{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 20.6252014179826,
  "eval_steps": 100,
  "global_step": 8000,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.12890750886239125,
      "grad_norm": 5.428058115945689,
      "learning_rate": 0.0001,
      "loss": 3.7217,
      "step": 50
    },
    {
      "epoch": 0.2578150177247825,
      "grad_norm": 3.6736343068699004,
      "learning_rate": 9.999743504733682e-05,
      "loss": 2.6957,
      "step": 100
    },
    {
      "epoch": 0.2578150177247825,
      "eval_loss": 2.472891092300415,
      "eval_runtime": 13.3767,
      "eval_samples_per_second": 74.757,
      "eval_steps_per_second": 2.392,
      "step": 100
    },
    {
      "epoch": 0.3867225265871737,
      "grad_norm": 3.9566460988411976,
      "learning_rate": 9.998974045516476e-05,
      "loss": 2.3982,
      "step": 150
    },
    {
      "epoch": 0.515630035449565,
      "grad_norm": 1.91000275173515,
      "learning_rate": 9.997691702090861e-05,
      "loss": 2.188,
      "step": 200
    },
    {
      "epoch": 0.515630035449565,
      "eval_loss": 2.072312593460083,
      "eval_runtime": 13.1385,
      "eval_samples_per_second": 76.112,
      "eval_steps_per_second": 2.436,
      "step": 200
    },
    {
      "epoch": 0.6445375443119562,
      "grad_norm": 3.7326458687273494,
      "learning_rate": 9.9958966073518e-05,
      "loss": 2.058,
      "step": 250
    },
    {
      "epoch": 0.7734450531743474,
      "grad_norm": 1.1877292780347308,
      "learning_rate": 9.993588947332947e-05,
      "loss": 2.0499,
      "step": 300
    },
    {
      "epoch": 0.7734450531743474,
      "eval_loss": 1.9889113903045654,
      "eval_runtime": 13.0859,
      "eval_samples_per_second": 76.418,
      "eval_steps_per_second": 2.445,
      "step": 300
    },
    {
      "epoch": 0.9023525620367386,
      "grad_norm": 2.896903582467451,
      "learning_rate": 9.990768961187381e-05,
      "loss": 1.9975,
      "step": 350
    },
    {
      "epoch": 1.03126007089913,
      "grad_norm": 1.1169559646807505,
      "learning_rate": 9.98743694116282e-05,
      "loss": 1.9593,
      "step": 400
    },
    {
      "epoch": 1.03126007089913,
      "eval_loss": 1.9284899234771729,
      "eval_runtime": 13.1076,
      "eval_samples_per_second": 76.292,
      "eval_steps_per_second": 2.441,
      "step": 400
    },
    {
      "epoch": 1.1601675797615212,
      "grad_norm": 0.9849310297267434,
      "learning_rate": 9.983593232571328e-05,
      "loss": 1.9186,
      "step": 450
    },
    {
      "epoch": 1.2890750886239124,
      "grad_norm": 0.9560509903560709,
      "learning_rate": 9.979238233753537e-05,
      "loss": 1.921,
      "step": 500
    },
    {
      "epoch": 1.2890750886239124,
      "eval_loss": 1.899874210357666,
      "eval_runtime": 13.1392,
      "eval_samples_per_second": 76.108,
      "eval_steps_per_second": 2.435,
      "step": 500
    },
    {
      "epoch": 1.4179825974863036,
      "grad_norm": 1.3741796182310861,
      "learning_rate": 9.974372396037357e-05,
      "loss": 1.9246,
      "step": 550
    },
    {
      "epoch": 1.5468901063486948,
      "grad_norm": 1.718108772048462,
      "learning_rate": 9.96899622369121e-05,
      "loss": 1.8882,
      "step": 600
    },
    {
      "epoch": 1.5468901063486948,
      "eval_loss": 1.8836181163787842,
      "eval_runtime": 13.1533,
      "eval_samples_per_second": 76.026,
      "eval_steps_per_second": 2.433,
      "step": 600
    },
    {
      "epoch": 1.675797615211086,
      "grad_norm": 1.3793562879868584,
      "learning_rate": 9.963110273871768e-05,
      "loss": 1.8823,
      "step": 650
    },
    {
      "epoch": 1.8047051240734773,
      "grad_norm": 1.7941360700066475,
      "learning_rate": 9.956715156566203e-05,
      "loss": 1.8786,
      "step": 700
    },
    {
      "epoch": 1.8047051240734773,
      "eval_loss": 1.874005913734436,
      "eval_runtime": 13.1279,
      "eval_samples_per_second": 76.174,
      "eval_steps_per_second": 2.438,
      "step": 700
    },
    {
      "epoch": 1.9336126329358685,
      "grad_norm": 1.2268118375629977,
      "learning_rate": 9.94981153452899e-05,
      "loss": 1.8764,
      "step": 750
    },
    {
      "epoch": 2.06252014179826,
      "grad_norm": 0.8694323872512157,
      "learning_rate": 9.94240012321321e-05,
      "loss": 1.8327,
      "step": 800
    },
    {
      "epoch": 2.06252014179826,
      "eval_loss": 1.8573421239852905,
      "eval_runtime": 13.1123,
      "eval_samples_per_second": 76.264,
      "eval_steps_per_second": 2.44,
      "step": 800
    },
    {
      "epoch": 2.191427650660651,
      "grad_norm": 1.6779131964119156,
      "learning_rate": 9.934481690696406e-05,
      "loss": 1.7937,
      "step": 850
    },
    {
      "epoch": 2.3203351595230424,
      "grad_norm": 1.2570407836555852,
      "learning_rate": 9.92605705760098e-05,
      "loss": 1.8047,
      "step": 900
    },
    {
      "epoch": 2.3203351595230424,
      "eval_loss": 1.8506410121917725,
      "eval_runtime": 13.1515,
      "eval_samples_per_second": 76.037,
      "eval_steps_per_second": 2.433,
      "step": 900
    },
    {
      "epoch": 2.4492426683854336,
      "grad_norm": 0.9012297719032618,
      "learning_rate": 9.917127097009165e-05,
      "loss": 1.8028,
      "step": 950
    },
    {
      "epoch": 2.578150177247825,
      "grad_norm": 2.071809054799254,
      "learning_rate": 9.907692734372522e-05,
      "loss": 1.8001,
      "step": 1000
    },
    {
      "epoch": 2.578150177247825,
      "eval_loss": 1.8604120016098022,
      "eval_runtime": 13.1616,
      "eval_samples_per_second": 75.978,
      "eval_steps_per_second": 2.431,
      "step": 1000
    },
    {
      "epoch": 2.707057686110216,
      "grad_norm": 1.8799616432709254,
      "learning_rate": 9.897754947416042e-05,
      "loss": 1.8009,
      "step": 1050
    },
    {
      "epoch": 2.8359651949726072,
      "grad_norm": 0.9523078941745302,
      "learning_rate": 9.887314766036823e-05,
      "loss": 1.8016,
      "step": 1100
    },
    {
      "epoch": 2.8359651949726072,
      "eval_loss": 1.8416086435317993,
      "eval_runtime": 13.1272,
      "eval_samples_per_second": 76.178,
      "eval_steps_per_second": 2.438,
      "step": 1100
    },
    {
      "epoch": 2.9648727038349985,
      "grad_norm": 0.7548354411956247,
      "learning_rate": 9.87637327219733e-05,
      "loss": 1.7941,
      "step": 1150
    },
    {
      "epoch": 3.0937802126973897,
      "grad_norm": 1.603556668939176,
      "learning_rate": 9.86493159981327e-05,
      "loss": 1.7221,
      "step": 1200
    },
    {
      "epoch": 3.0937802126973897,
      "eval_loss": 1.8572254180908203,
      "eval_runtime": 13.1348,
      "eval_samples_per_second": 76.134,
      "eval_steps_per_second": 2.436,
      "step": 1200
    },
    {
      "epoch": 3.222687721559781,
      "grad_norm": 1.1914026184365651,
      "learning_rate": 9.85299093463608e-05,
      "loss": 1.6919,
      "step": 1250
    },
    {
      "epoch": 3.351595230422172,
      "grad_norm": 1.108713427331391,
      "learning_rate": 9.840552514130043e-05,
      "loss": 1.6979,
      "step": 1300
    },
    {
      "epoch": 3.351595230422172,
      "eval_loss": 1.8494083881378174,
      "eval_runtime": 13.1546,
      "eval_samples_per_second": 76.019,
      "eval_steps_per_second": 2.433,
      "step": 1300
    },
    {
      "epoch": 3.4805027392845633,
      "grad_norm": 1.188923474560524,
      "learning_rate": 9.827617627344035e-05,
      "loss": 1.7012,
      "step": 1350
    },
    {
      "epoch": 3.6094102481469545,
      "grad_norm": 1.1069030836862073,
      "learning_rate": 9.814187614777952e-05,
      "loss": 1.7117,
      "step": 1400
    },
    {
      "epoch": 3.6094102481469545,
      "eval_loss": 1.8396939039230347,
      "eval_runtime": 13.1058,
      "eval_samples_per_second": 76.302,
      "eval_steps_per_second": 2.442,
      "step": 1400
    },
    {
      "epoch": 3.7383177570093458,
      "grad_norm": 1.4927257455688545,
      "learning_rate": 9.800263868243771e-05,
      "loss": 1.7115,
      "step": 1450
    },
    {
      "epoch": 3.867225265871737,
      "grad_norm": 0.9082276480595701,
      "learning_rate": 9.785847830721322e-05,
      "loss": 1.7181,
      "step": 1500
    },
    {
      "epoch": 3.867225265871737,
      "eval_loss": 1.8317677974700928,
      "eval_runtime": 13.1459,
      "eval_samples_per_second": 76.069,
      "eval_steps_per_second": 2.434,
      "step": 1500
    },
    {
      "epoch": 3.996132774734128,
      "grad_norm": 1.38818621871337,
      "learning_rate": 9.770940996208739e-05,
      "loss": 1.7195,
      "step": 1550
    },
    {
      "epoch": 4.12504028359652,
      "grad_norm": 1.1154867829165278,
      "learning_rate": 9.755544909567632e-05,
      "loss": 1.5548,
      "step": 1600
    },
    {
      "epoch": 4.12504028359652,
      "eval_loss": 1.8801182508468628,
      "eval_runtime": 13.1664,
      "eval_samples_per_second": 75.951,
      "eval_steps_per_second": 2.43,
      "step": 1600
    },
    {
      "epoch": 4.253947792458911,
      "grad_norm": 1.741656038399529,
      "learning_rate": 9.73966116636299e-05,
      "loss": 1.5498,
      "step": 1650
    },
    {
      "epoch": 4.382855301321302,
      "grad_norm": 0.9507921869813947,
      "learning_rate": 9.723291412697821e-05,
      "loss": 1.5728,
      "step": 1700
    },
    {
      "epoch": 4.382855301321302,
      "eval_loss": 1.8739728927612305,
      "eval_runtime": 13.1362,
      "eval_samples_per_second": 76.126,
      "eval_steps_per_second": 2.436,
      "step": 1700
    },
    {
      "epoch": 4.5117628101836935,
      "grad_norm": 1.1876157181384903,
      "learning_rate": 9.706437345042558e-05,
      "loss": 1.5896,
      "step": 1750
    },
    {
      "epoch": 4.640670319046085,
      "grad_norm": 1.365051813978068,
      "learning_rate": 9.689100710059251e-05,
      "loss": 1.5998,
      "step": 1800
    },
    {
      "epoch": 4.640670319046085,
      "eval_loss": 1.8598759174346924,
      "eval_runtime": 13.1287,
      "eval_samples_per_second": 76.169,
      "eval_steps_per_second": 2.437,
      "step": 1800
    },
    {
      "epoch": 4.769577827908476,
      "grad_norm": 1.3205776295240563,
      "learning_rate": 9.67128330442055e-05,
      "loss": 1.5974,
      "step": 1850
    },
    {
      "epoch": 4.898485336770867,
      "grad_norm": 1.4453793304970697,
      "learning_rate": 9.652986974623506e-05,
      "loss": 1.6127,
      "step": 1900
    },
    {
      "epoch": 4.898485336770867,
      "eval_loss": 1.8465975522994995,
      "eval_runtime": 13.1264,
      "eval_samples_per_second": 76.182,
      "eval_steps_per_second": 2.438,
      "step": 1900
    },
    {
      "epoch": 5.027392845633258,
      "grad_norm": 2.3043609697161633,
      "learning_rate": 9.634213616798214e-05,
      "loss": 1.5733,
      "step": 1950
    },
    {
      "epoch": 5.15630035449565,
      "grad_norm": 1.1111652873863385,
      "learning_rate": 9.614965176511308e-05,
      "loss": 1.3923,
      "step": 2000
    },
    {
      "epoch": 5.15630035449565,
      "eval_loss": 1.9901988506317139,
      "eval_runtime": 13.1434,
      "eval_samples_per_second": 76.084,
      "eval_steps_per_second": 2.435,
      "step": 2000
    },
    {
      "epoch": 5.285207863358041,
      "grad_norm": 1.8996098810923527,
      "learning_rate": 9.595243648564326e-05,
      "loss": 1.3904,
      "step": 2050
    },
    {
      "epoch": 5.414115372220432,
      "grad_norm": 0.8248818797353966,
      "learning_rate": 9.575051076786994e-05,
      "loss": 1.413,
      "step": 2100
    },
    {
      "epoch": 5.414115372220432,
      "eval_loss": 1.941409945487976,
      "eval_runtime": 13.1615,
      "eval_samples_per_second": 75.979,
      "eval_steps_per_second": 2.431,
      "step": 2100
    },
    {
      "epoch": 5.543022881082823,
      "grad_norm": 1.3796008879183341,
      "learning_rate": 9.554389553825404e-05,
      "loss": 1.4426,
      "step": 2150
    },
    {
      "epoch": 5.6719303899452145,
      "grad_norm": 1.2370414797788207,
      "learning_rate": 9.533261220925143e-05,
      "loss": 1.4441,
      "step": 2200
    },
    {
      "epoch": 5.6719303899452145,
      "eval_loss": 1.9324249029159546,
      "eval_runtime": 13.1482,
      "eval_samples_per_second": 76.056,
      "eval_steps_per_second": 2.434,
      "step": 2200
    },
    {
      "epoch": 5.800837898807606,
      "grad_norm": 0.7933476727900988,
      "learning_rate": 9.511668267709395e-05,
      "loss": 1.4578,
      "step": 2250
    },
    {
      "epoch": 5.929745407669997,
      "grad_norm": 0.9922228439520735,
      "learning_rate": 9.48961293195201e-05,
      "loss": 1.4729,
      "step": 2300
    },
    {
      "epoch": 5.929745407669997,
      "eval_loss": 1.9070981740951538,
      "eval_runtime": 13.1414,
      "eval_samples_per_second": 76.095,
      "eval_steps_per_second": 2.435,
      "step": 2300
    },
    {
      "epoch": 6.058652916532388,
      "grad_norm": 1.133228933331563,
      "learning_rate": 9.467097499345605e-05,
      "loss": 1.3338,
      "step": 2350
    },
    {
      "epoch": 6.187560425394779,
      "grad_norm": 1.1388176086742452,
      "learning_rate": 9.444124303264675e-05,
      "loss": 1.1768,
      "step": 2400
    },
    {
      "epoch": 6.187560425394779,
      "eval_loss": 2.123551607131958,
      "eval_runtime": 13.1885,
      "eval_samples_per_second": 75.824,
      "eval_steps_per_second": 2.426,
      "step": 2400
    },
    {
      "epoch": 6.316467934257171,
      "grad_norm": 1.0540569425798711,
      "learning_rate": 9.420695724523785e-05,
      "loss": 1.1952,
      "step": 2450
    },
    {
      "epoch": 6.445375443119562,
      "grad_norm": 1.0205037806931825,
      "learning_rate": 9.396814191130831e-05,
      "loss": 1.2409,
      "step": 2500
    },
    {
      "epoch": 6.445375443119562,
      "eval_loss": 2.0799171924591064,
      "eval_runtime": 13.1885,
      "eval_samples_per_second": 75.824,
      "eval_steps_per_second": 2.426,
      "step": 2500
    },
    {
      "epoch": 6.574282951981953,
      "grad_norm": 1.3287277537295186,
      "learning_rate": 9.37248217803542e-05,
      "loss": 1.2484,
      "step": 2550
    },
    {
      "epoch": 6.703190460844344,
      "grad_norm": 1.7324707650665503,
      "learning_rate": 9.347702206872368e-05,
      "loss": 1.2794,
      "step": 2600
    },
    {
      "epoch": 6.703190460844344,
      "eval_loss": 2.066530227661133,
      "eval_runtime": 13.1457,
      "eval_samples_per_second": 76.07,
      "eval_steps_per_second": 2.434,
      "step": 2600
    },
    {
      "epoch": 6.832097969706735,
      "grad_norm": 0.7821269122995969,
      "learning_rate": 9.322476845700381e-05,
      "loss": 1.2872,
      "step": 2650
    },
    {
      "epoch": 6.961005478569127,
      "grad_norm": 1.0321720718594696,
      "learning_rate": 9.296808708735924e-05,
      "loss": 1.2975,
      "step": 2700
    },
    {
      "epoch": 6.961005478569127,
      "eval_loss": 2.028914213180542,
      "eval_runtime": 13.188,
      "eval_samples_per_second": 75.827,
      "eval_steps_per_second": 2.426,
      "step": 2700
    },
    {
      "epoch": 7.089912987431518,
      "grad_norm": 0.8079357741453,
      "learning_rate": 9.270700456082274e-05,
      "loss": 1.0552,
      "step": 2750
    },
    {
      "epoch": 7.218820496293909,
      "grad_norm": 1.2220177415199573,
      "learning_rate": 9.244154793453865e-05,
      "loss": 0.9573,
      "step": 2800
    },
    {
      "epoch": 7.218820496293909,
      "eval_loss": 2.30826473236084,
      "eval_runtime": 13.14,
      "eval_samples_per_second": 76.103,
      "eval_steps_per_second": 2.435,
      "step": 2800
    },
    {
      "epoch": 7.3477280051563,
      "grad_norm": 1.0187532250301208,
      "learning_rate": 9.217174471895868e-05,
      "loss": 0.9933,
      "step": 2850
    },
    {
      "epoch": 7.4766355140186915,
      "grad_norm": 1.1142206529537857,
      "learning_rate": 9.189762287499101e-05,
      "loss": 1.0244,
      "step": 2900
    },
    {
      "epoch": 7.4766355140186915,
      "eval_loss": 2.2640790939331055,
      "eval_runtime": 13.1456,
      "eval_samples_per_second": 76.071,
      "eval_steps_per_second": 2.434,
      "step": 2900
    },
    {
      "epoch": 7.605543022881083,
      "grad_norm": 0.922470837605761,
      "learning_rate": 9.16192108111024e-05,
      "loss": 1.0455,
      "step": 2950
    },
    {
      "epoch": 7.734450531743474,
      "grad_norm": 1.4488181883296236,
      "learning_rate": 9.133653738037431e-05,
      "loss": 1.066,
      "step": 3000
    },
    {
      "epoch": 7.734450531743474,
      "eval_loss": 2.224378824234009,
      "eval_runtime": 13.1557,
      "eval_samples_per_second": 76.013,
      "eval_steps_per_second": 2.432,
      "step": 3000
    },
    {
      "epoch": 7.863358040605865,
      "grad_norm": 1.284431896370195,
      "learning_rate": 9.104963187751253e-05,
      "loss": 1.0771,
      "step": 3050
    },
    {
      "epoch": 7.992265549468256,
      "grad_norm": 1.3116181988612625,
      "learning_rate": 9.075852403581133e-05,
      "loss": 1.1098,
      "step": 3100
    },
    {
      "epoch": 7.992265549468256,
      "eval_loss": 2.198763608932495,
      "eval_runtime": 13.1699,
      "eval_samples_per_second": 75.931,
      "eval_steps_per_second": 2.43,
      "step": 3100
    },
    {
      "epoch": 8.121173058330648,
      "grad_norm": 1.185925605822691,
      "learning_rate": 9.046324402407213e-05,
      "loss": 0.733,
      "step": 3150
    },
    {
      "epoch": 8.25008056719304,
      "grad_norm": 2.353999238623636,
      "learning_rate": 9.016382244347679e-05,
      "loss": 0.7603,
      "step": 3200
    },
    {
      "epoch": 8.25008056719304,
      "eval_loss": 2.5014915466308594,
      "eval_runtime": 13.1569,
      "eval_samples_per_second": 76.006,
      "eval_steps_per_second": 2.432,
      "step": 3200
    },
    {
      "epoch": 8.37898807605543,
      "grad_norm": 1.3793811731790488,
      "learning_rate": 8.98602903244165e-05,
      "loss": 0.7843,
      "step": 3250
    },
    {
      "epoch": 8.507895584917822,
      "grad_norm": 1.0500789470659453,
      "learning_rate": 8.955267912327574e-05,
      "loss": 0.8025,
      "step": 3300
    },
    {
      "epoch": 8.507895584917822,
      "eval_loss": 2.4928858280181885,
      "eval_runtime": 13.1471,
      "eval_samples_per_second": 76.062,
      "eval_steps_per_second": 2.434,
      "step": 3300
    },
    {
      "epoch": 8.636803093780212,
      "grad_norm": 1.0687481728737136,
      "learning_rate": 8.924102071917248e-05,
      "loss": 0.8292,
      "step": 3350
    },
    {
      "epoch": 8.765710602642605,
      "grad_norm": 0.8890330524205967,
      "learning_rate": 8.89253474106544e-05,
      "loss": 0.8554,
      "step": 3400
    },
    {
      "epoch": 8.765710602642605,
      "eval_loss": 2.4535696506500244,
      "eval_runtime": 13.1668,
      "eval_samples_per_second": 75.949,
      "eval_steps_per_second": 2.43,
      "step": 3400
    },
    {
      "epoch": 8.894618111504995,
      "grad_norm": 1.1314521221393643,
      "learning_rate": 8.860569191235147e-05,
      "loss": 0.8804,
      "step": 3450
    },
    {
      "epoch": 9.023525620367387,
      "grad_norm": 1.1316860727966558,
      "learning_rate": 8.828208735158577e-05,
      "loss": 0.8251,
      "step": 3500
    },
    {
      "epoch": 9.023525620367387,
      "eval_loss": 2.771848440170288,
      "eval_runtime": 13.1585,
      "eval_samples_per_second": 75.996,
      "eval_steps_per_second": 2.432,
      "step": 3500
    },
    {
      "epoch": 9.152433129229777,
      "grad_norm": 2.2363185101646765,
      "learning_rate": 8.795456726493838e-05,
      "loss": 0.5115,
      "step": 3550
    },
    {
      "epoch": 9.28134063809217,
      "grad_norm": 1.1689864457962205,
      "learning_rate": 8.762316559477361e-05,
      "loss": 0.5483,
      "step": 3600
    },
    {
      "epoch": 9.28134063809217,
      "eval_loss": 2.738429307937622,
      "eval_runtime": 13.1693,
      "eval_samples_per_second": 75.934,
      "eval_steps_per_second": 2.43,
      "step": 3600
    },
    {
      "epoch": 9.41024814695456,
      "grad_norm": 1.1593603230103788,
      "learning_rate": 8.728791668572168e-05,
      "loss": 0.581,
      "step": 3650
    },
    {
      "epoch": 9.539155655816952,
      "grad_norm": 1.4682987033608552,
      "learning_rate": 8.694885528111918e-05,
      "loss": 0.6167,
      "step": 3700
    },
    {
      "epoch": 9.539155655816952,
      "eval_loss": 2.702000379562378,
      "eval_runtime": 13.1618,
      "eval_samples_per_second": 75.978,
      "eval_steps_per_second": 2.431,
      "step": 3700
    },
    {
      "epoch": 9.668063164679342,
      "grad_norm": 1.3807999009672582,
      "learning_rate": 8.66060165194087e-05,
      "loss": 0.6364,
      "step": 3750
    },
    {
      "epoch": 9.796970673541734,
      "grad_norm": 1.6910628067329174,
      "learning_rate": 8.625943593049708e-05,
      "loss": 0.664,
      "step": 3800
    },
    {
      "epoch": 9.796970673541734,
      "eval_loss": 2.6746528148651123,
      "eval_runtime": 13.1207,
      "eval_samples_per_second": 76.216,
      "eval_steps_per_second": 2.439,
      "step": 3800
    },
    {
      "epoch": 9.925878182404125,
      "grad_norm": 1.2810221531255845,
      "learning_rate": 8.590914943207338e-05,
      "loss": 0.6808,
      "step": 3850
    },
    {
      "epoch": 10.054785691266517,
      "grad_norm": 0.894817953292936,
      "learning_rate": 8.555519332588658e-05,
      "loss": 0.5501,
      "step": 3900
    },
    {
      "epoch": 10.054785691266517,
      "eval_loss": 2.9522013664245605,
      "eval_runtime": 13.1696,
      "eval_samples_per_second": 75.933,
      "eval_steps_per_second": 2.43,
      "step": 3900
    },
    {
      "epoch": 10.183693200128907,
      "grad_norm": 0.7673858183230434,
      "learning_rate": 8.519760429398344e-05,
      "loss": 0.3676,
      "step": 3950
    },
    {
      "epoch": 10.3126007089913,
      "grad_norm": 0.9258575274469117,
      "learning_rate": 8.4836419394907e-05,
      "loss": 0.3948,
      "step": 4000
    },
    {
      "epoch": 10.3126007089913,
      "eval_loss": 2.939727544784546,
      "eval_runtime": 13.261,
      "eval_samples_per_second": 75.409,
      "eval_steps_per_second": 2.413,
      "step": 4000
    },
    {
      "epoch": 10.44150821785369,
      "grad_norm": 0.9837761849309598,
      "learning_rate": 8.447167605985595e-05,
      "loss": 0.424,
      "step": 4050
    },
    {
      "epoch": 10.570415726716082,
      "grad_norm": 1.1827846876344605,
      "learning_rate": 8.410341208880562e-05,
      "loss": 0.4493,
      "step": 4100
    },
    {
      "epoch": 10.570415726716082,
      "eval_loss": 2.9346396923065186,
      "eval_runtime": 13.1851,
      "eval_samples_per_second": 75.843,
      "eval_steps_per_second": 2.427,
      "step": 4100
    },
    {
      "epoch": 10.699323235578472,
      "grad_norm": 1.41364563812005,
      "learning_rate": 8.373166564659048e-05,
      "loss": 0.4724,
      "step": 4150
    },
    {
      "epoch": 10.828230744440864,
      "grad_norm": 1.341861709508504,
      "learning_rate": 8.335647525894898e-05,
      "loss": 0.4991,
      "step": 4200
    },
    {
      "epoch": 10.828230744440864,
      "eval_loss": 2.8969509601593018,
      "eval_runtime": 13.153,
      "eval_samples_per_second": 76.028,
      "eval_steps_per_second": 2.433,
      "step": 4200
    },
    {
      "epoch": 10.957138253303254,
      "grad_norm": 0.8732212524533925,
      "learning_rate": 8.297787980853102e-05,
      "loss": 0.514,
      "step": 4250
    },
    {
      "epoch": 11.086045762165647,
      "grad_norm": 0.9580077889491428,
      "learning_rate": 8.259591853086822e-05,
      "loss": 0.3342,
      "step": 4300
    },
    {
      "epoch": 11.086045762165647,
      "eval_loss": 3.12947940826416,
      "eval_runtime": 13.164,
      "eval_samples_per_second": 75.964,
      "eval_steps_per_second": 2.431,
      "step": 4300
    },
    {
      "epoch": 11.214953271028037,
      "grad_norm": 1.0158413995480675,
      "learning_rate": 8.221063101030793e-05,
      "loss": 0.2608,
      "step": 4350
    },
    {
      "epoch": 11.343860779890429,
      "grad_norm": 1.066799301262293,
      "learning_rate": 8.182205717591083e-05,
      "loss": 0.2868,
      "step": 4400
    },
    {
      "epoch": 11.343860779890429,
      "eval_loss": 3.152151107788086,
      "eval_runtime": 13.1308,
      "eval_samples_per_second": 76.157,
      "eval_steps_per_second": 2.437,
      "step": 4400
    },
    {
      "epoch": 11.47276828875282,
      "grad_norm": 1.7853526972041276,
      "learning_rate": 8.143023729731294e-05,
      "loss": 0.3088,
      "step": 4450
    },
    {
      "epoch": 11.601675797615211,
      "grad_norm": 0.8686004168533566,
      "learning_rate": 8.103521198055229e-05,
      "loss": 0.3307,
      "step": 4500
    },
    {
      "epoch": 11.601675797615211,
      "eval_loss": 3.140380382537842,
      "eval_runtime": 13.1601,
      "eval_samples_per_second": 75.987,
      "eval_steps_per_second": 2.432,
      "step": 4500
    },
    {
      "epoch": 11.730583306477602,
      "grad_norm": 1.1842485940651084,
      "learning_rate": 8.063702216386072e-05,
      "loss": 0.3505,
      "step": 4550
    },
    {
      "epoch": 11.859490815339994,
      "grad_norm": 1.139988677136263,
      "learning_rate": 8.02357091134213e-05,
      "loss": 0.3798,
      "step": 4600
    },
    {
      "epoch": 11.859490815339994,
      "eval_loss": 3.1152422428131104,
      "eval_runtime": 13.1658,
      "eval_samples_per_second": 75.954,
      "eval_steps_per_second": 2.431,
      "step": 4600
    },
    {
      "epoch": 11.988398324202384,
      "grad_norm": 1.1317453381475029,
      "learning_rate": 7.983131441909169e-05,
      "loss": 0.383,
      "step": 4650
    },
    {
      "epoch": 12.117305833064776,
      "grad_norm": 0.8577623285954574,
      "learning_rate": 7.942387999009405e-05,
      "loss": 0.1955,
      "step": 4700
    },
    {
      "epoch": 12.117305833064776,
      "eval_loss": 3.292053699493408,
      "eval_runtime": 13.1656,
      "eval_samples_per_second": 75.955,
      "eval_steps_per_second": 2.431,
      "step": 4700
    },
    {
      "epoch": 12.246213341927167,
      "grad_norm": 0.792139936919041,
      "learning_rate": 7.901344805067176e-05,
      "loss": 0.2018,
      "step": 4750
    },
    {
      "epoch": 12.375120850789559,
      "grad_norm": 1.1197587931587643,
      "learning_rate": 7.860006113571356e-05,
      "loss": 0.2149,
      "step": 4800
    },
    {
      "epoch": 12.375120850789559,
      "eval_loss": 3.300734519958496,
      "eval_runtime": 13.2004,
      "eval_samples_per_second": 75.755,
      "eval_steps_per_second": 2.424,
      "step": 4800
    },
    {
      "epoch": 12.504028359651949,
      "grad_norm": 0.9715029062397967,
      "learning_rate": 7.818376208634544e-05,
      "loss": 0.2326,
      "step": 4850
    },
    {
      "epoch": 12.632935868514341,
      "grad_norm": 0.9363158756959897,
      "learning_rate": 7.776459404549084e-05,
      "loss": 0.2492,
      "step": 4900
    },
    {
      "epoch": 12.632935868514341,
      "eval_loss": 3.3016297817230225,
      "eval_runtime": 13.1894,
      "eval_samples_per_second": 75.819,
      "eval_steps_per_second": 2.426,
      "step": 4900
    },
    {
      "epoch": 12.761843377376731,
      "grad_norm": 0.8911732251342259,
      "learning_rate": 7.734260045339957e-05,
      "loss": 0.2662,
      "step": 4950
    },
    {
      "epoch": 12.890750886239124,
      "grad_norm": 1.0150499184539068,
      "learning_rate": 7.69178250431459e-05,
      "loss": 0.2797,
      "step": 5000
    },
    {
      "epoch": 12.890750886239124,
      "eval_loss": 3.2908613681793213,
      "eval_runtime": 13.193,
      "eval_samples_per_second": 75.798,
      "eval_steps_per_second": 2.426,
      "step": 5000
    },
    {
      "epoch": 13.019658395101514,
      "grad_norm": 0.9560026596743003,
      "learning_rate": 7.649031183609627e-05,
      "loss": 0.27,
      "step": 5050
    },
    {
      "epoch": 13.148565903963906,
      "grad_norm": 0.7460724556430167,
      "learning_rate": 7.606010513734722e-05,
      "loss": 0.1396,
      "step": 5100
    },
    {
      "epoch": 13.148565903963906,
      "eval_loss": 3.4375460147857666,
      "eval_runtime": 13.1426,
      "eval_samples_per_second": 76.088,
      "eval_steps_per_second": 2.435,
      "step": 5100
    },
    {
      "epoch": 13.277473412826296,
      "grad_norm": 1.2891823937464943,
      "learning_rate": 7.562724953113382e-05,
      "loss": 0.1548,
      "step": 5150
    },
    {
      "epoch": 13.406380921688688,
      "grad_norm": 0.8052288266291375,
      "learning_rate": 7.519178987620915e-05,
      "loss": 0.1697,
      "step": 5200
    },
    {
      "epoch": 13.406380921688688,
      "eval_loss": 3.4272212982177734,
      "eval_runtime": 13.1636,
      "eval_samples_per_second": 75.967,
      "eval_steps_per_second": 2.431,
      "step": 5200
    },
    {
      "epoch": 13.535288430551079,
      "grad_norm": 0.7626827120967569,
      "learning_rate": 7.475377130119553e-05,
      "loss": 0.1836,
      "step": 5250
    },
    {
      "epoch": 13.66419593941347,
      "grad_norm": 0.7817292246891677,
      "learning_rate": 7.431323919990754e-05,
      "loss": 0.1948,
      "step": 5300
    },
    {
      "epoch": 13.66419593941347,
      "eval_loss": 3.4388325214385986,
      "eval_runtime": 13.1888,
      "eval_samples_per_second": 75.822,
      "eval_steps_per_second": 2.426,
      "step": 5300
    },
    {
      "epoch": 13.793103448275861,
      "grad_norm": 0.7362472448079689,
      "learning_rate": 7.387023922664762e-05,
      "loss": 0.2081,
      "step": 5350
    },
    {
      "epoch": 13.922010957138253,
      "grad_norm": 0.7773529994490974,
      "learning_rate": 7.342481729147485e-05,
      "loss": 0.2165,
      "step": 5400
    },
    {
      "epoch": 13.922010957138253,
      "eval_loss": 3.435514450073242,
      "eval_runtime": 13.2029,
      "eval_samples_per_second": 75.741,
      "eval_steps_per_second": 2.424,
      "step": 5400
    },
    {
      "epoch": 14.050918466000645,
      "grad_norm": 1.3925624599380586,
      "learning_rate": 7.297701955544692e-05,
      "loss": 0.1793,
      "step": 5450
    },
    {
      "epoch": 14.179825974863036,
      "grad_norm": 0.6045993909446667,
      "learning_rate": 7.252689242583643e-05,
      "loss": 0.1302,
      "step": 5500
    },
    {
      "epoch": 14.179825974863036,
      "eval_loss": 3.5325989723205566,
      "eval_runtime": 13.2171,
      "eval_samples_per_second": 75.659,
      "eval_steps_per_second": 2.421,
      "step": 5500
    },
    {
      "epoch": 14.308733483725428,
      "grad_norm": 0.6099737603899711,
      "learning_rate": 7.20744825513213e-05,
      "loss": 0.1265,
      "step": 5550
    },
    {
      "epoch": 14.437640992587818,
      "grad_norm": 0.6596680228597308,
      "learning_rate": 7.161983681715056e-05,
      "loss": 0.138,
      "step": 5600
    },
    {
      "epoch": 14.437640992587818,
      "eval_loss": 3.557699680328369,
      "eval_runtime": 13.1774,
      "eval_samples_per_second": 75.888,
      "eval_steps_per_second": 2.428,
      "step": 5600
    },
    {
      "epoch": 14.56654850145021,
      "grad_norm": 0.6445903882034174,
      "learning_rate": 7.116300234028528e-05,
      "loss": 0.1486,
      "step": 5650
    },
    {
      "epoch": 14.6954560103126,
      "grad_norm": 0.6448511072922888,
      "learning_rate": 7.070402646451566e-05,
      "loss": 0.1569,
      "step": 5700
    },
    {
      "epoch": 14.6954560103126,
      "eval_loss": 3.5597221851348877,
      "eval_runtime": 13.1454,
      "eval_samples_per_second": 76.072,
      "eval_steps_per_second": 2.434,
      "step": 5700
    },
    {
      "epoch": 14.824363519174993,
      "grad_norm": 0.623499885141096,
      "learning_rate": 7.024295675555465e-05,
      "loss": 0.1675,
      "step": 5750
    },
    {
      "epoch": 14.953271028037383,
      "grad_norm": 0.6601246140904097,
      "learning_rate": 6.977984099610842e-05,
      "loss": 0.1757,
      "step": 5800
    },
    {
      "epoch": 14.953271028037383,
      "eval_loss": 3.555185556411743,
      "eval_runtime": 13.1511,
      "eval_samples_per_second": 76.039,
      "eval_steps_per_second": 2.433,
      "step": 5800
    },
    {
      "epoch": 15.082178536899775,
      "grad_norm": 0.5535844816659106,
      "learning_rate": 6.931472718092446e-05,
      "loss": 0.1223,
      "step": 5850
    },
    {
      "epoch": 15.211086045762165,
      "grad_norm": 0.4901789372000881,
      "learning_rate": 6.88476635118177e-05,
      "loss": 0.0964,
      "step": 5900
    },
    {
      "epoch": 15.211086045762165,
      "eval_loss": 3.6443583965301514,
      "eval_runtime": 13.1441,
      "eval_samples_per_second": 76.08,
      "eval_steps_per_second": 2.435,
      "step": 5900
    },
    {
      "epoch": 15.339993554624558,
      "grad_norm": 0.8805753683037839,
      "learning_rate": 6.837869839267507e-05,
      "loss": 0.107,
      "step": 5950
    },
    {
      "epoch": 15.468901063486948,
      "grad_norm": 0.6115833032151868,
      "learning_rate": 6.790788042443924e-05,
      "loss": 0.1157,
      "step": 6000
    },
    {
      "epoch": 15.468901063486948,
      "eval_loss": 3.6541991233825684,
      "eval_runtime": 13.1423,
      "eval_samples_per_second": 76.09,
      "eval_steps_per_second": 2.435,
      "step": 6000
    },
    {
      "epoch": 15.59780857234934,
      "grad_norm": 0.5710489600928594,
      "learning_rate": 6.743525840007191e-05,
      "loss": 0.1228,
      "step": 6050
    },
    {
      "epoch": 15.72671608121173,
      "grad_norm": 0.689652930712509,
      "learning_rate": 6.69608812994971e-05,
      "loss": 0.1293,
      "step": 6100
    },
    {
      "epoch": 15.72671608121173,
      "eval_loss": 3.656620979309082,
      "eval_runtime": 13.1483,
      "eval_samples_per_second": 76.055,
      "eval_steps_per_second": 2.434,
      "step": 6100
    },
    {
      "epoch": 15.855623590074122,
      "grad_norm": 0.5914349974251379,
      "learning_rate": 6.648479828452522e-05,
      "loss": 0.135,
      "step": 6150
    },
    {
      "epoch": 15.984531098936513,
      "grad_norm": 0.6189439108834661,
      "learning_rate": 6.600705869375817e-05,
      "loss": 0.1419,
      "step": 6200
    },
    {
      "epoch": 15.984531098936513,
      "eval_loss": 3.671614170074463,
      "eval_runtime": 13.1443,
      "eval_samples_per_second": 76.078,
      "eval_steps_per_second": 2.435,
      "step": 6200
    },
    {
      "epoch": 16.113438607798905,
      "grad_norm": 0.4321088458532425,
      "learning_rate": 6.55277120374762e-05,
      "loss": 0.0842,
      "step": 6250
    },
    {
      "epoch": 16.242346116661295,
      "grad_norm": 0.5785562511316525,
      "learning_rate": 6.504680799250689e-05,
      "loss": 0.0872,
      "step": 6300
    },
    {
      "epoch": 16.242346116661295,
      "eval_loss": 3.7362146377563477,
      "eval_runtime": 13.1708,
      "eval_samples_per_second": 75.925,
      "eval_steps_per_second": 2.43,
      "step": 6300
    },
    {
      "epoch": 16.371253625523686,
      "grad_norm": 0.45371653083862784,
      "learning_rate": 6.45643963970769e-05,
      "loss": 0.0899,
      "step": 6350
    },
    {
      "epoch": 16.50016113438608,
      "grad_norm": 0.5412656335319528,
      "learning_rate": 6.408052724564714e-05,
      "loss": 0.0954,
      "step": 6400
    },
    {
      "epoch": 16.50016113438608,
      "eval_loss": 3.7513670921325684,
      "eval_runtime": 13.1547,
      "eval_samples_per_second": 76.018,
      "eval_steps_per_second": 2.433,
      "step": 6400
    },
    {
      "epoch": 16.62906864324847,
      "grad_norm": 0.6489716287488846,
      "learning_rate": 6.359525068373147e-05,
      "loss": 0.1021,
      "step": 6450
    },
    {
      "epoch": 16.75797615211086,
      "grad_norm": 0.51515711941371,
      "learning_rate": 6.310861700270001e-05,
      "loss": 0.1074,
      "step": 6500
    },
    {
      "epoch": 16.75797615211086,
      "eval_loss": 3.7568299770355225,
      "eval_runtime": 13.1646,
      "eval_samples_per_second": 75.961,
      "eval_steps_per_second": 2.431,
      "step": 6500
    },
    {
      "epoch": 16.88688366097325,
      "grad_norm": 0.5784472468074615,
      "learning_rate": 6.262067663456714e-05,
      "loss": 0.1119,
      "step": 6550
    },
    {
      "epoch": 17.015791169835644,
      "grad_norm": 0.36703765035846453,
      "learning_rate": 6.2131480146765e-05,
      "loss": 0.1089,
      "step": 6600
    },
    {
      "epoch": 17.015791169835644,
      "eval_loss": 3.8403706550598145,
      "eval_runtime": 13.14,
      "eval_samples_per_second": 76.103,
      "eval_steps_per_second": 2.435,
      "step": 6600
    },
    {
      "epoch": 17.144698678698035,
      "grad_norm": 0.473234963059124,
      "learning_rate": 6.164107823690311e-05,
      "loss": 0.0653,
      "step": 6650
    },
    {
      "epoch": 17.273606187560425,
      "grad_norm": 0.4495142121889821,
      "learning_rate": 6.11495217275142e-05,
      "loss": 0.071,
      "step": 6700
    },
    {
      "epoch": 17.273606187560425,
      "eval_loss": 3.830505609512329,
      "eval_runtime": 13.1641,
      "eval_samples_per_second": 75.964,
      "eval_steps_per_second": 2.431,
      "step": 6700
    },
    {
      "epoch": 17.402513696422815,
      "grad_norm": 0.47554094680025083,
      "learning_rate": 6.065686156078723e-05,
      "loss": 0.0758,
      "step": 6750
    },
    {
      "epoch": 17.53142120528521,
      "grad_norm": 0.44446520713729815,
      "learning_rate": 6.0163148793288135e-05,
      "loss": 0.0803,
      "step": 6800
    },
    {
      "epoch": 17.53142120528521,
      "eval_loss": 3.8333582878112793,
      "eval_runtime": 13.3574,
      "eval_samples_per_second": 74.865,
      "eval_steps_per_second": 2.396,
      "step": 6800
    },
    {
      "epoch": 17.6603287141476,
      "grad_norm": 0.5076778674223996,
      "learning_rate": 5.966843459066858e-05,
      "loss": 0.0847,
      "step": 6850
    },
    {
      "epoch": 17.78923622300999,
      "grad_norm": 0.49074606040336083,
      "learning_rate": 5.917277022236333e-05,
      "loss": 0.0882,
      "step": 6900
    },
    {
      "epoch": 17.78923622300999,
      "eval_loss": 3.840571403503418,
      "eval_runtime": 13.1466,
      "eval_samples_per_second": 76.065,
      "eval_steps_per_second": 2.434,
      "step": 6900
    },
    {
      "epoch": 17.91814373187238,
      "grad_norm": 0.5550690055495523,
      "learning_rate": 5.86762070562771e-05,
      "loss": 0.0927,
      "step": 6950
    },
    {
      "epoch": 18.047051240734774,
      "grad_norm": 0.37606709531546206,
      "learning_rate": 5.817879655346103e-05,
      "loss": 0.0805,
      "step": 7000
    },
    {
      "epoch": 18.047051240734774,
      "eval_loss": 3.8941986560821533,
      "eval_runtime": 13.1572,
      "eval_samples_per_second": 76.004,
      "eval_steps_per_second": 2.432,
      "step": 7000
    },
    {
      "epoch": 18.175958749597164,
      "grad_norm": 0.48969762050004223,
      "learning_rate": 5.7680590262779535e-05,
      "loss": 0.0561,
      "step": 7050
    },
    {
      "epoch": 18.304866258459555,
      "grad_norm": 0.43070776806915745,
      "learning_rate": 5.718163981556802e-05,
      "loss": 0.061,
      "step": 7100
    },
    {
      "epoch": 18.304866258459555,
      "eval_loss": 3.9117181301116943,
      "eval_runtime": 13.1597,
      "eval_samples_per_second": 75.99,
      "eval_steps_per_second": 2.432,
      "step": 7100
    },
    {
      "epoch": 18.433773767321945,
      "grad_norm": 0.42717221620170337,
      "learning_rate": 5.6681996920282174e-05,
      "loss": 0.0649,
      "step": 7150
    },
    {
      "epoch": 18.56268127618434,
      "grad_norm": 0.5194058540782253,
      "learning_rate": 5.6181713357139184e-05,
      "loss": 0.0682,
      "step": 7200
    },
    {
      "epoch": 18.56268127618434,
      "eval_loss": 3.901904821395874,
      "eval_runtime": 13.1719,
      "eval_samples_per_second": 75.919,
      "eval_steps_per_second": 2.429,
      "step": 7200
    },
    {
      "epoch": 18.69158878504673,
      "grad_norm": 0.37203814221096937,
      "learning_rate": 5.568084097275149e-05,
      "loss": 0.0719,
      "step": 7250
    },
    {
      "epoch": 18.82049629390912,
      "grad_norm": 0.4355364201650638,
      "learning_rate": 5.517943167475367e-05,
      "loss": 0.0736,
      "step": 7300
    },
    {
      "epoch": 18.82049629390912,
      "eval_loss": 3.927724599838257,
      "eval_runtime": 13.1773,
      "eval_samples_per_second": 75.888,
      "eval_steps_per_second": 2.428,
      "step": 7300
    },
    {
      "epoch": 18.94940380277151,
      "grad_norm": 0.470475938907921,
      "learning_rate": 5.4677537426423044e-05,
      "loss": 0.0765,
      "step": 7350
    },
    {
      "epoch": 19.078311311633904,
      "grad_norm": 0.41878300792660983,
      "learning_rate": 5.4175210241294564e-05,
      "loss": 0.0584,
      "step": 7400
    },
    {
      "epoch": 19.078311311633904,
      "eval_loss": 3.983382225036621,
      "eval_runtime": 13.1436,
      "eval_samples_per_second": 76.082,
      "eval_steps_per_second": 2.435,
      "step": 7400
    },
    {
      "epoch": 19.207218820496294,
      "grad_norm": 0.363679410249128,
      "learning_rate": 5.367250217777029e-05,
      "loss": 0.0488,
      "step": 7450
    },
    {
      "epoch": 19.336126329358684,
      "grad_norm": 0.38290054329422646,
      "learning_rate": 5.316946533372439e-05,
      "loss": 0.0518,
      "step": 7500
    },
    {
      "epoch": 19.336126329358684,
      "eval_loss": 3.965801954269409,
      "eval_runtime": 13.1515,
      "eval_samples_per_second": 76.037,
      "eval_steps_per_second": 2.433,
      "step": 7500
    },
    {
      "epoch": 19.465033838221075,
      "grad_norm": 0.3547993700029784,
      "learning_rate": 5.266615184110405e-05,
      "loss": 0.0546,
      "step": 7550
    },
    {
      "epoch": 19.59394134708347,
      "grad_norm": 0.37584304866780927,
      "learning_rate": 5.216261386052682e-05,
      "loss": 0.0565,
      "step": 7600
    },
    {
      "epoch": 19.59394134708347,
      "eval_loss": 3.980562686920166,
      "eval_runtime": 13.1525,
      "eval_samples_per_second": 76.031,
      "eval_steps_per_second": 2.433,
      "step": 7600
    },
    {
      "epoch": 19.72284885594586,
      "grad_norm": 0.38295332236701923,
      "learning_rate": 5.165890357587486e-05,
      "loss": 0.059,
      "step": 7650
    },
    {
      "epoch": 19.85175636480825,
      "grad_norm": 0.41601237103360944,
      "learning_rate": 5.115507318888699e-05,
      "loss": 0.0608,
      "step": 7700
    },
    {
      "epoch": 19.85175636480825,
      "eval_loss": 3.9985480308532715,
      "eval_runtime": 13.1495,
      "eval_samples_per_second": 76.049,
      "eval_steps_per_second": 2.434,
      "step": 7700
    },
    {
      "epoch": 19.98066387367064,
      "grad_norm": 0.40935703530908246,
      "learning_rate": 5.0651174913748836e-05,
      "loss": 0.0631,
      "step": 7750
    },
    {
      "epoch": 20.109571382533034,
      "grad_norm": 0.3115920333441882,
      "learning_rate": 5.0147260971681466e-05,
      "loss": 0.0434,
      "step": 7800
    },
    {
      "epoch": 20.109571382533034,
      "eval_loss": 4.044510364532471,
      "eval_runtime": 13.1761,
      "eval_samples_per_second": 75.895,
      "eval_steps_per_second": 2.429,
      "step": 7800
    },
    {
      "epoch": 20.238478891395424,
      "grad_norm": 0.33928266884719477,
      "learning_rate": 4.9643383585529644e-05,
      "loss": 0.0417,
      "step": 7850
    },
    {
      "epoch": 20.367386400257814,
      "grad_norm": 0.366492032031904,
      "learning_rate": 4.9139594974349714e-05,
      "loss": 0.0469,
      "step": 7900
    },
    {
      "epoch": 20.367386400257814,
      "eval_loss": 4.039082050323486,
      "eval_runtime": 13.1597,
      "eval_samples_per_second": 75.989,
      "eval_steps_per_second": 2.432,
      "step": 7900
    },
    {
      "epoch": 20.496293909120205,
      "grad_norm": 0.314671889688573,
      "learning_rate": 4.863594734799778e-05,
      "loss": 0.0458,
      "step": 7950
    },
    {
      "epoch": 20.6252014179826,
      "grad_norm": 0.34654066388282007,
      "learning_rate": 4.813249290171915e-05,
      "loss": 0.0477,
      "step": 8000
    },
    {
      "epoch": 20.6252014179826,
      "eval_loss": 4.047826766967773,
      "eval_runtime": 13.1879,
      "eval_samples_per_second": 75.827,
      "eval_steps_per_second": 2.426,
      "step": 8000
    }
  ],
  "logging_steps": 50,
  "max_steps": 15480,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 40,
  "save_steps": 800,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": false
      },
      "attributes": {}
    }
  },
  "total_flos": 7114179689316352.0,
  "train_batch_size": 4,
  "trial_name": null,
  "trial_params": null
}