learning_rate": 0.00029999914201480634,
+      "loss": 0.4178,
+      "step": 8680
+    },
+    {
+      "epoch": 0.0012181866149748078,
+      "grad_norm": 0.71875,
+      "learning_rate": 0.00029999913977872604,
+      "loss": 0.4327,
+      "step": 8690
+    },
+    {
+      "epoch": 0.001219588440768795,
+      "grad_norm": 0.97265625,
+      "learning_rate": 0.0002999991375397357,
+      "loss": 0.4699,
+      "step": 8700
+    },
+    {
+      "epoch": 0.001220990266562782,
+      "grad_norm": 1.0078125,
+      "learning_rate": 0.00029999913529783534,
+      "loss": 0.5145,
+      "step": 8710
+    },
+    {
+      "epoch": 0.0012223920923567692,
+      "grad_norm": 0.5546875,
+      "learning_rate": 0.00029999913305302494,
+      "loss": 0.4242,
+      "step": 8720
+    },
+    {
+      "epoch": 0.0012237939181507563,
+      "grad_norm": 0.421875,
+      "learning_rate": 0.0002999991308053045,
+      "loss": 0.3744,
+      "step": 8730
+    },
+    {
+      "epoch": 0.0012251957439447434,
+      "grad_norm": 0.4375,
+      "learning_rate": 0.000299999128554674,
+      "loss": 0.4844,
+      "step": 8740
+    },
+    {
+      "epoch": 0.0012265975697387304,
+      "grad_norm": 0.57421875,
+      "learning_rate": 0.0002999991263011335,
+      "loss": 0.4545,
+      "step": 8750
+    },
+    {
+      "epoch": 0.0012279993955327177,
+      "grad_norm": 0.99609375,
+      "learning_rate": 0.00029999912404468296,
+      "loss": 0.4804,
+      "step": 8760
+    },
+    {
+      "epoch": 0.0012294012213267048,
+      "grad_norm": 0.875,
+      "learning_rate": 0.0002999991217853224,
+      "loss": 0.47,
+      "step": 8770
+    },
+    {
+      "epoch": 0.0012308030471206918,
+      "grad_norm": 0.99609375,
+      "learning_rate": 0.00029999911952305173,
+      "loss": 0.4163,
+      "step": 8780
+    },
+    {
+      "epoch": 0.0012322048729146789,
+      "grad_norm": 0.65234375,
+      "learning_rate": 0.0002999991172578711,
+      "loss": 0.5008,
+      "step": 8790
+    },
+    {
+      "epoch": 0.0012336066987086662,
+      "grad_norm": 1.0546875,
+      "learning_rate": 0.0002999991149897804,
+      "loss": 0.5043,
+      "step": 8800
+    },
+    {
+      "epoch": 0.0012350085245026532,
+      "grad_norm": 1.15625,
+      "learning_rate": 0.0002999991127187797,
+      "loss": 0.4966,
+      "step": 8810
+    },
+    {
+      "epoch": 0.0012364103502966403,
+      "grad_norm": 0.5859375,
+      "learning_rate": 0.0002999991104448689,
+      "loss": 0.4113,
+      "step": 8820
+    },
+    {
+      "epoch": 0.0012378121760906276,
+      "grad_norm": 0.76171875,
+      "learning_rate": 0.0002999991081680481,
+      "loss": 0.419,
+      "step": 8830
+    },
+    {
+      "epoch": 0.0012392140018846146,
+      "grad_norm": 0.408203125,
+      "learning_rate": 0.0002999991058883173,
+      "loss": 0.423,
+      "step": 8840
+    },
+    {
+      "epoch": 0.0012406158276786017,
+      "grad_norm": 0.33203125,
+      "learning_rate": 0.0002999991036056764,
+      "loss": 0.4298,
+      "step": 8850
+    },
+    {
+      "epoch": 0.0012420176534725887,
+      "grad_norm": 0.7578125,
+      "learning_rate": 0.00029999910132012553,
+      "loss": 0.4679,
+      "step": 8860
+    },
+    {
+      "epoch": 0.001243419479266576,
+      "grad_norm": 0.8359375,
+      "learning_rate": 0.00029999909903166455,
+      "loss": 0.4492,
+      "step": 8870
+    },
+    {
+      "epoch": 0.001244821305060563,
+      "grad_norm": 0.490234375,
+      "learning_rate": 0.00029999909674029363,
+      "loss": 0.5003,
+      "step": 8880
+    },
+    {
+      "epoch": 0.0012462231308545501,
+      "grad_norm": 0.29296875,
+      "learning_rate": 0.0002999990944460126,
+      "loss": 0.3791,
+      "step": 8890
+    },
+    {
+      "epoch": 0.0012476249566485374,
+      "grad_norm": 0.34765625,
+      "learning_rate": 0.0002999990921488215,
+      "loss": 0.4881,
+      "step": 8900
+    },
+    {
+      "epoch": 0.0012490267824425245,
+      "grad_norm": 0.8828125,
+      "learning_rate": 0.0002999990898487205,
+      "loss": 0.4312,
+      "step": 8910
+    },
+    {
+      "epoch": 0.0012504286082365115,
+      "grad_norm": 0.318359375,
+      "learning_rate": 0.00029999908754570936,
+      "loss": 0.4883,
+      "step": 8920
+    },
+    {
+      "epoch": 0.0012518304340304986,
+      "grad_norm": 0.42578125,
+      "learning_rate": 0.0002999990852397882,
+      "loss": 0.3525,
+      "step": 8930
+    },
+    {
+      "epoch": 0.0012532322598244859,
+      "grad_norm": 0.53125,
+      "learning_rate": 0.000299999082930957,
+      "loss": 0.5207,
+      "step": 8940
+    },
+    {
+      "epoch": 0.001254634085618473,
+      "grad_norm": 0.76171875,
+      "learning_rate": 0.0002999990806192158,
+      "loss": 0.3815,
+      "step": 8950
+    },
+    {
+      "epoch": 0.00125603591141246,
+      "grad_norm": 0.74609375,
+      "learning_rate": 0.00029999907830456456,
+      "loss": 0.4554,
+      "step": 8960
+    },
+    {
+      "epoch": 0.001257437737206447,
+      "grad_norm": 0.384765625,
+      "learning_rate": 0.00029999907598700327,
+      "loss": 0.4522,
+      "step": 8970
+    },
+    {
+      "epoch": 0.0012588395630004343,
+      "grad_norm": 0.609375,
+      "learning_rate": 0.0002999990736665319,
+      "loss": 0.4533,
+      "step": 8980
+    },
+    {
+      "epoch": 0.0012602413887944214,
+      "grad_norm": 0.462890625,
+      "learning_rate": 0.00029999907134315053,
+      "loss": 0.5037,
+      "step": 8990
+    },
+    {
+      "epoch": 0.0012616432145884085,
+      "grad_norm": 0.65625,
+      "learning_rate": 0.00029999906901685913,
+      "loss": 0.4848,
+      "step": 9000
+    },
+    {
+      "epoch": 0.0012616432145884085,
+      "eval_loss": 0.17965355515480042,
+      "eval_runtime": 113.876,
+      "eval_samples_per_second": 200.595,
+      "eval_steps_per_second": 1.572,
+      "step": 9000
+    },
+    {
+      "epoch": 0.0012630450403823957,
+      "grad_norm": 0.47265625,
+      "learning_rate": 0.00029999906668765774,
+      "loss": 0.4496,
+      "step": 9010
+    },
+    {
+      "epoch": 0.0012644468661763828,
+      "grad_norm": 0.63671875,
+      "learning_rate": 0.00029999906435554624,
+      "loss": 0.4478,
+      "step": 9020
+    },
+    {
+      "epoch": 0.0012658486919703699,
+      "grad_norm": 0.6796875,
+      "learning_rate": 0.00029999906202052474,
+      "loss": 0.4299,
+      "step": 9030
+    },
+    {
+      "epoch": 0.001267250517764357,
+      "grad_norm": 0.79296875,
+      "learning_rate": 0.00029999905968259324,
+      "loss": 0.5,
+      "step": 9040
+    },
+    {
+      "epoch": 0.0012686523435583442,
+      "grad_norm": 0.6171875,
+      "learning_rate": 0.00029999905734175164,
+      "loss": 0.4341,
+      "step": 9050
+    },
+    {
+      "epoch": 0.0012700541693523313,
+      "grad_norm": 0.87890625,
+      "learning_rate": 0.00029999905499800004,
+      "loss": 0.5121,
+      "step": 9060
+    },
+    {
+      "epoch": 0.0012714559951463183,
+      "grad_norm": 0.75,
+      "learning_rate": 0.00029999905265133843,
+      "loss": 0.423,
+      "step": 9070
+    },
+    {
+      "epoch": 0.0012728578209403054,
+      "grad_norm": 0.890625,
+      "learning_rate": 0.0002999990503017668,
+      "loss": 0.5157,
+      "step": 9080
+    },
+    {
+      "epoch": 0.0012742596467342927,
+      "grad_norm": 0.68359375,
+      "learning_rate": 0.000299999047949285,
+      "loss": 0.4593,
+      "step": 9090
+    },
+    {
+      "epoch": 0.0012756614725282797,
+      "grad_norm": 0.8046875,
+      "learning_rate": 0.0002999990455938933,
+      "loss": 0.4342,
+      "step": 9100
+    },
+    {
+      "epoch": 0.0012770632983222668,
+      "grad_norm": 0.8828125,
+      "learning_rate": 0.00029999904323559155,
+      "loss": 0.4865,
+      "step": 9110
+    },
+    {
+      "epoch": 0.001278465124116254,
+      "grad_norm": 0.255859375,
+      "learning_rate": 0.00029999904087437974,
+      "loss": 0.448,
+      "step": 9120
+    },
+    {
+      "epoch": 0.0012798669499102411,
+      "grad_norm": 1.03125,
+      "learning_rate": 0.0002999990385102579,
+      "loss": 0.4956,
+      "step": 9130
+    },
+    {
+      "epoch": 0.0012812687757042282,
+      "grad_norm": 0.6328125,
+      "learning_rate": 0.00029999903614322596,
+      "loss": 0.4379,
+      "step": 9140
+    },
+    {
+      "epoch": 0.0012826706014982152,
+      "grad_norm": 0.392578125,
+      "learning_rate": 0.0002999990337732841,
+      "loss": 0.5302,
+      "step": 9150
+    },
+    {
+      "epoch": 0.0012840724272922025,
+      "grad_norm": 0.7578125,
+      "learning_rate": 0.0002999990314004322,
+      "loss": 0.5031,
+      "step": 9160
+    },
+    {
+      "epoch": 0.0012854742530861896,
+      "grad_norm": 0.58203125,
+      "learning_rate": 0.00029999902902467016,
+      "loss": 0.3132,
+      "step": 9170
+    },
+    {
+      "epoch": 0.0012868760788801766,
+      "grad_norm": 0.72265625,
+      "learning_rate": 0.00029999902664599814,
+      "loss": 0.3933,
+      "step": 9180
+    },
+    {
+      "epoch": 0.001288277904674164,
+      "grad_norm": 0.296875,
+      "learning_rate": 0.0002999990242644161,
+      "loss": 0.4043,
+      "step": 9190
+    },
+    {
+      "epoch": 0.001289679730468151,
+      "grad_norm": 0.5859375,
+      "learning_rate": 0.00029999902187992405,
+      "loss": 0.5478,
+      "step": 9200
+    },
+    {
+      "epoch": 0.001291081556262138,
+      "grad_norm": 0.26171875,
+      "learning_rate": 0.0002999990194925219,
+      "loss": 0.495,
+      "step": 9210
+    },
+    {
+      "epoch": 0.001292483382056125,
+      "grad_norm": 0.345703125,
+      "learning_rate": 0.00029999901710220974,
+      "loss": 0.5168,
+      "step": 9220
+    },
+    {
+      "epoch": 0.0012938852078501124,
+      "grad_norm": 0.310546875,
+      "learning_rate": 0.00029999901470898757,
+      "loss": 0.3826,
+      "step": 9230
+    },
+    {
+      "epoch": 0.0012952870336440994,
+      "grad_norm": 0.470703125,
+      "learning_rate": 0.0002999990123128554,
+      "loss": 0.4363,
+      "step": 9240
+    },
+    {
+      "epoch": 0.0012966888594380865,
+      "grad_norm": 0.6171875,
+      "learning_rate": 0.0002999990099138131,
+      "loss": 0.4175,
+      "step": 9250
+    },
+    {
+      "epoch": 0.0012980906852320736,
+      "grad_norm": 0.34765625,
+      "learning_rate": 0.00029999900751186083,
+      "loss": 0.3774,
+      "step": 9260
+    },
+    {
+      "epoch": 0.0012994925110260608,
+      "grad_norm": 0.9140625,
+      "learning_rate": 0.00029999900510699855,
+      "loss": 0.4917,
+      "step": 9270
+    },
+    {
+      "epoch": 0.001300894336820048,
+      "grad_norm": 0.79296875,
+      "learning_rate": 0.00029999900269922616,
+      "loss": 0.5462,
+      "step": 9280
+    },
+    {
+      "epoch": 0.001302296162614035,
+      "grad_norm": 0.35546875,
+      "learning_rate": 0.0002999990002885438,
+      "loss": 0.344,
+      "step": 9290
+    },
+    {
+      "epoch": 0.0013036979884080222,
+      "grad_norm": 0.59765625,
+      "learning_rate": 0.00029999899787495134,
+      "loss": 0.499,
+      "step": 9300
+    },
+    {
+      "epoch": 0.0013050998142020093,
+      "grad_norm": 1.0703125,
+      "learning_rate": 0.0002999989954584489,
+      "loss": 0.4163,
+      "step": 9310
+    },
+    {
+      "epoch": 0.0013065016399959964,
+      "grad_norm": 0.296875,
+      "learning_rate": 0.0002999989930390364,
+      "loss": 0.4647,
+      "step": 9320
+    },
+    {
+      "epoch": 0.0013079034657899834,
+      "grad_norm": 0.44921875,
+      "learning_rate": 0.00029999899061671387,
+      "loss": 0.4569,
+      "step": 9330
+    },
+    {
+      "epoch": 0.0013093052915839707,
+      "grad_norm": 0.150390625,
+      "learning_rate": 0.00029999898819148133,
+      "loss": 0.4494,
+      "step": 9340
+    },
+    {
+      "epoch": 0.0013107071173779578,
+      "grad_norm": 0.66796875,
+      "learning_rate": 0.0002999989857633388,
+      "loss": 0.4393,
+      "step": 9350
+    },
+    {
+      "epoch": 0.0013121089431719448,
+      "grad_norm": 0.703125,
+      "learning_rate": 0.00029999898333228615,
+      "loss": 0.5231,
+      "step": 9360
+    },
+    {
+      "epoch": 0.0013135107689659319,
+      "grad_norm": 0.60546875,
+      "learning_rate": 0.0002999989808983235,
+      "loss": 0.3771,
+      "step": 9370
+    },
+    {
+      "epoch": 0.0013149125947599192,
+      "grad_norm": 1.1171875,
+      "learning_rate": 0.0002999989784614508,
+      "loss": 0.5389,
+      "step": 9380
+    },
+    {
+      "epoch": 0.0013163144205539062,
+      "grad_norm": 0.6484375,
+      "learning_rate": 0.0002999989760216681,
+      "loss": 0.4195,
+      "step": 9390
+    },
+    {
+      "epoch": 0.0013177162463478933,
+      "grad_norm": 0.283203125,
+      "learning_rate": 0.00029999897357897536,
+      "loss": 0.4314,
+      "step": 9400
+    },
+    {
+      "epoch": 0.0013191180721418806,
+      "grad_norm": 1.15625,
+      "learning_rate": 0.00029999897113337255,
+      "loss": 0.4837,
+      "step": 9410
+    },
+    {
+      "epoch": 0.0013205198979358676,
+      "grad_norm": 0.4609375,
+      "learning_rate": 0.00029999896868485975,
+      "loss": 0.4618,
+      "step": 9420
+    },
+    {
+      "epoch": 0.0013219217237298547,
+      "grad_norm": 1.171875,
+      "learning_rate": 0.0002999989662334369,
+      "loss": 0.4941,
+      "step": 9430
+    },
+    {
+      "epoch": 0.0013233235495238417,
+      "grad_norm": 0.75390625,
+      "learning_rate": 0.000299998963779104,
+      "loss": 0.4681,
+      "step": 9440
+    },
+    {
+      "epoch": 0.001324725375317829,
+      "grad_norm": 0.71875,
+      "learning_rate": 0.0002999989613218611,
+      "loss": 0.4324,
+      "step": 9450
+    },
+    {
+      "epoch": 0.001326127201111816,
+      "grad_norm": 0.85546875,
+      "learning_rate": 0.0002999989588617082,
+      "loss": 0.4494,
+      "step": 9460
+    },
+    {
+      "epoch": 0.0013275290269058031,
+      "grad_norm": 0.74609375,
+      "learning_rate": 0.00029999895639864517,
+      "loss": 0.4027,
+      "step": 9470
+    },
+    {
+      "epoch": 0.0013289308526997904,
+      "grad_norm": 0.6796875,
+      "learning_rate": 0.00029999895393267216,
+      "loss": 0.4776,
+      "step": 9480
+    },
+    {
+      "epoch": 0.0013303326784937775,
+      "grad_norm": 0.51171875,
+      "learning_rate": 0.0002999989514637891,
+      "loss": 0.4024,
+      "step": 9490
+    },
+    {
+      "epoch": 0.0013317345042877645,
+      "grad_norm": 0.91015625,
+      "learning_rate": 0.0002999989489919961,
+      "loss": 0.405,
+      "step": 9500
+    },
+    {
+      "epoch": 0.0013317345042877645,
+      "eval_loss": 0.1779126524925232,
+      "eval_runtime": 112.2872,
+      "eval_samples_per_second": 203.434,
+      "eval_steps_per_second": 1.594,
+      "step": 9500
+    },
+    {
+      "epoch": 0.0013331363300817516,
+      "grad_norm": 0.380859375,
+      "learning_rate": 0.000299998946517293,
+      "loss": 0.4613,
+      "step": 9510
+    },
+    {
+      "epoch": 0.0013345381558757389,
+      "grad_norm": 0.7890625,
+      "learning_rate": 0.0002999989440396798,
+      "loss": 0.4131,
+      "step": 9520
+    },
+    {
+      "epoch": 0.001335939981669726,
+      "grad_norm": 0.609375,
+      "learning_rate": 0.00029999894155915665,
+      "loss": 0.4716,
+      "step": 9530
+    },
+    {
+      "epoch": 0.001337341807463713,
+      "grad_norm": 0.234375,
+      "learning_rate": 0.00029999893907572343,
+      "loss": 0.3789,
+      "step": 9540
+    },
+    {
+      "epoch": 0.0013387436332577,
+      "grad_norm": 0.5546875,
+      "learning_rate": 0.0002999989365893802,
+      "loss": 0.4503,
+      "step": 9550
+    },
+    {
+      "epoch": 0.0013401454590516873,
+      "grad_norm": 0.44921875,
+      "learning_rate": 0.00029999893410012694,
+      "loss": 0.3946,
+      "step": 9560
+    },
+    {
+      "epoch": 0.0013415472848456744,
+      "grad_norm": 0.3359375,
+      "learning_rate": 0.0002999989316079636,
+      "loss": 0.3684,
+      "step": 9570
+    },
+    {
+      "epoch": 0.0013429491106396615,
+      "grad_norm": 0.64453125,
+      "learning_rate": 0.0002999989291128903,
+      "loss": 0.4323,
+      "step": 9580
+    },
+    {
+      "epoch": 0.0013443509364336487,
+      "grad_norm": 0.3359375,
+      "learning_rate": 0.0002999989266149069,
+      "loss": 0.403,
+      "step": 9590
+    },
+    {
+      "epoch": 0.0013457527622276358,
+      "grad_norm": 1.40625,
+      "learning_rate": 0.00029999892411401355,
+      "loss": 0.4677,
+      "step": 9600
+    },
+    {
+      "epoch": 0.0013471545880216229,
+      "grad_norm": 0.5390625,
+      "learning_rate": 0.00029999892161021013,
+      "loss": 0.4505,
+      "step": 9610
+    },
+    {
+      "epoch": 0.00134855641381561,
+      "grad_norm": 0.625,
+      "learning_rate": 0.00029999891910349665,
+      "loss": 0.3674,
+      "step": 9620
+    },
+    {
+      "epoch": 0.0013499582396095972,
+      "grad_norm": 0.6796875,
+      "learning_rate": 0.00029999891659387317,
+      "loss": 0.4835,
+      "step": 9630
+    },
+    {
+      "epoch": 0.0013513600654035843,
+      "grad_norm": 0.431640625,
+      "learning_rate": 0.00029999891408133964,
+      "loss": 0.4168,
+      "step": 9640
+    },
+    {
+      "epoch": 0.0013527618911975713,
+      "grad_norm": 0.45703125,
+      "learning_rate": 0.00029999891156589606,
+      "loss": 0.4533,
+      "step": 9650
+    },
+    {
+      "epoch": 0.0013541637169915584,
+      "grad_norm": 0.75390625,
+      "learning_rate": 0.0002999989090475425,
+      "loss": 0.4661,
+      "step": 9660
+    },
+    {
+      "epoch": 0.0013555655427855457,
+      "grad_norm": 0.73046875,
+      "learning_rate": 0.00029999890652627884,
+      "loss": 0.5244,
+      "step": 9670
+    },
+    {
+      "epoch": 0.0013569673685795327,
+      "grad_norm": 0.5234375,
+      "learning_rate": 0.0002999989040021052,
+      "loss": 0.4458,
+      "step": 9680
+    },
+    {
+      "epoch": 0.0013583691943735198,
+      "grad_norm": 0.65234375,
+      "learning_rate": 0.0002999989014750215,
+      "loss": 0.3922,
+      "step": 9690
+    },
+    {
+      "epoch": 0.001359771020167507,
+      "grad_norm": 0.404296875,
+      "learning_rate": 0.0002999988989450278,
+      "loss": 0.4443,
+      "step": 9700
+    },
+    {
+      "epoch": 0.0013611728459614941,
+      "grad_norm": 0.5703125,
+      "learning_rate": 0.00029999889641212405,
+      "loss": 0.3954,
+      "step": 9710
+    },
+    {
+      "epoch": 0.0013625746717554812,
+      "grad_norm": 0.32421875,
+      "learning_rate": 0.00029999889387631026,
+      "loss": 0.4527,
+      "step": 9720
+    },
+    {
+      "epoch": 0.0013639764975494682,
+      "grad_norm": 0.2294921875,
+      "learning_rate": 0.0002999988913375865,
+      "loss": 0.3806,
+      "step": 9730
+    },
+    {
+      "epoch": 0.0013653783233434555,
+      "grad_norm": 0.5390625,
+      "learning_rate": 0.0002999988887959527,
+      "loss": 0.46,
+      "step": 9740
+    },
+    {
+      "epoch": 0.0013667801491374426,
+      "grad_norm": 0.62890625,
+      "learning_rate": 0.00029999888625140874,
+      "loss": 0.4645,
+      "step": 9750
+    },
+    {
+      "epoch": 0.0013681819749314296,
+      "grad_norm": 0.5,
+      "learning_rate": 0.0002999988837039549,
+      "loss": 0.3977,
+      "step": 9760
+    },
+    {
+      "epoch": 0.001369583800725417,
+      "grad_norm": 0.70703125,
+      "learning_rate": 0.0002999988811535909,
+      "loss": 0.4475,
+      "step": 9770
+    },
+    {
+      "epoch": 0.001370985626519404,
+      "grad_norm": 0.73828125,
+      "learning_rate": 0.00029999887860031695,
+      "loss": 0.4696,
+      "step": 9780
+    },
+    {
+      "epoch": 0.001372387452313391,
+      "grad_norm": 0.796875,
+      "learning_rate": 0.000299998876044133,
+      "loss": 0.4126,
+      "step": 9790
+    },
+    {
+      "epoch": 0.001373789278107378,
+      "grad_norm": 0.64453125,
+      "learning_rate": 0.0002999988734850389,
+      "loss": 0.4177,
+      "step": 9800
+    },
+    {
+      "epoch": 0.0013751911039013654,
+      "grad_norm": 0.70703125,
+      "learning_rate": 0.00029999887092303486,
+      "loss": 0.4654,
+      "step": 9810
+    },
+    {
+      "epoch": 0.0013765929296953524,
+      "grad_norm": 0.69921875,
+      "learning_rate": 0.0002999988683581208,
+      "loss": 0.4431,
+      "step": 9820
+    },
+    {
+      "epoch": 0.0013779947554893395,
+      "grad_norm": 0.80859375,
+      "learning_rate": 0.00029999886579029666,
+      "loss": 0.4776,
+      "step": 9830
+    },
+    {
+      "epoch": 0.0013793965812833266,
+      "grad_norm": 0.6015625,
+      "learning_rate": 0.0002999988632195625,
+      "loss": 0.507,
+      "step": 9840
+    },
+    {
+      "epoch": 0.0013807984070773138,
+      "grad_norm": 0.359375,
+      "learning_rate": 0.00029999886064591835,
+      "loss": 0.496,
+      "step": 9850
+    },
+    {
+      "epoch": 0.001382200232871301,
+      "grad_norm": 0.5234375,
+      "learning_rate": 0.00029999885806936415,
+      "loss": 0.5075,
+      "step": 9860
+    },
+    {
+      "epoch": 0.001383602058665288,
+      "grad_norm": 0.328125,
+      "learning_rate": 0.0002999988554898999,
+      "loss": 0.3726,
+      "step": 9870
+    },
+    {
+      "epoch": 0.0013850038844592752,
+      "grad_norm": 0.77734375,
+      "learning_rate": 0.00029999885290752564,
+      "loss": 0.4425,
+      "step": 9880
+    },
+    {
+      "epoch": 0.0013864057102532623,
+      "grad_norm": 0.435546875,
+      "learning_rate": 0.0002999988503222413,
+      "loss": 0.4121,
+      "step": 9890
+    },
+    {
+      "epoch": 0.0013878075360472494,
+      "grad_norm": 0.66015625,
+      "learning_rate": 0.00029999884773404697,
+      "loss": 0.3684,
+      "step": 9900
+    },
+    {
+      "epoch": 0.0013892093618412364,
+      "grad_norm": 0.49609375,
+      "learning_rate": 0.0002999988451429426,
+      "loss": 0.3725,
+      "step": 9910
+    },
+    {
+      "epoch": 0.0013906111876352237,
+      "grad_norm": 0.419921875,
+      "learning_rate": 0.00029999884254892825,
+      "loss": 0.469,
+      "step": 9920
+    },
+    {
+      "epoch": 0.0013920130134292108,
+      "grad_norm": 0.8515625,
+      "learning_rate": 0.00029999883995200383,
+      "loss": 0.4056,
+      "step": 9930
+    },
+    {
+      "epoch": 0.0013934148392231978,
+      "grad_norm": 0.3828125,
+      "learning_rate": 0.00029999883735216937,
+      "loss": 0.415,
+      "step": 9940
+    },
+    {
+      "epoch": 0.0013948166650171849,
+      "grad_norm": 0.5078125,
+      "learning_rate": 0.00029999883474942485,
+      "loss": 0.4442,
+      "step": 9950
+    },
+    {
+      "epoch": 0.0013962184908111722,
+      "grad_norm": 0.1845703125,
+      "learning_rate": 0.00029999883214377034,
+      "loss": 0.4451,
+      "step": 9960
+    },
+    {
+      "epoch": 0.0013976203166051592,
+      "grad_norm": 0.25390625,
+      "learning_rate": 0.0002999988295352058,
+      "loss": 0.4303,
+      "step": 9970
+    },
+    {
+      "epoch": 0.0013990221423991463,
+      "grad_norm": 0.99609375,
+      "learning_rate": 0.00029999882692373125,
+      "loss": 0.4909,
+      "step": 9980
+    },
+    {
+      "epoch": 0.0014004239681931336,
+      "grad_norm": 0.7734375,
+      "learning_rate": 0.00029999882430934663,
+      "loss": 0.3695,
+      "step": 9990
+    },
+    {
+      "epoch": 0.0014018257939871206,
+      "grad_norm": 0.9921875,
+      "learning_rate": 0.000299998821692052,
+      "loss": 0.489,
+      "step": 10000
+    },
+    {
+      "epoch": 0.0014018257939871206,
+      "eval_loss": 0.17730337381362915,
+      "eval_runtime": 112.2399,
+      "eval_samples_per_second": 203.519,
+      "eval_steps_per_second": 1.595,
+      "step": 10000
+    },
+    {
+      "epoch": 0.0014032276197811077,
+      "grad_norm": 0.88671875,
+      "learning_rate": 0.0002999988190718473,
+      "loss": 0.4894,
+      "step": 10010
+    },
+    {
+      "epoch": 0.0014046294455750947,
+      "grad_norm": 0.69921875,
+      "learning_rate": 0.0002999988164487327,
+      "loss": 0.427,
+      "step": 10020
+    },
+    {
+      "epoch": 0.001406031271369082,
+      "grad_norm": 0.609375,
+      "learning_rate": 0.0002999988138227079,
+      "loss": 0.404,
+      "step": 10030
+    },
+    {
+      "epoch": 0.001407433097163069,
+      "grad_norm": 0.6640625,
+      "learning_rate": 0.0002999988111937732,
+      "loss": 0.3942,
+      "step": 10040
+    },
+    {
+      "epoch": 0.0014088349229570561,
+      "grad_norm": 0.79296875,
+      "learning_rate": 0.00029999880856192835,
+      "loss": 0.3519,
+      "step": 10050
+    },
+    {
+      "epoch": 0.0014102367487510434,
+      "grad_norm": 0.259765625,
+      "learning_rate": 0.00029999880592717357,
+      "loss": 0.4191,
+      "step": 10060
+    },
+    {
+      "epoch": 0.0014116385745450305,
+      "grad_norm": 0.7578125,
+      "learning_rate": 0.0002999988032895087,
+      "loss": 0.4562,
+      "step": 10070
+    },
+    {
+      "epoch": 0.0014130404003390175,
+      "grad_norm": 0.9765625,
+      "learning_rate": 0.00029999880064893387,
+      "loss": 0.4004,
+      "step": 10080
+    },
+    {
+      "epoch": 0.0014144422261330046,
+      "grad_norm": 0.796875,
+      "learning_rate": 0.00029999879800544894,
+      "loss": 0.4163,
+      "step": 10090
+    },
+    {
+      "epoch": 0.0014158440519269919,
+      "grad_norm": 0.30859375,
+      "learning_rate": 0.000299998795359054,
+      "loss": 0.292,
+      "step": 10100
+    },
+    {
+      "epoch": 0.001417245877720979,
+      "grad_norm": 0.263671875,
+      "learning_rate": 0.000299998792709749,
+      "loss": 0.3318,
+      "step": 10110
+    },
+    {
+      "epoch": 0.001418647703514966,
+      "grad_norm": 0.69921875,
+      "learning_rate": 0.00029999879005753404,
+      "loss": 0.4539,
+      "step": 10120
+    },
+    {
+      "epoch": 0.001420049529308953,
+      "grad_norm": 0.361328125,
+      "learning_rate": 0.00029999878740240906,
+      "loss": 0.3881,
+      "step": 10130
+    },
+    {
+      "epoch": 0.0014214513551029403,
+      "grad_norm": 0.66796875,
+      "learning_rate": 0.00029999878474437397,
+      "loss": 0.4689,
+      "step": 10140
+    },
+    {
+      "epoch": 0.0014228531808969274,
+      "grad_norm": 0.5625,
+      "learning_rate": 0.00029999878208342894,
+      "loss": 0.4459,
+      "step": 10150
+    },
+    {
+      "epoch": 0.0014242550066909145,
+      "grad_norm": 0.73828125,
+      "learning_rate": 0.0002999987794195738,
+      "loss": 0.3846,
+      "step": 10160
+    },
+    {
+      "epoch": 0.0014256568324849017,
+      "grad_norm": 0.8203125,
+      "learning_rate": 0.0002999987767528086,
+      "loss": 0.4947,
+      "step": 10170
+    },
+    {
+      "epoch": 0.0014270586582788888,
+      "grad_norm": 0.79296875,
+      "learning_rate": 0.00029999877408313347,
+      "loss": 0.5135,
+      "step": 10180
+    },
+    {
+      "epoch": 0.0014284604840728759,
+      "grad_norm": 0.796875,
+      "learning_rate": 0.0002999987714105483,
+      "loss": 0.4745,
+      "step": 10190
+    },
+    {
+      "epoch": 0.001429862309866863,
+      "grad_norm": 0.88671875,
+      "learning_rate": 0.00029999876873505304,
+      "loss": 0.4396,
+      "step": 10200
+    },
+    {
+      "epoch": 0.0014312641356608502,
+      "grad_norm": 1.0,
+      "learning_rate": 0.0002999987660566478,
+      "loss": 0.4613,
+      "step": 10210
+    },
+    {
+      "epoch": 0.0014326659614548373,
+      "grad_norm": 1.015625,
+      "learning_rate": 0.0002999987633753325,
+      "loss": 0.4413,
+      "step": 10220
+    },
+    {
+      "epoch": 0.0014340677872488243,
+      "grad_norm": 0.609375,
+      "learning_rate": 0.0002999987606911072,
+      "loss": 0.3807,
+      "step": 10230
+    },
+    {
+      "epoch": 0.0014354696130428114,
+      "grad_norm": 0.275390625,
+      "learning_rate": 0.00029999875800397187,
+      "loss": 0.3948,
+      "step": 10240
+    },
+    {
+      "epoch": 0.0014368714388367987,
+      "grad_norm": 0.2197265625,
+      "learning_rate": 0.00029999875531392647,
+      "loss": 0.354,
+      "step": 10250
+    },
+    {
+      "epoch": 0.0014382732646307857,
+      "grad_norm": 0.6640625,
+      "learning_rate": 0.00029999875262097113,
+      "loss": 0.4706,
+      "step": 10260
+    },
+    {
+      "epoch": 0.0014396750904247728,
+      "grad_norm": 0.39453125,
+      "learning_rate": 0.0002999987499251057,
+      "loss": 0.3163,
+      "step": 10270
+    },
+    {
+      "epoch": 0.00144107691621876,
+      "grad_norm": 0.65234375,
+      "learning_rate": 0.00029999874722633023,
+      "loss": 0.3933,
+      "step": 10280
+    },
+    {
+      "epoch": 0.0014424787420127471,
+      "grad_norm": 0.431640625,
+      "learning_rate": 0.0002999987445246448,
+      "loss": 0.3804,
+      "step": 10290
+    },
+    {
+      "epoch": 0.0014438805678067342,
+      "grad_norm": 0.70703125,
+      "learning_rate": 0.00029999874182004923,
+      "loss": 0.4831,
+      "step": 10300
+    },
+    {
+      "epoch": 0.0014452823936007212,
+      "grad_norm": 0.62109375,
+      "learning_rate": 0.0002999987391125437,
+      "loss": 0.3895,
+      "step": 10310
+    },
+    {
+      "epoch": 0.0014466842193947085,
+      "grad_norm": 0.298828125,
+      "learning_rate": 0.00029999873640212813,
+      "loss": 0.4482,
+      "step": 10320
+    },
+    {
+      "epoch": 0.0014480860451886956,
+      "grad_norm": 0.26953125,
+      "learning_rate": 0.0002999987336888025,
+      "loss": 0.4603,
+      "step": 10330
+    },
+    {
+      "epoch": 0.0014494878709826826,
+      "grad_norm": 0.44140625,
+      "learning_rate": 0.0002999987309725669,
+      "loss": 0.4359,
+      "step": 10340
+    },
+    {
+      "epoch": 0.00145088969677667,
+      "grad_norm": 0.609375,
+      "learning_rate": 0.00029999872825342127,
+      "loss": 0.4599,
+      "step": 10350
+    },
+    {
+      "epoch": 0.001452291522570657,
+      "grad_norm": 0.5625,
+      "learning_rate": 0.00029999872553136556,
+      "loss": 0.3849,
+      "step": 10360
+    },
+    {
+      "epoch": 0.001453693348364644,
+      "grad_norm": 0.2578125,
+      "learning_rate": 0.00029999872280639985,
+      "loss": 0.3681,
+      "step": 10370
+    },
+    {
+      "epoch": 0.001455095174158631,
+      "grad_norm": 0.431640625,
+      "learning_rate": 0.0002999987200785241,
+      "loss": 0.3418,
+      "step": 10380
+    },
+    {
+      "epoch": 0.0014564969999526184,
+      "grad_norm": 0.5078125,
+      "learning_rate": 0.0002999987173477384,
+      "loss": 0.4339,
+      "step": 10390
+    },
+    {
+      "epoch": 0.0014578988257466054,
+      "grad_norm": 0.435546875,
+      "learning_rate": 0.0002999987146140426,
+      "loss": 0.3566,
+      "step": 10400
+    },
+    {
+      "epoch": 0.0014593006515405925,
+      "grad_norm": 1.234375,
+      "learning_rate": 0.00029999871187743677,
+      "loss": 0.4247,
+      "step": 10410
+    },
+    {
+      "epoch": 0.0014607024773345796,
+      "grad_norm": 0.96875,
+      "learning_rate": 0.00029999870913792085,
+      "loss": 0.5284,
+      "step": 10420
+    },
+    {
+      "epoch": 0.0014621043031285668,
+      "grad_norm": 0.59765625,
+      "learning_rate": 0.00029999870639549505,
+      "loss": 0.4373,
+      "step": 10430
+    },
+    {
+      "epoch": 0.0014635061289225539,
+      "grad_norm": 0.48828125,
+      "learning_rate": 0.0002999987036501591,
+      "loss": 0.3596,
+      "step": 10440
+    },
+    {
+      "epoch": 0.001464907954716541,
+      "grad_norm": 0.236328125,
+      "learning_rate": 0.0002999987009019132,
+      "loss": 0.4211,
+      "step": 10450
+    },
+    {
+      "epoch": 0.0014663097805105282,
+      "grad_norm": 0.546875,
+      "learning_rate": 0.00029999869815075725,
+      "loss": 0.487,
+      "step": 10460
+    },
+    {
+      "epoch": 0.0014677116063045153,
+      "grad_norm": 0.79296875,
+      "learning_rate": 0.00029999869539669124,
+      "loss": 0.4045,
+      "step": 10470
+    },
+    {
+      "epoch": 0.0014691134320985024,
+      "grad_norm": 0.380859375,
+      "learning_rate": 0.0002999986926397152,
+      "loss": 0.4048,
+      "step": 10480
+    },
+    {
+      "epoch": 0.0014705152578924894,
+      "grad_norm": 0.6328125,
+      "learning_rate": 0.00029999868987982915,
+      "loss": 0.4103,
+      "step": 10490
+    },
+    {
+      "epoch": 0.0014719170836864767,
+      "grad_norm": 0.5703125,
+      "learning_rate": 0.0002999986871170331,
+      "loss": 0.3853,
+      "step": 10500
+    },
+    {
+      "epoch": 0.0014719170836864767,
+      "eval_loss": 0.1777743548154831,
+      "eval_runtime": 113.9971,
+      "eval_samples_per_second": 200.382,
+      "eval_steps_per_second": 1.57,
+      "step": 10500
+    },
+    {
+      "epoch": 0.0014733189094804637,
+      "grad_norm": 0.69921875,
+      "learning_rate": 0.000299998684351327,
+      "loss": 0.4171,
+      "step": 10510
+    },
+    {
+      "epoch": 0.0014747207352744508,
+      "grad_norm": 0.361328125,
+      "learning_rate": 0.0002999986815827109,
+      "loss": 0.4019,
+      "step": 10520
+    },
+    {
+      "epoch": 0.0014761225610684379,
+      "grad_norm": 0.59765625,
+      "learning_rate": 0.0002999986788111848,
+      "loss": 0.5203,
+      "step": 10530
+    },
+    {
+      "epoch": 0.0014775243868624251,
+      "grad_norm": 0.640625,
+      "learning_rate": 0.00029999867603674855,
+      "loss": 0.4741,
+      "step": 10540
+    },
+    {
+      "epoch": 0.0014789262126564122,
+      "grad_norm": 0.5703125,
+      "learning_rate": 0.0002999986732594024,
+      "loss": 0.4593,
+      "step": 10550
+    },
+    {
+      "epoch": 0.0014803280384503993,
+      "grad_norm": 0.66796875,
+      "learning_rate": 0.0002999986704791461,
+      "loss": 0.45,
+      "step": 10560
+    },
+    {
+      "epoch": 0.0014817298642443865,
+      "grad_norm": 0.54296875,
+      "learning_rate": 0.0002999986676959799,
+      "loss": 0.4173,
+      "step": 10570
+    },
+    {
+      "epoch": 0.0014831316900383736,
+      "grad_norm": 0.8515625,
+      "learning_rate": 0.00029999866490990355,
+      "loss": 0.4825,
+      "step": 10580
+    },
+    {
+      "epoch": 0.0014845335158323607,
+      "grad_norm": 0.71875,
+      "learning_rate": 0.0002999986621209173,
+      "loss": 0.4325,
+      "step": 10590
+    },
+    {
+      "epoch": 0.0014859353416263477,
+      "grad_norm": 0.65234375,
+      "learning_rate": 0.00029999865932902096,
+      "loss": 0.3705,
+      "step": 10600
+    },
+    {
+      "epoch": 0.001487337167420335,
+      "grad_norm": 0.703125,
+      "learning_rate": 0.0002999986565342146,
+      "loss": 0.4873,
+      "step": 10610
+    },
+    {
+      "epoch": 0.001488738993214322,
+      "grad_norm": 0.765625,
+      "learning_rate": 0.00029999865373649815,
+      "loss": 0.4666,
+      "step": 10620
+    },
+    {
+      "epoch": 0.0014901408190083091,
+      "grad_norm": 0.2177734375,
+      "learning_rate": 0.00029999865093587177,
+      "loss": 0.3025,
+      "step": 10630
+    },
+    {
+      "epoch": 0.0014915426448022964,
+      "grad_norm": 0.49609375,
+      "learning_rate": 0.0002999986481323353,
+      "loss": 0.4764,
+      "step": 10640
+    },
+    {
+      "epoch": 0.0014929444705962835,
+      "grad_norm": 1.0234375,
+      "learning_rate": 0.0002999986453258888,
+      "loss": 0.4676,
+      "step": 10650
+    },
+    {
+      "epoch": 0.0014943462963902705,
+      "grad_norm": 0.26171875,
+      "learning_rate": 0.0002999986425165323,
+      "loss": 0.4282,
+      "step": 10660
+    },
+    {
+      "epoch": 0.0014957481221842576,
+      "grad_norm": 0.3046875,
+      "learning_rate": 0.0002999986397042658,
+      "loss": 0.3724,
+      "step": 10670
+    },
+    {
+      "epoch": 0.0014971499479782449,
+      "grad_norm": 0.423828125,
+      "learning_rate": 0.0002999986368890892,
+      "loss": 0.3169,
+      "step": 10680
+    },
+    {
+      "epoch": 0.001498551773772232,
+      "grad_norm": 0.33984375,
+      "learning_rate": 0.0002999986340710027,
+      "loss": 0.3254,
+      "step": 10690
+    },
+    {
+      "epoch": 0.001499953599566219,
+      "grad_norm": 0.2197265625,
+      "learning_rate": 0.00029999863125000604,
+      "loss": 0.4022,
+      "step": 10700
+    },
+    {
+      "epoch": 0.001501355425360206,
+      "grad_norm": 0.81640625,
+      "learning_rate": 0.00029999862842609946,
+      "loss": 0.5181,
+      "step": 10710
+    },
+    {
+      "epoch": 0.0015027572511541933,
+      "grad_norm": 0.7890625,
+      "learning_rate": 0.00029999862559928277,
+      "loss": 0.5158,
+      "step": 10720
+    },
+    {
+      "epoch": 0.0015041590769481804,
+      "grad_norm": 0.81640625,
+      "learning_rate": 0.0002999986227695561,
+      "loss": 0.3475,
+      "step": 10730
+    },
+    {
+      "epoch": 0.0015055609027421674,
+      "grad_norm": 0.71875,
+      "learning_rate": 0.0002999986199369194,
+      "loss": 0.4349,
+      "step": 10740
+    },
+    {
+      "epoch": 0.0015069627285361547,
+      "grad_norm": 0.63671875,
+      "learning_rate": 0.0002999986171013726,
+      "loss": 0.4544,
+      "step": 10750
+    },
+    {
+      "epoch": 0.0015083645543301418,
+      "grad_norm": 1.2109375,
+      "learning_rate": 0.0002999986142629159,
+      "loss": 0.5036,
+      "step": 10760
+    },
+    {
+      "epoch": 0.0015097663801241288,
+      "grad_norm": 0.5,
+      "learning_rate": 0.0002999986114215491,
+      "loss": 0.4059,
+      "step": 10770
+    },
+    {
+      "epoch": 0.001511168205918116,
+      "grad_norm": 0.341796875,
+      "learning_rate": 0.0002999986085772723,
+      "loss": 0.3344,
+      "step": 10780
+    },
+    {
+      "epoch": 0.0015125700317121032,
+      "grad_norm": 0.75,
+      "learning_rate": 0.00029999860573008544,
+      "loss": 0.486,
+      "step": 10790
+    },
+    {
+      "epoch": 0.0015139718575060902,
+      "grad_norm": 0.25,
+      "learning_rate": 0.00029999860287998855,
+      "loss": 0.347,
+      "step": 10800
+    },
+    {
+      "epoch": 0.0015153736833000773,
+      "grad_norm": 0.4453125,
+      "learning_rate": 0.0002999986000269817,
+      "loss": 0.3681,
+      "step": 10810
+    },
+    {
+      "epoch": 0.0015167755090940644,
+      "grad_norm": 0.8671875,
+      "learning_rate": 0.00029999859717106476,
+      "loss": 0.4981,
+      "step": 10820
+    },
+    {
+      "epoch": 0.0015181773348880516,
+      "grad_norm": 0.69140625,
+      "learning_rate": 0.0002999985943122379,
+      "loss": 0.4436,
+      "step": 10830
+    },
+    {
+      "epoch": 0.0015195791606820387,
+      "grad_norm": 1.0859375,
+      "learning_rate": 0.0002999985914505009,
+      "loss": 0.4291,
+      "step": 10840
+    },
+    {
+      "epoch": 0.0015209809864760258,
+      "grad_norm": 0.71875,
+      "learning_rate": 0.0002999985885858539,
+      "loss": 0.4227,
+      "step": 10850
+    },
+    {
+      "epoch": 0.001522382812270013,
+      "grad_norm": 0.7734375,
+      "learning_rate": 0.00029999858571829683,
+      "loss": 0.4109,
+      "step": 10860
+    },
+    {
+      "epoch": 0.001523784638064,
+      "grad_norm": 0.5859375,
+      "learning_rate": 0.00029999858284782984,
+      "loss": 0.487,
+      "step": 10870
+    },
+    {
+      "epoch": 0.0015251864638579872,
+      "grad_norm": 0.6875,
+      "learning_rate": 0.00029999857997445274,
+      "loss": 0.4839,
+      "step": 10880
+    },
+    {
+      "epoch": 0.0015265882896519742,
+      "grad_norm": 0.7109375,
+      "learning_rate": 0.00029999857709816564,
+      "loss": 0.4406,
+      "step": 10890
+    },
+    {
+      "epoch": 0.0015279901154459615,
+      "grad_norm": 0.52734375,
+      "learning_rate": 0.00029999857421896854,
+      "loss": 0.4348,
+      "step": 10900
+    },
+    {
+      "epoch": 0.0015293919412399486,
+      "grad_norm": 0.6328125,
+      "learning_rate": 0.0002999985713368614,
+      "loss": 0.4313,
+      "step": 10910
+    },
+    {
+      "epoch": 0.0015307937670339356,
+      "grad_norm": 0.92578125,
+      "learning_rate": 0.00029999856845184424,
+      "loss": 0.4236,
+      "step": 10920
+    },
+    {
+      "epoch": 0.001532195592827923,
+      "grad_norm": 0.88671875,
+      "learning_rate": 0.00029999856556391704,
+      "loss": 0.4851,
+      "step": 10930
+    },
+    {
+      "epoch": 0.00153359741862191,
+      "grad_norm": 0.95703125,
+      "learning_rate": 0.00029999856267307984,
+      "loss": 0.3592,
+      "step": 10940
+    },
+    {
+      "epoch": 0.001534999244415897,
+      "grad_norm": 0.462890625,
+      "learning_rate": 0.00029999855977933253,
+      "loss": 0.4506,
+      "step": 10950
+    },
+    {
+      "epoch": 0.001536401070209884,
+      "grad_norm": 0.4375,
+      "learning_rate": 0.0002999985568826753,
+      "loss": 0.467,
+      "step": 10960
+    },
+    {
+      "epoch": 0.0015378028960038714,
+      "grad_norm": 0.9765625,
+      "learning_rate": 0.00029999855398310803,
+      "loss": 0.4379,
+      "step": 10970
+    },
+    {
+      "epoch": 0.0015392047217978584,
+      "grad_norm": 0.326171875,
+      "learning_rate": 0.00029999855108063067,
+      "loss": 0.3838,
+      "step": 10980
+    },
+    {
+      "epoch": 0.0015406065475918455,
+      "grad_norm": 0.8203125,
+      "learning_rate": 0.0002999985481752433,
+      "loss": 0.4903,
+      "step": 10990
+    },
+    {
+      "epoch": 0.0015420083733858325,
+      "grad_norm": 0.8515625,
+      "learning_rate": 0.00029999854526694596,
+      "loss": 0.4298,
+      "step": 11000
+    },
+    {
+      "epoch": 0.0015420083733858325,
+      "eval_loss": 0.1755325049161911,
+      "eval_runtime": 112.2725,
+      "eval_samples_per_second": 203.46,
+      "eval_steps_per_second": 1.594,
+      "step": 11000
+    },
+    {
+      "epoch": 0.0015434101991798198,
+      "grad_norm": 0.69921875,
+      "learning_rate": 0.00029999854235573855,
+      "loss": 0.3867,
+      "step": 11010
+    },
+    {
+      "epoch": 0.0015448120249738069,
+      "grad_norm": 0.54296875,
+      "learning_rate": 0.00029999853944162114,
+      "loss": 0.4283,
+      "step": 11020
+    },
+    {
+      "epoch": 0.001546213850767794,
+      "grad_norm": 0.89453125,
+      "learning_rate": 0.00029999853652459374,
+      "loss": 0.4642,
+      "step": 11030
+    },
+    {
+      "epoch": 0.0015476156765617812,
+      "grad_norm": 0.83984375,
+      "learning_rate": 0.0002999985336046562,
+      "loss": 0.3961,
+      "step": 11040
+    },
+    {
+      "epoch": 0.0015490175023557683,
+      "grad_norm": 0.73046875,
+      "learning_rate": 0.00029999853068180877,
+      "loss": 0.4539,
+      "step": 11050
+    },
+    {
+      "epoch": 0.0015504193281497553,
+      "grad_norm": 0.984375,
+      "learning_rate": 0.0002999985277560512,
+      "loss": 0.45,
+      "step": 11060
+    },
+    {
+      "epoch": 0.0015518211539437424,
+      "grad_norm": 0.61328125,
+      "learning_rate": 0.0002999985248273837,
+      "loss": 0.432,
+      "step": 11070
+    },
+    {
+      "epoch": 0.0015532229797377297,
+      "grad_norm": 1.0546875,
+      "learning_rate": 0.00029999852189580614,
+      "loss": 0.499,
+      "step": 11080
+    },
+    {
+      "epoch": 0.0015546248055317167,
+      "grad_norm": 0.4296875,
+      "learning_rate": 0.0002999985189613185,
+      "loss": 0.3849,
+      "step": 11090
+    },
+    {
+      "epoch": 0.0015560266313257038,
+      "grad_norm": 0.5390625,
+      "learning_rate": 0.00029999851602392096,
+      "loss": 0.3608,
+      "step": 11100
+    },
+    {
+      "epoch": 0.0015574284571196909,
+      "grad_norm": 0.3828125,
+      "learning_rate": 0.0002999985130836133,
+      "loss": 0.4467,
+      "step": 11110
+    },
+    {
+      "epoch": 0.0015588302829136781,
+      "grad_norm": 0.87109375,
+      "learning_rate": 0.00029999851014039563,
+      "loss": 0.4526,
+      "step": 11120
+    },
+    {
+      "epoch": 0.0015602321087076652,
+      "grad_norm": 0.97265625,
+      "learning_rate": 0.00029999850719426797,
+      "loss": 0.4376,
+      "step": 11130
+    },
+    {
+      "epoch": 0.0015616339345016523,
+      "grad_norm": 0.91796875,
+      "learning_rate": 0.00029999850424523026,
+      "loss": 0.3923,
+      "step": 11140
+    },
+    {
+      "epoch": 0.0015630357602956395,
+      "grad_norm": 0.67578125,
+      "learning_rate": 0.0002999985012932825,
+      "loss": 0.4162,
+      "step": 11150
+    },
+    {
+      "epoch": 0.0015644375860896266,
+      "grad_norm": 0.578125,
+      "learning_rate": 0.0002999984983384248,
+      "loss": 0.4105,
+      "step": 11160
+    },
+    {
+      "epoch": 0.0015658394118836137,
+      "grad_norm": 0.87109375,
+      "learning_rate": 0.000299998495380657,
+      "loss": 0.4134,
+      "step": 11170
+    },
+    {
+      "epoch": 0.0015672412376776007,
+      "grad_norm": 0.90625,
+      "learning_rate": 0.0002999984924199792,
+      "loss": 0.4347,
+      "step": 11180
+    },
+    {
+      "epoch": 0.001568643063471588,
+      "grad_norm": 0.8046875,
+      "learning_rate": 0.0002999984894563914,
+      "loss": 0.4853,
+      "step": 11190
+    },
+    {
+      "epoch": 0.001570044889265575,
+      "grad_norm": 0.84375,
+      "learning_rate": 0.0002999984864898935,
+      "loss": 0.4014,
+      "step": 11200
+    },
+    {
+      "epoch": 0.0015714467150595621,
+      "grad_norm": 0.95703125,
+      "learning_rate": 0.00029999848352048564,
+      "loss": 0.4483,
+      "step": 11210
+    },
+    {
+      "epoch": 0.0015728485408535494,
+      "grad_norm": 0.376953125,
+      "learning_rate": 0.00029999848054816777,
+      "loss": 0.4587,
+      "step": 11220
+    },
+    {
+      "epoch": 0.0015742503666475365,
+      "grad_norm": 0.5390625,
+      "learning_rate": 0.0002999984775729398,
+      "loss": 0.3971,
+      "step": 11230
+    },
+    {
+      "epoch": 0.0015756521924415235,
+      "grad_norm": 0.9375,
+      "learning_rate": 0.0002999984745948019,
+      "loss": 0.5006,
+      "step": 11240
+    },
+    {
+      "epoch": 0.0015770540182355106,
+      "grad_norm": 0.2470703125,
+      "learning_rate": 0.00029999847161375396,
+      "loss": 0.3465,
+      "step": 11250
+    },
+    {
+      "epoch": 0.0015784558440294979,
+      "grad_norm": 0.65234375,
+      "learning_rate": 0.00029999846862979594,
+      "loss": 0.4071,
+      "step": 11260
+    },
+    {
+      "epoch": 0.001579857669823485,
+      "grad_norm": 0.609375,
+      "learning_rate": 0.0002999984656429279,
+      "loss": 0.4812,
+      "step": 11270
+    },
+    {
+      "epoch": 0.001581259495617472,
+      "grad_norm": 0.55859375,
+      "learning_rate": 0.0002999984626531499,
+      "loss": 0.4278,
+      "step": 11280
+    },
+    {
+      "epoch": 0.001582661321411459,
+      "grad_norm": 0.96484375,
+      "learning_rate": 0.0002999984596604618,
+      "loss": 0.4164,
+      "step": 11290
+    },
+    {
+      "epoch": 0.0015840631472054463,
+      "grad_norm": 0.8046875,
+      "learning_rate": 0.00029999845666486374,
+      "loss": 0.4443,
+      "step": 11300
+    },
+    {
+      "epoch": 0.0015854649729994334,
+      "grad_norm": 0.59765625,
+      "learning_rate": 0.0002999984536663556,
+      "loss": 0.5397,
+      "step": 11310
+    },
+    {
+      "epoch": 0.0015868667987934204,
+      "grad_norm": 1.6015625,
+      "learning_rate": 0.00029999845066493755,
+      "loss": 0.4188,
+      "step": 11320
+    },
+    {
+      "epoch": 0.0015882686245874077,
+      "grad_norm": 0.27734375,
+      "learning_rate": 0.00029999844766060937,
+      "loss": 0.397,
+      "step": 11330
+    },
+    {
+      "epoch": 0.0015896704503813948,
+      "grad_norm": 0.984375,
+      "learning_rate": 0.0002999984446533712,
+      "loss": 0.3977,
+      "step": 11340
+    },
+    {
+      "epoch": 0.0015910722761753818,
+      "grad_norm": 0.83984375,
+      "learning_rate": 0.000299998441643223,
+      "loss": 0.4544,
+      "step": 11350
+    },
+    {
+      "epoch": 0.001592474101969369,
+      "grad_norm": 0.96484375,
+      "learning_rate": 0.0002999984386301648,
+      "loss": 0.4528,
+      "step": 11360
+    },
+    {
+      "epoch": 0.0015938759277633562,
+      "grad_norm": 0.4921875,
+      "learning_rate": 0.00029999843561419657,
+      "loss": 0.3638,
+      "step": 11370
+    },
+    {
+      "epoch": 0.0015952777535573432,
+      "grad_norm": 0.640625,
+      "learning_rate": 0.0002999984325953183,
+      "loss": 0.4077,
+      "step": 11380
+    },
+    {
+      "epoch": 0.0015966795793513303,
+      "grad_norm": 0.5546875,
+      "learning_rate": 0.00029999842957353,
+      "loss": 0.4546,
+      "step": 11390
+    },
+    {
+      "epoch": 0.0015980814051453174,
+      "grad_norm": 0.51953125,
+      "learning_rate": 0.0002999984265488317,
+      "loss": 0.4067,
+      "step": 11400
+    },
+    {
+      "epoch": 0.0015994832309393046,
+      "grad_norm": 0.19140625,
+      "learning_rate": 0.0002999984235212234,
+      "loss": 0.4523,
+      "step": 11410
+    },
+    {
+      "epoch": 0.0016008850567332917,
+      "grad_norm": 0.5625,
+      "learning_rate": 0.000299998420490705,
+      "loss": 0.4503,
+      "step": 11420
+    },
+    {
+      "epoch": 0.0016022868825272788,
+      "grad_norm": 0.953125,
+      "learning_rate": 0.00029999841745727664,
+      "loss": 0.3976,
+      "step": 11430
+    },
+    {
+      "epoch": 0.001603688708321266,
+      "grad_norm": 0.77734375,
+      "learning_rate": 0.00029999841442093826,
+      "loss": 0.445,
+      "step": 11440
+    },
+    {
+      "epoch": 0.001605090534115253,
+      "grad_norm": 0.1728515625,
+      "learning_rate": 0.00029999841138168983,
+      "loss": 0.3924,
+      "step": 11450
+    },
+    {
+      "epoch": 0.0016064923599092402,
+      "grad_norm": 0.7890625,
+      "learning_rate": 0.0002999984083395314,
+      "loss": 0.5265,
+      "step": 11460
+    },
+    {
+      "epoch": 0.0016078941857032272,
+      "grad_norm": 0.439453125,
+      "learning_rate": 0.00029999840529446297,
+      "loss": 0.4448,
+      "step": 11470
+    },
+    {
+      "epoch": 0.0016092960114972145,
+      "grad_norm": 1.0234375,
+      "learning_rate": 0.00029999840224648444,
+      "loss": 0.3815,
+      "step": 11480
+    },
+    {
+      "epoch": 0.0016106978372912016,
+      "grad_norm": 0.384765625,
+      "learning_rate": 0.00029999839919559596,
+      "loss": 0.4106,
+      "step": 11490
+    },
+    {
+      "epoch": 0.0016120996630851886,
+      "grad_norm": 0.3125,
+      "learning_rate": 0.00029999839614179737,
+      "loss": 0.4216,
+      "step": 11500
+    },
+    {
+      "epoch": 0.0016120996630851886,
+      "eval_loss": 0.175471693277359,
+      "eval_runtime": 112.4366,
+      "eval_samples_per_second": 203.163,
+      "eval_steps_per_second": 1.592,
+      "step": 11500
+    },
+    {
+      "epoch": 0.001613501488879176,
+      "grad_norm": 0.61328125,
+      "learning_rate": 0.0002999983930850889,
+      "loss": 0.3826,
+      "step": 11510
+    },
+    {
+      "epoch": 0.001614903314673163,
+      "grad_norm": 0.4609375,
+      "learning_rate": 0.0002999983900254703,
+      "loss": 0.4388,
+      "step": 11520
+    },
+    {
+      "epoch": 0.00161630514046715,
+      "grad_norm": 0.81640625,
+      "learning_rate": 0.0002999983869629417,
+      "loss": 0.4358,
+      "step": 11530
+    },
+    {
+      "epoch": 0.001617706966261137,
+      "grad_norm": 0.3984375,
+      "learning_rate": 0.0002999983838975031,
+      "loss": 0.3945,
+      "step": 11540
+    },
+    {
+      "epoch": 0.0016191087920551244,
+      "grad_norm": 0.6171875,
+      "learning_rate": 0.00029999838082915445,
+      "loss": 0.4987,
+      "step": 11550
+    },
+    {
+      "epoch": 0.0016205106178491114,
+      "grad_norm": 0.62890625,
+      "learning_rate": 0.0002999983777578958,
+      "loss": 0.4542,
+      "step": 11560
+    },
+    {
+      "epoch": 0.0016219124436430985,
+      "grad_norm": 0.25,
+      "learning_rate": 0.00029999837468372713,
+      "loss": 0.4642,
+      "step": 11570
+    },
+    {
+      "epoch": 0.0016233142694370855,
+      "grad_norm": 0.2431640625,
+      "learning_rate": 0.0002999983716066484,
+      "loss": 0.4516,
+      "step": 11580
+    },
+    {
+      "epoch": 0.0016247160952310728,
+      "grad_norm": 0.5,
+      "learning_rate": 0.0002999983685266597,
+      "loss": 0.425,
+      "step": 11590
+    },
+    {
+      "epoch": 0.0016261179210250599,
+      "grad_norm": 1.109375,
+      "learning_rate": 0.00029999836544376097,
+      "loss": 0.488,
+      "step": 11600
+    },
+    {
+      "epoch": 0.001627519746819047,
+      "grad_norm": 0.62890625,
+      "learning_rate": 0.0002999983623579522,
+      "loss": 0.3917,
+      "step": 11610
+    },
+    {
+      "epoch": 0.0016289215726130342,
+      "grad_norm": 0.75,
+      "learning_rate": 0.00029999835926923345,
+      "loss": 0.3768,
+      "step": 11620
+    },
+    {
+      "epoch": 0.0016303233984070213,
+      "grad_norm": 0.5234375,
+      "learning_rate": 0.00029999835617760466,
+      "loss": 0.3811,
+      "step": 11630
+    },
+    {
+      "epoch": 0.0016317252242010083,
+      "grad_norm": 0.890625,
+      "learning_rate": 0.0002999983530830658,
+      "loss": 0.4377,
+      "step": 11640
+    },
+    {
+      "epoch": 0.0016331270499949954,
+      "grad_norm": 1.125,
+      "learning_rate": 0.00029999834998561693,
+      "loss": 0.4541,
+      "step": 11650
+    },
+    {
+      "epoch": 0.0016345288757889827,
+      "grad_norm": 1.0546875,
+      "learning_rate": 0.0002999983468852581,
+      "loss": 0.4223,
+      "step": 11660
+    },
+    {
+      "epoch": 0.0016359307015829697,
+      "grad_norm": 0.79296875,
+      "learning_rate": 0.0002999983437819892,
+      "loss": 0.5063,
+      "step": 11670
+    },
+    {
+      "epoch": 0.0016373325273769568,
+      "grad_norm": 0.578125,
+      "learning_rate": 0.0002999983406758103,
+      "loss": 0.4157,
+      "step": 11680
+    },
+    {
+      "epoch": 0.0016387343531709439,
+      "grad_norm": 0.5234375,
+      "learning_rate": 0.00029999833756672137,
+      "loss": 0.4603,
+      "step": 11690
+    },
+    {
+      "epoch": 0.0016401361789649311,
+      "grad_norm": 0.53515625,
+      "learning_rate": 0.00029999833445472243,
+      "loss": 0.4326,
+      "step": 11700
+    },
+    {
+      "epoch": 0.0016415380047589182,
+      "grad_norm": 0.77734375,
+      "learning_rate": 0.00029999833133981344,
+      "loss": 0.355,
+      "step": 11710
+    },
+    {
+      "epoch": 0.0016429398305529053,
+      "grad_norm": 0.31640625,
+      "learning_rate": 0.0002999983282219945,
+      "loss": 0.3694,
+      "step": 11720
+    },
+    {
+      "epoch": 0.0016443416563468925,
+      "grad_norm": 0.5703125,
+      "learning_rate": 0.00029999832510126546,
+      "loss": 0.4451,
+      "step": 11730
+    },
+    {
+      "epoch": 0.0016457434821408796,
+      "grad_norm": 1.046875,
+      "learning_rate": 0.0002999983219776264,
+      "loss": 0.5196,
+      "step": 11740
+    },
+    {
+      "epoch": 0.0016471453079348667,
+      "grad_norm": 0.486328125,
+      "learning_rate": 0.0002999983188510774,
+      "loss": 0.4154,
+      "step": 11750
+    },
+    {
+      "epoch": 0.0016485471337288537,
+      "grad_norm": 0.51953125,
+      "learning_rate": 0.0002999983157216183,
+      "loss": 0.4227,
+      "step": 11760
+    },
+    {
+      "epoch": 0.001649948959522841,
+      "grad_norm": 0.25,
+      "learning_rate": 0.0002999983125892492,
+      "loss": 0.4026,
+      "step": 11770
+    },
+    {
+      "epoch": 0.001651350785316828,
+      "grad_norm": 0.75390625,
+      "learning_rate": 0.0002999983094539701,
+      "loss": 0.5898,
+      "step": 11780
+    },
+    {
+      "epoch": 0.0016527526111108151,
+      "grad_norm": 0.734375,
+      "learning_rate": 0.00029999830631578095,
+      "loss": 0.3777,
+      "step": 11790
+    },
+    {
+      "epoch": 0.0016541544369048024,
+      "grad_norm": 0.68359375,
+      "learning_rate": 0.0002999983031746818,
+      "loss": 0.3929,
+      "step": 11800
+    },
+    {
+      "epoch": 0.0016555562626987895,
+      "grad_norm": 0.21875,
+      "learning_rate": 0.00029999830003067267,
+      "loss": 0.3358,
+      "step": 11810
+    },
+    {
+      "epoch": 0.0016569580884927765,
+      "grad_norm": 0.75390625,
+      "learning_rate": 0.00029999829688375347,
+      "loss": 0.4753,
+      "step": 11820
+    },
+    {
+      "epoch": 0.0016583599142867636,
+      "grad_norm": 0.71875,
+      "learning_rate": 0.0002999982937339242,
+      "loss": 0.4277,
+      "step": 11830
+    },
+    {
+      "epoch": 0.0016597617400807509,
+      "grad_norm": 0.201171875,
+      "learning_rate": 0.00029999829058118503,
+      "loss": 0.3802,
+      "step": 11840
+    },
+    {
+      "epoch": 0.001661163565874738,
+      "grad_norm": 0.478515625,
+      "learning_rate": 0.0002999982874255358,
+      "loss": 0.4961,
+      "step": 11850
+    },
+    {
+      "epoch": 0.001662565391668725,
+      "grad_norm": 0.69921875,
+      "learning_rate": 0.00029999828426697655,
+      "loss": 0.4635,
+      "step": 11860
+    },
+    {
+      "epoch": 0.001663967217462712,
+      "grad_norm": 0.40625,
+      "learning_rate": 0.0002999982811055072,
+      "loss": 0.4261,
+      "step": 11870
+    },
+    {
+      "epoch": 0.0016653690432566993,
+      "grad_norm": 0.46875,
+      "learning_rate": 0.0002999982779411279,
+      "loss": 0.4546,
+      "step": 11880
+    },
+    {
+      "epoch": 0.0016667708690506864,
+      "grad_norm": 0.77734375,
+      "learning_rate": 0.00029999827477383855,
+      "loss": 0.3615,
+      "step": 11890
+    },
+    {
+      "epoch": 0.0016681726948446734,
+      "grad_norm": 0.82421875,
+      "learning_rate": 0.0002999982716036392,
+      "loss": 0.5243,
+      "step": 11900
+    },
+    {
+      "epoch": 0.0016695745206386607,
+      "grad_norm": 0.216796875,
+      "learning_rate": 0.00029999826843052986,
+      "loss": 0.4004,
+      "step": 11910
+    },
+    {
+      "epoch": 0.0016709763464326478,
+      "grad_norm": 0.419921875,
+      "learning_rate": 0.00029999826525451046,
+      "loss": 0.4181,
+      "step": 11920
+    },
+    {
+      "epoch": 0.0016723781722266348,
+      "grad_norm": 0.515625,
+      "learning_rate": 0.00029999826207558107,
+      "loss": 0.4921,
+      "step": 11930
+    },
+    {
+      "epoch": 0.001673779998020622,
+      "grad_norm": 0.9921875,
+      "learning_rate": 0.0002999982588937416,
+      "loss": 0.3902,
+      "step": 11940
+    },
+    {
+      "epoch": 0.0016751818238146092,
+      "grad_norm": 0.43359375,
+      "learning_rate": 0.0002999982557089922,
+      "loss": 0.4505,
+      "step": 11950
+    },
+    {
+      "epoch": 0.0016765836496085962,
+      "grad_norm": 0.5859375,
+      "learning_rate": 0.0002999982525213327,
+      "loss": 0.4233,
+      "step": 11960
+    },
+    {
+      "epoch": 0.0016779854754025833,
+      "grad_norm": 0.51953125,
+      "learning_rate": 0.0002999982493307633,
+      "loss": 0.442,
+      "step": 11970
+    },
+    {
+      "epoch": 0.0016793873011965704,
+      "grad_norm": 0.44921875,
+      "learning_rate": 0.0002999982461372838,
+      "loss": 0.3414,
+      "step": 11980
+    },
+    {
+      "epoch": 0.0016807891269905576,
+      "grad_norm": 0.3984375,
+      "learning_rate": 0.00029999824294089423,
+      "loss": 0.4397,
+      "step": 11990
+    },
+    {
+      "epoch": 0.0016821909527845447,
+      "grad_norm": 0.8828125,
+      "learning_rate": 0.0002999982397415947,
+      "loss": 0.4557,
+      "step": 12000
+    },
+    {
+      "epoch": 0.0016821909527845447,
+      "eval_loss": 0.1744946837425232,
+      "eval_runtime": 114.0463,
+      "eval_samples_per_second": 200.296,
+      "eval_steps_per_second": 1.57,
+      "step": 12000
+    },
+    {
+      "epoch": 0.0016835927785785318,
+      "grad_norm": 0.59765625,
+      "learning_rate": 0.0002999982365393852,
+      "loss": 0.4089,
+      "step": 12010
+    },
+    {
+      "epoch": 0.001684994604372519,
+      "grad_norm": 0.60546875,
+      "learning_rate": 0.0002999982333342656,
+      "loss": 0.441,
+      "step": 12020
+    },
+    {
+      "epoch": 0.001686396430166506,
+      "grad_norm": 0.5703125,
+      "learning_rate": 0.000299998230126236,
+      "loss": 0.3261,
+      "step": 12030
+    },
+    {
+      "epoch": 0.0016877982559604932,
+      "grad_norm": 0.60546875,
+      "learning_rate": 0.0002999982269152964,
+      "loss": 0.4775,
+      "step": 12040
+    },
+    {
+      "epoch": 0.0016892000817544802,
+      "grad_norm": 0.80078125,
+      "learning_rate": 0.00029999822370144674,
+      "loss": 0.4376,
+      "step": 12050
+    },
+    {
+      "epoch": 0.0016906019075484675,
+      "grad_norm": 0.494140625,
+      "learning_rate": 0.0002999982204846871,
+      "loss": 0.4219,
+      "step": 12060
+    },
+    {
+      "epoch": 0.0016920037333424546,
+      "grad_norm": 0.65625,
+      "learning_rate": 0.00029999821726501744,
+      "loss": 0.3873,
+      "step": 12070
+    },
+    {
+      "epoch": 0.0016934055591364416,
+      "grad_norm": 0.181640625,
+      "learning_rate": 0.00029999821404243774,
+      "loss": 0.4289,
+      "step": 12080
+    },
+    {
+      "epoch": 0.001694807384930429,
+      "grad_norm": 0.26171875,
+      "learning_rate": 0.00029999821081694804,
+      "loss": 0.4635,
+      "step": 12090
+    },
+    {
+      "epoch": 0.001696209210724416,
+      "grad_norm": 0.88671875,
+      "learning_rate": 0.0002999982075885483,
+      "loss": 0.4446,
+      "step": 12100
+    },
+    {
+      "epoch": 0.001697611036518403,
+      "grad_norm": 0.5234375,
+      "learning_rate": 0.0002999982043572386,
+      "loss": 0.3795,
+      "step": 12110
+    },
+    {
+      "epoch": 0.00169901286231239,
+      "grad_norm": 0.8828125,
+      "learning_rate": 0.00029999820112301884,
+      "loss": 0.4547,
+      "step": 12120
+    },
+    {
+      "epoch": 0.0017004146881063774,
+      "grad_norm": 0.78125,
+      "learning_rate": 0.00029999819788588904,
+      "loss": 0.4917,
+      "step": 12130
+    },
+    {
+      "epoch": 0.0017018165139003644,
+      "grad_norm": 0.63671875,
+      "learning_rate": 0.00029999819464584924,
+      "loss": 0.3711,
+      "step": 12140
+    },
+    {
+      "epoch": 0.0017032183396943515,
+      "grad_norm": 0.28515625,
+      "learning_rate": 0.00029999819140289944,
+      "loss": 0.4076,
+      "step": 12150
+    },
+    {
+      "epoch": 0.0017046201654883385,
+      "grad_norm": 0.8359375,
+      "learning_rate": 0.00029999818815703964,
+      "loss": 0.4338,
+      "step": 12160
+    },
+    {
+      "epoch": 0.0017060219912823258,
+      "grad_norm": 0.404296875,
+      "learning_rate": 0.0002999981849082698,
+      "loss": 0.3822,
+      "step": 12170
+    },
+    {
+      "epoch": 0.0017074238170763129,
+      "grad_norm": 0.734375,
+      "learning_rate": 0.00029999818165658994,
+      "loss": 0.4759,
+      "step": 12180
+    },
+    {
+      "epoch": 0.0017088256428703,
+      "grad_norm": 0.353515625,
+      "learning_rate": 0.0002999981784020001,
+      "loss": 0.3642,
+      "step": 12190
+    },
+    {
+      "epoch": 0.0017102274686642872,
+      "grad_norm": 0.6484375,
+      "learning_rate": 0.00029999817514450014,
+      "loss": 0.461,
+      "step": 12200
+    },
+    {
+      "epoch": 0.0017116292944582743,
+      "grad_norm": 0.8828125,
+      "learning_rate": 0.00029999817188409024,
+      "loss": 0.4225,
+      "step": 12210
+    },
+    {
+      "epoch": 0.0017130311202522613,
+      "grad_norm": 0.62109375,
+      "learning_rate": 0.0002999981686207703,
+      "loss": 0.4653,
+      "step": 12220
+    },
+    {
+      "epoch": 0.0017144329460462484,
+      "grad_norm": 0.396484375,
+      "learning_rate": 0.00029999816535454033,
+      "loss": 0.4366,
+      "step": 12230
+    },
+    {
+      "epoch": 0.0017158347718402357,
+      "grad_norm": 0.6953125,
+      "learning_rate": 0.0002999981620854004,
+      "loss": 0.3372,
+      "step": 12240
+    },
+    {
+      "epoch": 0.0017172365976342227,
+      "grad_norm": 0.74609375,
+      "learning_rate": 0.0002999981588133504,
+      "loss": 0.3731,
+      "step": 12250
+    },
+    {
+      "epoch": 0.0017186384234282098,
+      "grad_norm": 0.484375,
+      "learning_rate": 0.0002999981555383904,
+      "loss": 0.4849,
+      "step": 12260
+    },
+    {
+      "epoch": 0.0017200402492221969,
+      "grad_norm": 0.796875,
+      "learning_rate": 0.0002999981522605204,
+      "loss": 0.4366,
+      "step": 12270
+    },
+    {
+      "epoch": 0.0017214420750161841,
+      "grad_norm": 0.9375,
+      "learning_rate": 0.0002999981489797403,
+      "loss": 0.4627,
+      "step": 12280
+    },
+    {
+      "epoch": 0.0017228439008101712,
+      "grad_norm": 0.79296875,
+      "learning_rate": 0.0002999981456960503,
+      "loss": 0.4922,
+      "step": 12290
+    },
+    {
+      "epoch": 0.0017242457266041583,
+      "grad_norm": 0.640625,
+      "learning_rate": 0.0002999981424094502,
+      "loss": 0.5141,
+      "step": 12300
+    },
+    {
+      "epoch": 0.0017256475523981455,
+      "grad_norm": 0.58984375,
+      "learning_rate": 0.0002999981391199401,
+      "loss": 0.4268,
+      "step": 12310
+    },
+    {
+      "epoch": 0.0017270493781921326,
+      "grad_norm": 0.5625,
+      "learning_rate": 0.00029999813582752007,
+      "loss": 0.4026,
+      "step": 12320
+    },
+    {
+      "epoch": 0.0017284512039861197,
+      "grad_norm": 0.61328125,
+      "learning_rate": 0.0002999981325321899,
+      "loss": 0.3239,
+      "step": 12330
+    },
+    {
+      "epoch": 0.0017298530297801067,
+      "grad_norm": 0.244140625,
+      "learning_rate": 0.00029999812923394976,
+      "loss": 0.3925,
+      "step": 12340
+    },
+    {
+      "epoch": 0.001731254855574094,
+      "grad_norm": 0.4765625,
+      "learning_rate": 0.0002999981259327996,
+      "loss": 0.3677,
+      "step": 12350
+    },
+    {
+      "epoch": 0.001732656681368081,
+      "grad_norm": 0.703125,
+      "learning_rate": 0.00029999812262873946,
+      "loss": 0.4404,
+      "step": 12360
+    },
+    {
+      "epoch": 0.0017340585071620681,
+      "grad_norm": 0.83203125,
+      "learning_rate": 0.0002999981193217693,
+      "loss": 0.4245,
+      "step": 12370
+    },
+    {
+      "epoch": 0.0017354603329560554,
+      "grad_norm": 0.361328125,
+      "learning_rate": 0.00029999811601188906,
+      "loss": 0.4636,
+      "step": 12380
+    },
+    {
+      "epoch": 0.0017368621587500424,
+      "grad_norm": 0.30859375,
+      "learning_rate": 0.00029999811269909886,
+      "loss": 0.3831,
+      "step": 12390
+    },
+    {
+      "epoch": 0.0017382639845440295,
+      "grad_norm": 0.69140625,
+      "learning_rate": 0.00029999810938339866,
+      "loss": 0.4938,
+      "step": 12400
+    },
+    {
+      "epoch": 0.0017396658103380166,
+      "grad_norm": 1.0,
+      "learning_rate": 0.0002999981060647884,
+      "loss": 0.4036,
+      "step": 12410
+    },
+    {
+      "epoch": 0.0017410676361320038,
+      "grad_norm": 0.578125,
+      "learning_rate": 0.0002999981027432681,
+      "loss": 0.4091,
+      "step": 12420
+    },
+    {
+      "epoch": 0.001742469461925991,
+      "grad_norm": 0.443359375,
+      "learning_rate": 0.00029999809941883785,
+      "loss": 0.4134,
+      "step": 12430
+    },
+    {
+      "epoch": 0.001743871287719978,
+      "grad_norm": 0.67578125,
+      "learning_rate": 0.0002999980960914975,
+      "loss": 0.4591,
+      "step": 12440
+    },
+    {
+      "epoch": 0.001745273113513965,
+      "grad_norm": 0.228515625,
+      "learning_rate": 0.00029999809276124725,
+      "loss": 0.4345,
+      "step": 12450
+    },
+    {
+      "epoch": 0.0017466749393079523,
+      "grad_norm": 0.71484375,
+      "learning_rate": 0.0002999980894280869,
+      "loss": 0.3948,
+      "step": 12460
+    },
+    {
+      "epoch": 0.0017480767651019394,
+      "grad_norm": 0.3515625,
+      "learning_rate": 0.00029999808609201654,
+      "loss": 0.4106,
+      "step": 12470
+    },
+    {
+      "epoch": 0.0017494785908959264,
+      "grad_norm": 0.24609375,
+      "learning_rate": 0.0002999980827530362,
+      "loss": 0.4008,
+      "step": 12480
+    },
+    {
+      "epoch": 0.0017508804166899137,
+      "grad_norm": 0.23046875,
+      "learning_rate": 0.00029999807941114584,
+      "loss": 0.3922,
+      "step": 12490
+    },
+    {
+      "epoch": 0.0017522822424839008,
+      "grad_norm": 0.189453125,
+      "learning_rate": 0.00029999807606634544,
+      "loss": 0.4446,
+      "step": 12500
+    },
+    {
+      "epoch": 0.0017522822424839008,
+      "eval_loss": 0.17290730774402618,
+      "eval_runtime": 112.4498,
+      "eval_samples_per_second": 203.139,
+      "eval_steps_per_second": 1.592,
+      "step": 12500
+    },
+    {
+      "epoch": 0.0017536840682778878,
+      "grad_norm": 0.77734375,
+      "learning_rate": 0.000299998072718635,
+      "loss": 0.4245,
+      "step": 12510
+    },
+    {
+      "epoch": 0.0017550858940718749,
+      "grad_norm": 0.75390625,
+      "learning_rate": 0.0002999980693680146,
+      "loss": 0.3517,
+      "step": 12520
+    },
+    {
+      "epoch": 0.0017564877198658622,
+      "grad_norm": 0.6328125,
+      "learning_rate": 0.0002999980660144842,
+      "loss": 0.4871,
+      "step": 12530
+    },
+    {
+      "epoch": 0.0017578895456598492,
+      "grad_norm": 0.298828125,
+      "learning_rate": 0.00029999806265804373,
+      "loss": 0.4103,
+      "step": 12540
+    },
+    {
+      "epoch": 0.0017592913714538363,
+      "grad_norm": 0.50390625,
+      "learning_rate": 0.00029999805929869323,
+      "loss": 0.4575,
+      "step": 12550
+    },
+    {
+      "epoch": 0.0017606931972478234,
+      "grad_norm": 0.8984375,
+      "learning_rate": 0.00029999805593643273,
+      "loss": 0.5051,
+      "step": 12560
+    },
+    {
+      "epoch": 0.0017620950230418106,
+      "grad_norm": 0.59765625,
+      "learning_rate": 0.0002999980525712622,
+      "loss": 0.4247,
+      "step": 12570
+    },
+    {
+      "epoch": 0.0017634968488357977,
+      "grad_norm": 1.0390625,
+      "learning_rate": 0.0002999980492031818,
+      "loss": 0.3596,
+      "step": 12580
+    },
+    {
+      "epoch": 0.0017648986746297847,
+      "grad_norm": 0.82421875,
+      "learning_rate": 0.00029999804583219123,
+      "loss": 0.4019,
+      "step": 12590
+    },
+    {
+      "epoch": 0.001766300500423772,
+      "grad_norm": 0.375,
+      "learning_rate": 0.0002999980424582906,
+      "loss": 0.3925,
+      "step": 12600
+    },
+    {
+      "epoch": 0.001767702326217759,
+      "grad_norm": 0.37109375,
+      "learning_rate": 0.00029999803908148013,
+      "loss": 0.492,
+      "step": 12610
+    },
+    {
+      "epoch": 0.0017691041520117461,
+      "grad_norm": 0.88671875,
+      "learning_rate": 0.0002999980357017595,
+      "loss": 0.4479,
+      "step": 12620
+    },
+    {
+      "epoch": 0.0017705059778057332,
+      "grad_norm": 0.462890625,
+      "learning_rate": 0.0002999980323191289,
+      "loss": 0.4261,
+      "step": 12630
+    },
+    {
+      "epoch": 0.0017719078035997205,
+      "grad_norm": 0.9375,
+      "learning_rate": 0.0002999980289335884,
+      "loss": 0.5056,
+      "step": 12640
+    },
+    {
+      "epoch": 0.0017733096293937075,
+      "grad_norm": 0.62109375,
+      "learning_rate": 0.0002999980255451377,
+      "loss": 0.4311,
+      "step": 12650
+    },
+    {
+      "epoch": 0.0017747114551876946,
+      "grad_norm": 0.73046875,
+      "learning_rate": 0.00029999802215377713,
+      "loss": 0.5066,
+      "step": 12660
+    },
+    {
+      "epoch": 0.0017761132809816819,
+      "grad_norm": 0.78515625,
+      "learning_rate": 0.0002999980187595064,
+      "loss": 0.4447,
+      "step": 12670
+    },
+    {
+      "epoch": 0.001777515106775669,
+      "grad_norm": 0.85546875,
+      "learning_rate": 0.0002999980153623258,
+      "loss": 0.4514,
+      "step": 12680
+    },
+    {
+      "epoch": 0.001778916932569656,
+      "grad_norm": 0.4296875,
+      "learning_rate": 0.0002999980119622351,
+      "loss": 0.3377,
+      "step": 12690
+    },
+    {
+      "epoch": 0.001780318758363643,
+      "grad_norm": 0.8203125,
+      "learning_rate": 0.0002999980085592344,
+      "loss": 0.4139,
+      "step": 12700
+    },
+    {
+      "epoch": 0.0017817205841576303,
+      "grad_norm": 0.2578125,
+      "learning_rate": 0.00029999800515332373,
+      "loss": 0.3494,
+      "step": 12710
+    },
+    {
+      "epoch": 0.0017831224099516174,
+      "grad_norm": 0.4375,
+      "learning_rate": 0.00029999800174450303,
+      "loss": 0.3762,
+      "step": 12720
+    },
+    {
+      "epoch": 0.0017845242357456045,
+      "grad_norm": 1.328125,
+      "learning_rate": 0.0002999979983327723,
+      "loss": 0.4438,
+      "step": 12730
+    },
+    {
+      "epoch": 0.0017859260615395915,
+      "grad_norm": 0.859375,
+      "learning_rate": 0.00029999799491813153,
+      "loss": 0.359,
+      "step": 12740
+    },
+    {
+      "epoch": 0.0017873278873335788,
+      "grad_norm": 0.9375,
+      "learning_rate": 0.0002999979915005808,
+      "loss": 0.5009,
+      "step": 12750
+    },
+    {
+      "epoch": 0.0017887297131275659,
+      "grad_norm": 1.015625,
+      "learning_rate": 0.00029999798808012004,
+      "loss": 0.4575,
+      "step": 12760
+    },
+    {
+      "epoch": 0.001790131538921553,
+      "grad_norm": 0.5859375,
+      "learning_rate": 0.00029999798465674924,
+      "loss": 0.4969,
+      "step": 12770
+    },
+    {
+      "epoch": 0.0017915333647155402,
+      "grad_norm": 0.62109375,
+      "learning_rate": 0.00029999798123046844,
+      "loss": 0.5035,
+      "step": 12780
+    },
+    {
+      "epoch": 0.0017929351905095273,
+      "grad_norm": 0.47265625,
+      "learning_rate": 0.00029999797780127764,
+      "loss": 0.3995,
+      "step": 12790
+    },
+    {
+      "epoch": 0.0017943370163035143,
+      "grad_norm": 0.8203125,
+      "learning_rate": 0.00029999797436917674,
+      "loss": 0.5084,
+      "step": 12800
+    },
+    {
+      "epoch": 0.0017957388420975014,
+      "grad_norm": 0.408203125,
+      "learning_rate": 0.00029999797093416595,
+      "loss": 0.4455,
+      "step": 12810
+    },
+    {
+      "epoch": 0.0017971406678914887,
+      "grad_norm": 0.87890625,
+      "learning_rate": 0.00029999796749624505,
+      "loss": 0.4266,
+      "step": 12820
+    },
+    {
+      "epoch": 0.0017985424936854757,
+      "grad_norm": 0.470703125,
+      "learning_rate": 0.0002999979640554142,
+      "loss": 0.4251,
+      "step": 12830
+    },
+    {
+      "epoch": 0.0017999443194794628,
+      "grad_norm": 0.63671875,
+      "learning_rate": 0.00029999796061167336,
+      "loss": 0.4692,
+      "step": 12840
+    },
+    {
+      "epoch": 0.0018013461452734498,
+      "grad_norm": 0.734375,
+      "learning_rate": 0.00029999795716502246,
+      "loss": 0.5269,
+      "step": 12850
+    },
+    {
+      "epoch": 0.0018027479710674371,
+      "grad_norm": 0.765625,
+      "learning_rate": 0.00029999795371546157,
+      "loss": 0.436,
+      "step": 12860
+    },
+    {
+      "epoch": 0.0018041497968614242,
+      "grad_norm": 0.353515625,
+      "learning_rate": 0.0002999979502629906,
+      "loss": 0.4383,
+      "step": 12870
+    },
+    {
+      "epoch": 0.0018055516226554112,
+      "grad_norm": 0.6953125,
+      "learning_rate": 0.00029999794680760967,
+      "loss": 0.4605,
+      "step": 12880
+    },
+    {
+      "epoch": 0.0018069534484493985,
+      "grad_norm": 0.63671875,
+      "learning_rate": 0.0002999979433493187,
+      "loss": 0.4385,
+      "step": 12890
+    },
+    {
+      "epoch": 0.0018083552742433856,
+      "grad_norm": 0.69140625,
+      "learning_rate": 0.0002999979398881178,
+      "loss": 0.4601,
+      "step": 12900
+    },
+    {
+      "epoch": 0.0018097571000373726,
+      "grad_norm": 0.578125,
+      "learning_rate": 0.00029999793642400684,
+      "loss": 0.4269,
+      "step": 12910
+    },
+    {
+      "epoch": 0.0018111589258313597,
+      "grad_norm": 0.5,
+      "learning_rate": 0.00029999793295698584,
+      "loss": 0.3948,
+      "step": 12920
+    },
+    {
+      "epoch": 0.001812560751625347,
+      "grad_norm": 0.56640625,
+      "learning_rate": 0.0002999979294870548,
+      "loss": 0.412,
+      "step": 12930
+    },
+    {
+      "epoch": 0.001813962577419334,
+      "grad_norm": 0.6484375,
+      "learning_rate": 0.0002999979260142138,
+      "loss": 0.4961,
+      "step": 12940
+    },
+    {
+      "epoch": 0.001815364403213321,
+      "grad_norm": 0.33984375,
+      "learning_rate": 0.00029999792253846276,
+      "loss": 0.4089,
+      "step": 12950
+    },
+    {
+      "epoch": 0.0018167662290073084,
+      "grad_norm": 0.91015625,
+      "learning_rate": 0.00029999791905980177,
+      "loss": 0.388,
+      "step": 12960
+    },
+    {
+      "epoch": 0.0018181680548012954,
+      "grad_norm": 0.267578125,
+      "learning_rate": 0.0002999979155782307,
+      "loss": 0.4039,
+      "step": 12970
+    },
+    {
+      "epoch": 0.0018195698805952825,
+      "grad_norm": 0.318359375,
+      "learning_rate": 0.00029999791209374963,
+      "loss": 0.4272,
+      "step": 12980
+    },
+    {
+      "epoch": 0.0018209717063892696,
+      "grad_norm": 0.7734375,
+      "learning_rate": 0.0002999979086063586,
+      "loss": 0.4296,
+      "step": 12990
+    },
+    {
+      "epoch": 0.0018223735321832568,
+      "grad_norm": 0.5078125,
+      "learning_rate": 0.0002999979051160575,
+      "loss": 0.4481,
+      "step": 13000
+    },
+    {
+      "epoch": 0.0018223735321832568,
+      "eval_loss": 0.17369048297405243,
+      "eval_runtime": 112.6443,
+      "eval_samples_per_second": 202.789,
+      "eval_steps_per_second": 1.589,
+      "step": 13000
+    },
+    {
+      "epoch": 0.001823775357977244,
+      "grad_norm": 1.125,
+      "learning_rate": 0.0002999979016228464,
+      "loss": 0.4938,
+      "step": 13010
+    },
+    {
+      "epoch": 0.001825177183771231,
+      "grad_norm": 0.67578125,
+      "learning_rate": 0.00029999789812672526,
+      "loss": 0.3418,
+      "step": 13020
+    },
+    {
+      "epoch": 0.001826579009565218,
+      "grad_norm": 0.56640625,
+      "learning_rate": 0.00029999789462769417,
+      "loss": 0.4171,
+      "step": 13030
+    },
+    {
+      "epoch": 0.0018279808353592053,
+      "grad_norm": 0.6953125,
+      "learning_rate": 0.00029999789112575303,
+      "loss": 0.4075,
+      "step": 13040
+    },
+    {
+      "epoch": 0.0018293826611531924,
+      "grad_norm": 0.57421875,
+      "learning_rate": 0.0002999978876209019,
+      "loss": 0.3656,
+      "step": 13050
+    },
+    {
+      "epoch": 0.0018307844869471794,
+      "grad_norm": 0.76171875,
+      "learning_rate": 0.0002999978841131407,
+      "loss": 0.4442,
+      "step": 13060
+    },
+    {
+      "epoch": 0.0018321863127411667,
+      "grad_norm": 0.333984375,
+      "learning_rate": 0.00029999788060246956,
+      "loss": 0.349,
+      "step": 13070
+    },
+    {
+      "epoch": 0.0018335881385351538,
+      "grad_norm": 0.328125,
+      "learning_rate": 0.0002999978770888884,
+      "loss": 0.3392,
+      "step": 13080
+    },
+    {
+      "epoch": 0.0018349899643291408,
+      "grad_norm": 0.169921875,
+      "learning_rate": 0.0002999978735723972,
+      "loss": 0.3594,
+      "step": 13090
+    },
+    {
+      "epoch": 0.0018363917901231279,
+      "grad_norm": 0.79296875,
+      "learning_rate": 0.000299997870052996,
+      "loss": 0.4715,
+      "step": 13100
+    },
+    {
+      "epoch": 0.0018377936159171152,
+      "grad_norm": 0.150390625,
+      "learning_rate": 0.0002999978665306848,
+      "loss": 0.4352,
+      "step": 13110
+    },
+    {
+      "epoch": 0.0018391954417111022,
+      "grad_norm": 0.2099609375,
+      "learning_rate": 0.00029999786300546357,
+      "loss": 0.3783,
+      "step": 13120
+    },
+    {
+      "epoch": 0.0018405972675050893,
+      "grad_norm": 0.5625,
+      "learning_rate": 0.00029999785947733233,
+      "loss": 0.4245,
+      "step": 13130
+    },
+    {
+      "epoch": 0.0018419990932990763,
+      "grad_norm": 0.7578125,
+      "learning_rate": 0.00029999785594629104,
+      "loss": 0.3549,
+      "step": 13140
+    },
+    {
+      "epoch": 0.0018434009190930636,
+      "grad_norm": 0.859375,
+      "learning_rate": 0.0002999978524123398,
+      "loss": 0.4832,
+      "step": 13150
+    },
+    {
+      "epoch": 0.0018448027448870507,
+      "grad_norm": 0.423828125,
+      "learning_rate": 0.0002999978488754785,
+      "loss": 0.3666,
+      "step": 13160
+    },
+    {
+      "epoch": 0.0018462045706810377,
+      "grad_norm": 0.50390625,
+      "learning_rate": 0.00029999784533570723,
+      "loss": 0.3677,
+      "step": 13170
+    },
+    {
+      "epoch": 0.001847606396475025,
+      "grad_norm": 0.4453125,
+      "learning_rate": 0.000299997841793026,
+      "loss": 0.465,
+      "step": 13180
+    },
+    {
+      "epoch": 0.001849008222269012,
+      "grad_norm": 0.72265625,
+      "learning_rate": 0.00029999783824743466,
+      "loss": 0.388,
+      "step": 13190
+    },
+    {
+      "epoch": 0.0018504100480629991,
+      "grad_norm": 0.59765625,
+      "learning_rate": 0.0002999978346989334,
+      "loss": 0.3845,
+      "step": 13200
+    },
+    {
+      "epoch": 0.0018518118738569862,
+      "grad_norm": 1.015625,
+      "learning_rate": 0.00029999783114752204,
+      "loss": 0.3664,
+      "step": 13210
+    },
+    {
+      "epoch": 0.0018532136996509735,
+      "grad_norm": 0.87109375,
+      "learning_rate": 0.0002999978275932007,
+      "loss": 0.5365,
+      "step": 13220
+    },
+    {
+      "epoch": 0.0018546155254449605,
+      "grad_norm": 0.25,
+      "learning_rate": 0.0002999978240359693,
+      "loss": 0.4392,
+      "step": 13230
+    },
+    {
+      "epoch": 0.0018560173512389476,
+      "grad_norm": 0.6171875,
+      "learning_rate": 0.000299997820475828,
+      "loss": 0.3636,
+      "step": 13240
+    },
+    {
+      "epoch": 0.0018574191770329349,
+      "grad_norm": 0.74609375,
+      "learning_rate": 0.00029999781691277666,
+      "loss": 0.4864,
+      "step": 13250
+    },
+    {
+      "epoch": 0.001858821002826922,
+      "grad_norm": 0.275390625,
+      "learning_rate": 0.0002999978133468152,
+      "loss": 0.4072,
+      "step": 13260
+    },
+    {
+      "epoch": 0.001860222828620909,
+      "grad_norm": 0.55859375,
+      "learning_rate": 0.00029999780977794385,
+      "loss": 0.4495,
+      "step": 13270
+    },
+    {
+      "epoch": 0.001861624654414896,
+      "grad_norm": 0.2890625,
+      "learning_rate": 0.00029999780620616247,
+      "loss": 0.3696,
+      "step": 13280
+    },
+    {
+      "epoch": 0.0018630264802088833,
+      "grad_norm": 0.5234375,
+      "learning_rate": 0.00029999780263147103,
+      "loss": 0.4009,
+      "step": 13290
+    },
+    {
+      "epoch": 0.0018644283060028704,
+      "grad_norm": 0.58203125,
+      "learning_rate": 0.0002999977990538696,
+      "loss": 0.3759,
+      "step": 13300
+    },
+    {
+      "epoch": 0.0018658301317968575,
+      "grad_norm": 0.166015625,
+      "learning_rate": 0.0002999977954733582,
+      "loss": 0.3558,
+      "step": 13310
+    },
+    {
+      "epoch": 0.0018672319575908445,
+      "grad_norm": 0.69921875,
+      "learning_rate": 0.0002999977918899368,
+      "loss": 0.3603,
+      "step": 13320
+    },
+    {
+      "epoch": 0.0018686337833848318,
+      "grad_norm": 0.6875,
+      "learning_rate": 0.0002999977883036053,
+      "loss": 0.4357,
+      "step": 13330
+    },
+    {
+      "epoch": 0.0018700356091788189,
+      "grad_norm": 1.0234375,
+      "learning_rate": 0.00029999778471436384,
+      "loss": 0.4183,
+      "step": 13340
+    },
+    {
+      "epoch": 0.001871437434972806,
+      "grad_norm": 1.0703125,
+      "learning_rate": 0.0002999977811222124,
+      "loss": 0.4305,
+      "step": 13350
+    },
+    {
+      "epoch": 0.0018728392607667932,
+      "grad_norm": 0.66015625,
+      "learning_rate": 0.0002999977775271509,
+      "loss": 0.4278,
+      "step": 13360
+    },
+    {
+      "epoch": 0.0018742410865607803,
+      "grad_norm": 1.1640625,
+      "learning_rate": 0.0002999977739291794,
+      "loss": 0.4415,
+      "step": 13370
+    },
+    {
+      "epoch": 0.0018756429123547673,
+      "grad_norm": 0.84765625,
+      "learning_rate": 0.00029999777032829793,
+      "loss": 0.4681,
+      "step": 13380
+    },
+    {
+      "epoch": 0.0018770447381487544,
+      "grad_norm": 0.83203125,
+      "learning_rate": 0.00029999776672450646,
+      "loss": 0.4305,
+      "step": 13390
+    },
+    {
+      "epoch": 0.0018784465639427417,
+      "grad_norm": 1.0859375,
+      "learning_rate": 0.0002999977631178049,
+      "loss": 0.4568,
+      "step": 13400
+    },
+    {
+      "epoch": 0.0018798483897367287,
+      "grad_norm": 0.74609375,
+      "learning_rate": 0.00029999775950819335,
+      "loss": 0.4561,
+      "step": 13410
+    },
+    {
+      "epoch": 0.0018812502155307158,
+      "grad_norm": 0.7421875,
+      "learning_rate": 0.00029999775589567183,
+      "loss": 0.4982,
+      "step": 13420
+    },
+    {
+      "epoch": 0.0018826520413247028,
+      "grad_norm": 0.3828125,
+      "learning_rate": 0.0002999977522802403,
+      "loss": 0.4583,
+      "step": 13430
+    },
+    {
+      "epoch": 0.0018840538671186901,
+      "grad_norm": 1.015625,
+      "learning_rate": 0.0002999977486618988,
+      "loss": 0.5271,
+      "step": 13440
+    },
+    {
+      "epoch": 0.0018854556929126772,
+      "grad_norm": 0.69921875,
+      "learning_rate": 0.0002999977450406472,
+      "loss": 0.4311,
+      "step": 13450
+    },
+    {
+      "epoch": 0.0018868575187066642,
+      "grad_norm": 0.6640625,
+      "learning_rate": 0.00029999774141648564,
+      "loss": 0.4316,
+      "step": 13460
+    },
+    {
+      "epoch": 0.0018882593445006515,
+      "grad_norm": 0.58203125,
+      "learning_rate": 0.000299997737789414,
+      "loss": 0.3692,
+      "step": 13470
+    },
+    {
+      "epoch": 0.0018896611702946386,
+      "grad_norm": 0.72265625,
+      "learning_rate": 0.00029999773415943245,
+      "loss": 0.4599,
+      "step": 13480
+    },
+    {
+      "epoch": 0.0018910629960886256,
+      "grad_norm": 0.8828125,
+      "learning_rate": 0.0002999977305265408,
+      "loss": 0.5223,
+      "step": 13490
+    },
+    {
+      "epoch": 0.0018924648218826127,
+      "grad_norm": 1.1875,
+      "learning_rate": 0.0002999977268907392,
+      "loss": 0.4116,
+      "step": 13500
+    },
+    {
+      "epoch": 0.0018924648218826127,
+      "eval_loss": 0.17309853434562683,
+      "eval_runtime": 114.1817,
+      "eval_samples_per_second": 200.058,
+      "eval_steps_per_second": 1.568,
+      "step": 13500
+    },
+    {
+      "epoch": 0.0018938666476766,
+      "grad_norm": 0.94921875,
+      "learning_rate": 0.00029999772325202764,
+      "loss": 0.519,
+      "step": 13510
+    },
+    {
+      "epoch": 0.001895268473470587,
+      "grad_norm": 0.6484375,
+      "learning_rate": 0.00029999771961040597,
+      "loss": 0.4413,
+      "step": 13520
+    },
+    {
+      "epoch": 0.001896670299264574,
+      "grad_norm": 0.765625,
+      "learning_rate": 0.00029999771596587435,
+      "loss": 0.4443,
+      "step": 13530
+    },
+    {
+      "epoch": 0.0018980721250585614,
+      "grad_norm": 0.81640625,
+      "learning_rate": 0.00029999771231843273,
+      "loss": 0.486,
+      "step": 13540
+    },
+    {
+      "epoch": 0.0018994739508525484,
+      "grad_norm": 0.5234375,
+      "learning_rate": 0.000299997708668081,
+      "loss": 0.3188,
+      "step": 13550
+    },
+    {
+      "epoch": 0.0019008757766465355,
+      "grad_norm": 0.294921875,
+      "learning_rate": 0.0002999977050148194,
+      "loss": 0.3781,
+      "step": 13560
+    },
+    {
+      "epoch": 0.0019022776024405226,
+      "grad_norm": 0.73046875,
+      "learning_rate": 0.00029999770135864773,
+      "loss": 0.4946,
+      "step": 13570
+    },
+    {
+      "epoch": 0.0019036794282345098,
+      "grad_norm": 0.75390625,
+      "learning_rate": 0.00029999769769956607,
+      "loss": 0.3202,
+      "step": 13580
+    },
+    {
+      "epoch": 0.001905081254028497,
+      "grad_norm": 0.77734375,
+      "learning_rate": 0.00029999769403757435,
+      "loss": 0.4782,
+      "step": 13590
+    },
+    {
+      "epoch": 0.001906483079822484,
+      "grad_norm": 0.65625,
+      "learning_rate": 0.0002999976903726727,
+      "loss": 0.4282,
+      "step": 13600
+    },
+    {
+      "epoch": 0.001907884905616471,
+      "grad_norm": 0.54296875,
+      "learning_rate": 0.000299997686704861,
+      "loss": 0.4034,
+      "step": 13610
+    },
+    {
+      "epoch": 0.0019092867314104583,
+      "grad_norm": 0.62109375,
+      "learning_rate": 0.00029999768303413927,
+      "loss": 0.4568,
+      "step": 13620
+    },
+    {
+      "epoch": 0.0019106885572044454,
+      "grad_norm": 0.291015625,
+      "learning_rate": 0.00029999767936050755,
+      "loss": 0.3956,
+      "step": 13630
+    },
+    {
+      "epoch": 0.0019120903829984324,
+      "grad_norm": 0.66015625,
+      "learning_rate": 0.00029999767568396584,
+      "loss": 0.4545,
+      "step": 13640
+    },
+    {
+      "epoch": 0.0019134922087924197,
+      "grad_norm": 0.30078125,
+      "learning_rate": 0.00029999767200451414,
+      "loss": 0.489,
+      "step": 13650
+    },
+    {
+      "epoch": 0.0019148940345864068,
+      "grad_norm": 1.28125,
+      "learning_rate": 0.0002999976683221524,
+      "loss": 0.536,
+      "step": 13660
+    },
+    {
+      "epoch": 0.0019162958603803938,
+      "grad_norm": 0.2578125,
+      "learning_rate": 0.0002999976646368806,
+      "loss": 0.3801,
+      "step": 13670
+    },
+    {
+      "epoch": 0.0019176976861743809,
+      "grad_norm": 0.52734375,
+      "learning_rate": 0.00029999766094869886,
+      "loss": 0.4321,
+      "step": 13680
+    },
+    {
+      "epoch": 0.0019190995119683682,
+      "grad_norm": 0.25390625,
+      "learning_rate": 0.0002999976572576071,
+      "loss": 0.3094,
+      "step": 13690
+    },
+    {
+      "epoch": 0.0019205013377623552,
+      "grad_norm": 0.39453125,
+      "learning_rate": 0.00029999765356360534,
+      "loss": 0.3829,
+      "step": 13700
+    },
+    {
+      "epoch": 0.0019219031635563423,
+      "grad_norm": 0.423828125,
+      "learning_rate": 0.00029999764986669353,
+      "loss": 0.4466,
+      "step": 13710
+    },
+    {
+      "epoch": 0.0019233049893503293,
+      "grad_norm": 0.66015625,
+      "learning_rate": 0.0002999976461668718,
+      "loss": 0.4148,
+      "step": 13720
+    },
+    {
+      "epoch": 0.0019247068151443166,
+      "grad_norm": 0.83984375,
+      "learning_rate": 0.00029999764246414,
+      "loss": 0.4261,
+      "step": 13730
+    },
+    {
+      "epoch": 0.0019261086409383037,
+      "grad_norm": 0.48828125,
+      "learning_rate": 0.00029999763875849817,
+      "loss": 0.4347,
+      "step": 13740
+    },
+    {
+      "epoch": 0.0019275104667322907,
+      "grad_norm": 0.515625,
+      "learning_rate": 0.00029999763504994637,
+      "loss": 0.4352,
+      "step": 13750
+    },
+    {
+      "epoch": 0.001928912292526278,
+      "grad_norm": 0.5703125,
+      "learning_rate": 0.00029999763133848456,
+      "loss": 0.428,
+      "step": 13760
+    },
+    {
+      "epoch": 0.001930314118320265,
+      "grad_norm": 0.45703125,
+      "learning_rate": 0.00029999762762411276,
+      "loss": 0.337,
+      "step": 13770
+    },
+    {
+      "epoch": 0.0019317159441142521,
+      "grad_norm": 0.60546875,
+      "learning_rate": 0.00029999762390683096,
+      "loss": 0.3705,
+      "step": 13780
+    },
+    {
+      "epoch": 0.0019331177699082392,
+      "grad_norm": 0.486328125,
+      "learning_rate": 0.0002999976201866391,
+      "loss": 0.4369,
+      "step": 13790
+    },
+    {
+      "epoch": 0.0019345195957022265,
+      "grad_norm": 0.609375,
+      "learning_rate": 0.00029999761646353726,
+      "loss": 0.4915,
+      "step": 13800
+    },
+    {
+      "epoch": 0.0019359214214962135,
+      "grad_norm": 0.984375,
+      "learning_rate": 0.0002999976127375254,
+      "loss": 0.4433,
+      "step": 13810
+    },
+    {
+      "epoch": 0.0019373232472902006,
+      "grad_norm": 0.443359375,
+      "learning_rate": 0.00029999760900860356,
+      "loss": 0.465,
+      "step": 13820
+    },
+    {
+      "epoch": 0.0019387250730841879,
+      "grad_norm": 0.9140625,
+      "learning_rate": 0.0002999976052767717,
+      "loss": 0.4211,
+      "step": 13830
+    },
+    {
+      "epoch": 0.001940126898878175,
+      "grad_norm": 0.357421875,
+      "learning_rate": 0.00029999760154202987,
+      "loss": 0.413,
+      "step": 13840
+    },
+    {
+      "epoch": 0.001941528724672162,
+      "grad_norm": 0.490234375,
+      "learning_rate": 0.000299997597804378,
+      "loss": 0.3958,
+      "step": 13850
+    },
+    {
+      "epoch": 0.001942930550466149,
+      "grad_norm": 0.89453125,
+      "learning_rate": 0.00029999759406381613,
+      "loss": 0.3903,
+      "step": 13860
+    },
+    {
+      "epoch": 0.0019443323762601363,
+      "grad_norm": 0.33984375,
+      "learning_rate": 0.00029999759032034424,
+      "loss": 0.3797,
+      "step": 13870
+    },
+    {
+      "epoch": 0.0019457342020541234,
+      "grad_norm": 0.66796875,
+      "learning_rate": 0.00029999758657396234,
+      "loss": 0.5,
+      "step": 13880
+    },
+    {
+      "epoch": 0.0019471360278481105,
+      "grad_norm": 0.69140625,
+      "learning_rate": 0.0002999975828246705,
+      "loss": 0.4632,
+      "step": 13890
+    },
+    {
+      "epoch": 0.0019485378536420975,
+      "grad_norm": 1.09375,
+      "learning_rate": 0.00029999757907246856,
+      "loss": 0.4769,
+      "step": 13900
+    },
+    {
+      "epoch": 0.0019499396794360848,
+      "grad_norm": 0.1591796875,
+      "learning_rate": 0.0002999975753173566,
+      "loss": 0.409,
+      "step": 13910
+    },
+    {
+      "epoch": 0.0019513415052300719,
+      "grad_norm": 0.625,
+      "learning_rate": 0.0002999975715593347,
+      "loss": 0.4453,
+      "step": 13920
+    },
+    {
+      "epoch": 0.001952743331024059,
+      "grad_norm": 0.77734375,
+      "learning_rate": 0.00029999756779840284,
+      "loss": 0.4729,
+      "step": 13930
+    },
+    {
+      "epoch": 0.001954145156818046,
+      "grad_norm": 1.0390625,
+      "learning_rate": 0.0002999975640345609,
+      "loss": 0.4493,
+      "step": 13940
+    },
+    {
+      "epoch": 0.0019555469826120333,
+      "grad_norm": 0.341796875,
+      "learning_rate": 0.00029999756026780896,
+      "loss": 0.3759,
+      "step": 13950
+    },
+    {
+      "epoch": 0.0019569488084060205,
+      "grad_norm": 1.171875,
+      "learning_rate": 0.0002999975564981471,
+      "loss": 0.4879,
+      "step": 13960
+    },
+    {
+      "epoch": 0.0019583506342000074,
+      "grad_norm": 0.9140625,
+      "learning_rate": 0.0002999975527255751,
+      "loss": 0.4767,
+      "step": 13970
+    },
+    {
+      "epoch": 0.0019597524599939947,
+      "grad_norm": 0.28515625,
+      "learning_rate": 0.00029999754895009315,
+      "loss": 0.4501,
+      "step": 13980
+    },
+    {
+      "epoch": 0.0019611542857879815,
+      "grad_norm": 0.7109375,
+      "learning_rate": 0.0002999975451717012,
+      "loss": 0.4216,
+      "step": 13990
+    },
+    {
+      "epoch": 0.0019625561115819688,
+      "grad_norm": 0.345703125,
+      "learning_rate": 0.0002999975413903993,
+      "loss": 0.3885,
+      "step": 14000
+    },
+    {
+      "epoch": 0.0019625561115819688,
+      "eval_loss": 0.17219418287277222,
+      "eval_runtime": 112.3232,
+      "eval_samples_per_second": 203.369,
+      "eval_steps_per_second": 1.594,
+      "step": 14000
+    },
+    {
+      "epoch": 0.001963957937375956,
+      "grad_norm": 0.56640625,
+      "learning_rate": 0.0002999975376061873,
+      "loss": 0.412,
+      "step": 14010
+    },
+    {
+      "epoch": 0.001965359763169943,
+      "grad_norm": 0.337890625,
+      "learning_rate": 0.0002999975338190653,
+      "loss": 0.3503,
+      "step": 14020
+    },
+    {
+      "epoch": 0.00196676158896393,
+      "grad_norm": 0.83984375,
+      "learning_rate": 0.00029999753002903334,
+      "loss": 0.3654,
+      "step": 14030
+    },
+    {
+      "epoch": 0.0019681634147579175,
+      "grad_norm": 0.6484375,
+      "learning_rate": 0.0002999975262360914,
+      "loss": 0.5271,
+      "step": 14040
+    },
+    {
+      "epoch": 0.0019695652405519043,
+      "grad_norm": 0.6171875,
+      "learning_rate": 0.00029999752244023943,
+      "loss": 0.3611,
+      "step": 14050
+    },
+    {
+      "epoch": 0.0019709670663458916,
+      "grad_norm": 0.8671875,
+      "learning_rate": 0.0002999975186414774,
+      "loss": 0.475,
+      "step": 14060
+    },
+    {
+      "epoch": 0.001972368892139879,
+      "grad_norm": 0.6015625,
+      "learning_rate": 0.0002999975148398054,
+      "loss": 0.4544,
+      "step": 14070
+    },
+    {
+      "epoch": 0.0019737707179338657,
+      "grad_norm": 0.75390625,
+      "learning_rate": 0.00029999751103522345,
+      "loss": 0.371,
+      "step": 14080
+    },
+    {
+      "epoch": 0.001975172543727853,
+      "grad_norm": 0.2470703125,
+      "learning_rate": 0.0002999975072277314,
+      "loss": 0.4124,
+      "step": 14090
+    },
+    {
+      "epoch": 0.00197657436952184,
+      "grad_norm": 0.298828125,
+      "learning_rate": 0.00029999750341732945,
+      "loss": 0.4765,
+      "step": 14100
+    },
+    {
+      "epoch": 0.001977976195315827,
+      "grad_norm": 0.68359375,
+      "learning_rate": 0.0002999974996040174,
+      "loss": 0.4405,
+      "step": 14110
+    },
+    {
+      "epoch": 0.0019793780211098144,
+      "grad_norm": 0.69140625,
+      "learning_rate": 0.00029999749578779546,
+      "loss": 0.5118,
+      "step": 14120
+    },
+    {
+      "epoch": 0.001980779846903801,
+      "grad_norm": 0.53125,
+      "learning_rate": 0.00029999749196866343,
+      "loss": 0.4831,
+      "step": 14130
+    },
+    {
+      "epoch": 0.0019821816726977885,
+      "grad_norm": 0.6328125,
+      "learning_rate": 0.00029999748814662136,
+      "loss": 0.5109,
+      "step": 14140
+    },
+    {
+      "epoch": 0.0019835834984917758,
+      "grad_norm": 0.58984375,
+      "learning_rate": 0.00029999748432166934,
+      "loss": 0.3219,
+      "step": 14150
+    },
+    {
+      "epoch": 0.0019849853242857626,
+      "grad_norm": 0.515625,
+      "learning_rate": 0.0002999974804938073,
+      "loss": 0.3513,
+      "step": 14160
+    },
+    {
+      "epoch": 0.00198638715007975,
+      "grad_norm": 0.90625,
+      "learning_rate": 0.0002999974766630353,
+      "loss": 0.4944,
+      "step": 14170
+    },
+    {
+      "epoch": 0.001987788975873737,
+      "grad_norm": 0.91015625,
+      "learning_rate": 0.0002999974728293533,
+      "loss": 0.4509,
+      "step": 14180
+    },
+    {
+      "epoch": 0.001989190801667724,
+      "grad_norm": 1.1171875,
+      "learning_rate": 0.0002999974689927613,
+      "loss": 0.4779,
+      "step": 14190
+    },
+    {
+      "epoch": 0.0019905926274617113,
+      "grad_norm": 0.578125,
+      "learning_rate": 0.0002999974651532592,
+      "loss": 0.4435,
+      "step": 14200
+    },
+    {
+      "epoch": 0.0019919944532556986,
+      "grad_norm": 0.62109375,
+      "learning_rate": 0.00029999746131084714,
+      "loss": 0.4092,
+      "step": 14210
+    },
+    {
+      "epoch": 0.0019933962790496854,
+      "grad_norm": 1.1015625,
+      "learning_rate": 0.0002999974574655251,
+      "loss": 0.4802,
+      "step": 14220
+    },
+    {
+      "epoch": 0.0019947981048436727,
+      "grad_norm": 0.78125,
+      "learning_rate": 0.000299997453617293,
+      "loss": 0.4033,
+      "step": 14230
+    },
+    {
+      "epoch": 0.0019961999306376595,
+      "grad_norm": 0.494140625,
+      "learning_rate": 0.00029999744976615095,
+      "loss": 0.3894,
+      "step": 14240
+    },
+    {
+      "epoch": 0.001997601756431647,
+      "grad_norm": 0.671875,
+      "learning_rate": 0.00029999744591209895,
+      "loss": 0.35,
+      "step": 14250
+    },
+    {
+      "epoch": 0.001999003582225634,
+      "grad_norm": 0.1357421875,
+      "learning_rate": 0.00029999744205513683,
+      "loss": 0.4277,
+      "step": 14260
+    },
+    {
+      "epoch": 0.002000405408019621,
+      "grad_norm": 0.193359375,
+      "learning_rate": 0.0002999974381952648,
+      "loss": 0.3722,
+      "step": 14270
+    },
+    {
+      "epoch": 0.002001807233813608,
+      "grad_norm": 0.83203125,
+      "learning_rate": 0.0002999974343324827,
+      "loss": 0.4317,
+      "step": 14280
+    },
+    {
+      "epoch": 0.0020032090596075955,
+      "grad_norm": 1.1171875,
+      "learning_rate": 0.0002999974304667906,
+      "loss": 0.4286,
+      "step": 14290
+    },
+    {
+      "epoch": 0.0020046108854015823,
+      "grad_norm": 0.8359375,
+      "learning_rate": 0.0002999974265981886,
+      "loss": 0.4539,
+      "step": 14300
+    },
+    {
+      "epoch": 0.0020060127111955696,
+      "grad_norm": 0.7265625,
+      "learning_rate": 0.0002999974227266765,
+      "loss": 0.4675,
+      "step": 14310
+    },
+    {
+      "epoch": 0.002007414536989557,
+      "grad_norm": 0.23046875,
+      "learning_rate": 0.00029999741885225446,
+      "loss": 0.4397,
+      "step": 14320
+    },
+    {
+      "epoch": 0.0020088163627835437,
+      "grad_norm": 0.890625,
+      "learning_rate": 0.0002999974149749223,
+      "loss": 0.4109,
+      "step": 14330
+    },
+    {
+      "epoch": 0.002010218188577531,
+      "grad_norm": 0.66015625,
+      "learning_rate": 0.00029999741109468025,
+      "loss": 0.3804,
+      "step": 14340
+    },
+    {
+      "epoch": 0.002011620014371518,
+      "grad_norm": 0.435546875,
+      "learning_rate": 0.00029999740721152815,
+      "loss": 0.3754,
+      "step": 14350
+    },
+    {
+      "epoch": 0.002013021840165505,
+      "grad_norm": 0.234375,
+      "learning_rate": 0.00029999740332546605,
+      "loss": 0.3823,
+      "step": 14360
+    },
+    {
+      "epoch": 0.0020144236659594924,
+      "grad_norm": 0.765625,
+      "learning_rate": 0.00029999739943649395,
+      "loss": 0.3991,
+      "step": 14370
+    },
+    {
+      "epoch": 0.0020158254917534793,
+      "grad_norm": 0.8671875,
+      "learning_rate": 0.0002999973955446119,
+      "loss": 0.4227,
+      "step": 14380
+    },
+    {
+      "epoch": 0.0020172273175474665,
+      "grad_norm": 1.1640625,
+      "learning_rate": 0.0002999973916498198,
+      "loss": 0.3951,
+      "step": 14390
+    },
+    {
+      "epoch": 0.002018629143341454,
+      "grad_norm": 0.55078125,
+      "learning_rate": 0.0002999973877521177,
+      "loss": 0.4999,
+      "step": 14400
+    },
+    {
+      "epoch": 0.0020200309691354407,
+      "grad_norm": 0.89453125,
+      "learning_rate": 0.0002999973838515056,
+      "loss": 0.4185,
+      "step": 14410
+    },
+    {
+      "epoch": 0.002021432794929428,
+      "grad_norm": 0.53515625,
+      "learning_rate": 0.0002999973799479835,
+      "loss": 0.4898,
+      "step": 14420
+    },
+    {
+      "epoch": 0.002022834620723415,
+      "grad_norm": 0.388671875,
+      "learning_rate": 0.0002999973760415514,
+      "loss": 0.4813,
+      "step": 14430
+    },
+    {
+      "epoch": 0.002024236446517402,
+      "grad_norm": 0.5390625,
+      "learning_rate": 0.00029999737213220925,
+      "loss": 0.4236,
+      "step": 14440
+    },
+    {
+      "epoch": 0.0020256382723113893,
+      "grad_norm": 0.3125,
+      "learning_rate": 0.00029999736821995716,
+      "loss": 0.4339,
+      "step": 14450
+    },
+    {
+      "epoch": 0.002027040098105376,
+      "grad_norm": 0.6796875,
+      "learning_rate": 0.000299997364304795,
+      "loss": 0.4094,
+      "step": 14460
+    },
+    {
+      "epoch": 0.0020284419238993634,
+      "grad_norm": 0.19921875,
+      "learning_rate": 0.00029999736038672294,
+      "loss": 0.4625,
+      "step": 14470
+    },
+    {
+      "epoch": 0.0020298437496933507,
+      "grad_norm": 0.66796875,
+      "learning_rate": 0.0002999973564657408,
+      "loss": 0.33,
+      "step": 14480
+    },
+    {
+      "epoch": 0.0020312455754873376,
+      "grad_norm": 0.5,
+      "learning_rate": 0.00029999735254184867,
+      "loss": 0.4033,
+      "step": 14490
+    },
+    {
+      "epoch": 0.002032647401281325,
+      "grad_norm": 0.64453125,
+      "learning_rate": 0.0002999973486150466,
+      "loss": 0.4106,
+      "step": 14500
+    },
+    {
+      "epoch": 0.002032647401281325,
+      "eval_loss": 0.1713680624961853,
+      "eval_runtime": 112.5849,
+      "eval_samples_per_second": 202.896,
+      "eval_steps_per_second": 1.59,
+      "step": 14500
+    },
+    {
+      "epoch": 0.002034049227075312,
+      "grad_norm": 0.84375,
+      "learning_rate": 0.00029999734468533445,
+      "loss": 0.3933,
+      "step": 14510
+    },
+    {
+      "epoch": 0.002035451052869299,
+      "grad_norm": 0.39453125,
+      "learning_rate": 0.0002999973407527123,
+      "loss": 0.4153,
+      "step": 14520
+    },
+    {
+      "epoch": 0.0020368528786632862,
+      "grad_norm": 0.234375,
+      "learning_rate": 0.0002999973368171802,
+      "loss": 0.4338,
+      "step": 14530
+    },
+    {
+      "epoch": 0.0020382547044572735,
+      "grad_norm": 0.470703125,
+      "learning_rate": 0.00029999733287873806,
+      "loss": 0.4352,
+      "step": 14540
+    },
+    {
+      "epoch": 0.0020396565302512604,
+      "grad_norm": 0.828125,
+      "learning_rate": 0.00029999732893738594,
+      "loss": 0.4688,
+      "step": 14550
+    },
+    {
+      "epoch": 0.0020410583560452476,
+      "grad_norm": 0.35546875,
+      "learning_rate": 0.0002999973249931238,
+      "loss": 0.436,
+      "step": 14560
+    },
+    {
+      "epoch": 0.0020424601818392345,
+      "grad_norm": 0.65625,
+      "learning_rate": 0.0002999973210459517,
+      "loss": 0.3661,
+      "step": 14570
+    },
+    {
+      "epoch": 0.0020438620076332218,
+      "grad_norm": 0.482421875,
+      "learning_rate": 0.00029999731709586956,
+      "loss": 0.3806,
+      "step": 14580
+    },
+    {
+      "epoch": 0.002045263833427209,
+      "grad_norm": 0.66015625,
+      "learning_rate": 0.00029999731314287744,
+      "loss": 0.4362,
+      "step": 14590
+    },
+    {
+      "epoch": 0.002046665659221196,
+      "grad_norm": 0.75390625,
+      "learning_rate": 0.0002999973091869753,
+      "loss": 0.3717,
+      "step": 14600
+    },
+    {
+      "epoch": 0.002048067485015183,
+      "grad_norm": 0.412109375,
+      "learning_rate": 0.0002999973052281632,
+      "loss": 0.4599,
+      "step": 14610
+    },
+    {
+      "epoch": 0.0020494693108091704,
+      "grad_norm": 0.291015625,
+      "learning_rate": 0.0002999973012664411,
+      "loss": 0.4119,
+      "step": 14620
+    },
+    {
+      "epoch": 0.0020508711366031573,
+      "grad_norm": 0.376953125,
+      "learning_rate": 0.0002999972973018089,
+      "loss": 0.3842,
+      "step": 14630
+    },
+    {
+      "epoch": 0.0020522729623971446,
+      "grad_norm": 0.53515625,
+      "learning_rate": 0.00029999729333426685,
+      "loss": 0.5399,
+      "step": 14640
+    },
+    {
+      "epoch": 0.002053674788191132,
+      "grad_norm": 0.78515625,
+      "learning_rate": 0.0002999972893638147,
+      "loss": 0.3464,
+      "step": 14650
+    },
+    {
+      "epoch": 0.0020550766139851187,
+      "grad_norm": 0.4609375,
+      "learning_rate": 0.00029999728539045257,
+      "loss": 0.466,
+      "step": 14660
+    },
+    {
+      "epoch": 0.002056478439779106,
+      "grad_norm": 0.455078125,
+      "learning_rate": 0.0002999972814141804,
+      "loss": 0.4817,
+      "step": 14670
+    },
+    {
+      "epoch": 0.002057880265573093,
+      "grad_norm": 0.32421875,
+      "learning_rate": 0.00029999727743499824,
+      "loss": 0.3673,
+      "step": 14680
+    },
+    {
+      "epoch": 0.00205928209136708,
+      "grad_norm": 1.0,
+      "learning_rate": 0.00029999727345290613,
+      "loss": 0.4989,
+      "step": 14690
+    },
+    {
+      "epoch": 0.0020606839171610674,
+      "grad_norm": 0.72265625,
+      "learning_rate": 0.000299997269467904,
+      "loss": 0.4511,
+      "step": 14700
+    },
+    {
+      "epoch": 0.002062085742955054,
+      "grad_norm": 0.78515625,
+      "learning_rate": 0.00029999726547999187,
+      "loss": 0.5199,
+      "step": 14710
+    },
+    {
+      "epoch": 0.0020634875687490415,
+      "grad_norm": 0.9140625,
+      "learning_rate": 0.00029999726148916976,
+      "loss": 0.4247,
+      "step": 14720
+    },
+    {
+      "epoch": 0.0020648893945430288,
+      "grad_norm": 0.2158203125,
+      "learning_rate": 0.0002999972574954376,
+      "loss": 0.3859,
+      "step": 14730
+    },
+    {
+      "epoch": 0.0020662912203370156,
+      "grad_norm": 0.6171875,
+      "learning_rate": 0.0002999972534987955,
+      "loss": 0.4503,
+      "step": 14740
+    },
+    {
+      "epoch": 0.002067693046131003,
+      "grad_norm": 0.73828125,
+      "learning_rate": 0.00029999724949924335,
+      "loss": 0.4461,
+      "step": 14750
+    },
+    {
+      "epoch": 0.00206909487192499,
+      "grad_norm": 0.8671875,
+      "learning_rate": 0.00029999724549678125,
+      "loss": 0.4595,
+      "step": 14760
+    },
+    {
+      "epoch": 0.002070496697718977,
+      "grad_norm": 0.59765625,
+      "learning_rate": 0.0002999972414914091,
+      "loss": 0.325,
+      "step": 14770
+    },
+    {
+      "epoch": 0.0020718985235129643,
+      "grad_norm": 0.62890625,
+      "learning_rate": 0.000299997237483127,
+      "loss": 0.384,
+      "step": 14780
+    },
+    {
+      "epoch": 0.0020733003493069516,
+      "grad_norm": 0.64453125,
+      "learning_rate": 0.00029999723347193485,
+      "loss": 0.3865,
+      "step": 14790
+    },
+    {
+      "epoch": 0.0020747021751009384,
+      "grad_norm": 0.318359375,
+      "learning_rate": 0.00029999722945783276,
+      "loss": 0.4034,
+      "step": 14800
+    },
+    {
+      "epoch": 0.0020761040008949257,
+      "grad_norm": 0.373046875,
+      "learning_rate": 0.0002999972254408206,
+      "loss": 0.3206,
+      "step": 14810
+    },
+    {
+      "epoch": 0.0020775058266889125,
+      "grad_norm": 0.73828125,
+      "learning_rate": 0.0002999972214208985,
+      "loss": 0.5055,
+      "step": 14820
+    },
+    {
+      "epoch": 0.0020789076524829,
+      "grad_norm": 1.09375,
+      "learning_rate": 0.0002999972173980664,
+      "loss": 0.5426,
+      "step": 14830
+    },
+    {
+      "epoch": 0.002080309478276887,
+      "grad_norm": 0.416015625,
+      "learning_rate": 0.00029999721337232424,
+      "loss": 0.5022,
+      "step": 14840
+    },
+    {
+      "epoch": 0.002081711304070874,
+      "grad_norm": 1.0390625,
+      "learning_rate": 0.0002999972093436721,
+      "loss": 0.4795,
+      "step": 14850
+    },
+    {
+      "epoch": 0.002083113129864861,
+      "grad_norm": 0.73828125,
+      "learning_rate": 0.00029999720531211,
+      "loss": 0.5243,
+      "step": 14860
+    },
+    {
+      "epoch": 0.0020845149556588485,
+      "grad_norm": 0.51953125,
+      "learning_rate": 0.00029999720127763793,
+      "loss": 0.4725,
+      "step": 14870
+    },
+    {
+      "epoch": 0.0020859167814528353,
+      "grad_norm": 0.73046875,
+      "learning_rate": 0.00029999719724025574,
+      "loss": 0.4518,
+      "step": 14880
+    },
+    {
+      "epoch": 0.0020873186072468226,
+      "grad_norm": 0.78515625,
+      "learning_rate": 0.00029999719319996366,
+      "loss": 0.3732,
+      "step": 14890
+    },
+    {
+      "epoch": 0.00208872043304081,
+      "grad_norm": 0.67578125,
+      "learning_rate": 0.0002999971891567615,
+      "loss": 0.5069,
+      "step": 14900
+    },
+    {
+      "epoch": 0.0020901222588347967,
+      "grad_norm": 0.486328125,
+      "learning_rate": 0.00029999718511064945,
+      "loss": 0.4604,
+      "step": 14910
+    },
+    {
+      "epoch": 0.002091524084628784,
+      "grad_norm": 0.5078125,
+      "learning_rate": 0.0002999971810616273,
+      "loss": 0.4016,
+      "step": 14920
+    },
+    {
+      "epoch": 0.002092925910422771,
+      "grad_norm": 0.2080078125,
+      "learning_rate": 0.00029999717700969524,
+      "loss": 0.3554,
+      "step": 14930
+    },
+    {
+      "epoch": 0.002094327736216758,
+      "grad_norm": 0.703125,
+      "learning_rate": 0.0002999971729548531,
+      "loss": 0.3677,
+      "step": 14940
+    },
+    {
+      "epoch": 0.0020957295620107454,
+      "grad_norm": 0.703125,
+      "learning_rate": 0.000299997168897101,
+      "loss": 0.4108,
+      "step": 14950
+    },
+    {
+      "epoch": 0.0020971313878047322,
+      "grad_norm": 0.82421875,
+      "learning_rate": 0.0002999971648364389,
+      "loss": 0.5519,
+      "step": 14960
+    },
+    {
+      "epoch": 0.0020985332135987195,
+      "grad_norm": 0.296875,
+      "learning_rate": 0.0002999971607728668,
+      "loss": 0.4011,
+      "step": 14970
+    },
+    {
+      "epoch": 0.002099935039392707,
+      "grad_norm": 0.20703125,
+      "learning_rate": 0.0002999971567063847,
+      "loss": 0.2955,
+      "step": 14980
+    },
+    {
+      "epoch": 0.0021013368651866936,
+      "grad_norm": 0.859375,
+      "learning_rate": 0.0002999971526369926,
+      "loss": 0.4985,
+      "step": 14990
+    },
+    {
+      "epoch": 0.002102738690980681,
+      "grad_norm": 0.87890625,
+      "learning_rate": 0.00029999714856469053,
+      "loss": 0.3839,
+      "step": 15000
+    },
+    {
+      "epoch": 0.002102738690980681,
+      "eval_loss": 0.17203140258789062,
+      "eval_runtime": 114.2354,
+      "eval_samples_per_second": 199.964,
+      "eval_steps_per_second": 1.567,
+      "step": 15000
+    }
+  ],
+  "logging_steps": 10,
+  "max_steps": 7133554,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 1,
+  "save_steps": 5000,
+  "stateful_callbacks": {
+    "TrainerControl": {
+      "args": {
+        "should_epoch_stop": false,
+        "should_evaluate": false,
+        "should_log": false,
+        "should_save": true,
+        "should_training_stop": false
+      },
+      "attributes": {}
+    }
+  },
+  "total_flos": 2.0473171539266765e+17,
+  "train_batch_size": 16,
+  "trial_name": null,
+  "trial_params": null
+}