{
    "best_metric": 0.8043478260869565,
    "best_model_checkpoint": "CT-MAE-RIS2-Phase2-V1/checkpoint-520",
    "epoch": 49.02,
    "eval_steps": 500,
    "global_step": 3250,
    "is_hyper_param_search": false,
    "is_local_process_zero": true,
    "is_world_process_zero": true,
    "log_history": [
      {
        "epoch": 0.003076923076923077,
        "grad_norm": 2.7215452194213867,
        "learning_rate": 3.0769230769230774e-07,
        "loss": 0.6876,
        "step": 10
      },
      {
        "epoch": 0.006153846153846154,
        "grad_norm": 2.571739673614502,
        "learning_rate": 6.153846153846155e-07,
        "loss": 0.6748,
        "step": 20
      },
      {
        "epoch": 0.009230769230769232,
        "grad_norm": 4.409002780914307,
        "learning_rate": 9.230769230769232e-07,
        "loss": 0.6605,
        "step": 30
      },
      {
        "epoch": 0.012307692307692308,
        "grad_norm": 4.084295272827148,
        "learning_rate": 1.230769230769231e-06,
        "loss": 0.6389,
        "step": 40
      },
      {
        "epoch": 0.015384615384615385,
        "grad_norm": 11.590015411376953,
        "learning_rate": 1.5384615384615387e-06,
        "loss": 0.5934,
        "step": 50
      },
      {
        "epoch": 0.018461538461538463,
        "grad_norm": 8.831100463867188,
        "learning_rate": 1.8461538461538465e-06,
        "loss": 0.6186,
        "step": 60
      },
      {
        "epoch": 0.02,
        "eval_accuracy": 0.5434782608695652,
        "eval_loss": 0.7367333173751831,
        "eval_runtime": 11.5008,
        "eval_samples_per_second": 4.0,
        "eval_steps_per_second": 1.043,
        "step": 65
      },
      {
        "epoch": 1.0015384615384615,
        "grad_norm": 11.279766082763672,
        "learning_rate": 2.153846153846154e-06,
        "loss": 0.5907,
        "step": 70
      },
      {
        "epoch": 1.0046153846153847,
        "grad_norm": 5.7519145011901855,
        "learning_rate": 2.461538461538462e-06,
        "loss": 0.617,
        "step": 80
      },
      {
        "epoch": 1.0076923076923077,
        "grad_norm": 4.054751873016357,
        "learning_rate": 2.7692307692307697e-06,
        "loss": 0.6291,
        "step": 90
      },
      {
        "epoch": 1.0107692307692309,
        "grad_norm": 5.010909080505371,
        "learning_rate": 3.0769230769230774e-06,
        "loss": 0.528,
        "step": 100
      },
      {
        "epoch": 1.0138461538461538,
        "grad_norm": 14.651056289672852,
        "learning_rate": 3.384615384615385e-06,
        "loss": 0.4573,
        "step": 110
      },
      {
        "epoch": 1.0169230769230768,
        "grad_norm": 11.85365104675293,
        "learning_rate": 3.692307692307693e-06,
        "loss": 0.6294,
        "step": 120
      },
      {
        "epoch": 1.02,
        "grad_norm": 8.004859924316406,
        "learning_rate": 4.000000000000001e-06,
        "loss": 0.5974,
        "step": 130
      },
      {
        "epoch": 1.02,
        "eval_accuracy": 0.5434782608695652,
        "eval_loss": 0.8185142874717712,
        "eval_runtime": 10.3925,
        "eval_samples_per_second": 4.426,
        "eval_steps_per_second": 1.155,
        "step": 130
      },
      {
        "epoch": 2.003076923076923,
        "grad_norm": 14.822097778320312,
        "learning_rate": 4.307692307692308e-06,
        "loss": 0.5002,
        "step": 140
      },
      {
        "epoch": 2.006153846153846,
        "grad_norm": 10.74063491821289,
        "learning_rate": 4.615384615384616e-06,
        "loss": 0.6184,
        "step": 150
      },
      {
        "epoch": 2.0092307692307694,
        "grad_norm": 9.221219062805176,
        "learning_rate": 4.923076923076924e-06,
        "loss": 0.5821,
        "step": 160
      },
      {
        "epoch": 2.0123076923076924,
        "grad_norm": 4.616571426391602,
        "learning_rate": 5.230769230769232e-06,
        "loss": 0.617,
        "step": 170
      },
      {
        "epoch": 2.0153846153846153,
        "grad_norm": 8.066702842712402,
        "learning_rate": 5.538461538461539e-06,
        "loss": 0.561,
        "step": 180
      },
      {
        "epoch": 2.0184615384615383,
        "grad_norm": 7.593764781951904,
        "learning_rate": 5.846153846153847e-06,
        "loss": 0.5491,
        "step": 190
      },
      {
        "epoch": 2.02,
        "eval_accuracy": 0.5434782608695652,
        "eval_loss": 0.8371568918228149,
        "eval_runtime": 10.653,
        "eval_samples_per_second": 4.318,
        "eval_steps_per_second": 1.126,
        "step": 195
      },
      {
        "epoch": 3.0015384615384617,
        "grad_norm": 6.1916985511779785,
        "learning_rate": 6.153846153846155e-06,
        "loss": 0.5725,
        "step": 200
      },
      {
        "epoch": 3.0046153846153847,
        "grad_norm": 6.418662071228027,
        "learning_rate": 6.461538461538463e-06,
        "loss": 0.4794,
        "step": 210
      },
      {
        "epoch": 3.0076923076923077,
        "grad_norm": 9.704038619995117,
        "learning_rate": 6.76923076923077e-06,
        "loss": 0.4911,
        "step": 220
      },
      {
        "epoch": 3.0107692307692306,
        "grad_norm": 5.40470027923584,
        "learning_rate": 7.076923076923078e-06,
        "loss": 0.5041,
        "step": 230
      },
      {
        "epoch": 3.013846153846154,
        "grad_norm": 7.060332775115967,
        "learning_rate": 7.384615384615386e-06,
        "loss": 0.591,
        "step": 240
      },
      {
        "epoch": 3.016923076923077,
        "grad_norm": 34.150779724121094,
        "learning_rate": 7.692307692307694e-06,
        "loss": 0.5028,
        "step": 250
      },
      {
        "epoch": 3.02,
        "grad_norm": 6.146841526031494,
        "learning_rate": 8.000000000000001e-06,
        "loss": 0.6156,
        "step": 260
      },
      {
        "epoch": 3.02,
        "eval_accuracy": 0.5869565217391305,
        "eval_loss": 0.6619947552680969,
        "eval_runtime": 10.4311,
        "eval_samples_per_second": 4.41,
        "eval_steps_per_second": 1.15,
        "step": 260
      },
      {
        "epoch": 4.003076923076923,
        "grad_norm": 2.565047025680542,
        "learning_rate": 8.307692307692309e-06,
        "loss": 0.4261,
        "step": 270
      },
      {
        "epoch": 4.006153846153846,
        "grad_norm": 7.034364700317383,
        "learning_rate": 8.615384615384617e-06,
        "loss": 0.647,
        "step": 280
      },
      {
        "epoch": 4.009230769230769,
        "grad_norm": 6.253409385681152,
        "learning_rate": 8.923076923076925e-06,
        "loss": 0.5691,
        "step": 290
      },
      {
        "epoch": 4.012307692307692,
        "grad_norm": 1.1683131456375122,
        "learning_rate": 9.230769230769232e-06,
        "loss": 0.4258,
        "step": 300
      },
      {
        "epoch": 4.015384615384615,
        "grad_norm": 18.69379234313965,
        "learning_rate": 9.53846153846154e-06,
        "loss": 0.9862,
        "step": 310
      },
      {
        "epoch": 4.018461538461539,
        "grad_norm": 9.839452743530273,
        "learning_rate": 9.846153846153848e-06,
        "loss": 0.6255,
        "step": 320
      },
      {
        "epoch": 4.02,
        "eval_accuracy": 0.5434782608695652,
        "eval_loss": 0.6834591031074524,
        "eval_runtime": 10.7712,
        "eval_samples_per_second": 4.271,
        "eval_steps_per_second": 1.114,
        "step": 325
      },
      {
        "epoch": 5.001538461538462,
        "grad_norm": 11.986108779907227,
        "learning_rate": 9.982905982905984e-06,
        "loss": 0.5837,
        "step": 330
      },
      {
        "epoch": 5.004615384615384,
        "grad_norm": 10.73773193359375,
        "learning_rate": 9.94871794871795e-06,
        "loss": 0.6487,
        "step": 340
      },
      {
        "epoch": 5.007692307692308,
        "grad_norm": 2.132850170135498,
        "learning_rate": 9.914529914529915e-06,
        "loss": 0.5756,
        "step": 350
      },
      {
        "epoch": 5.010769230769231,
        "grad_norm": 3.829684257507324,
        "learning_rate": 9.880341880341882e-06,
        "loss": 0.4838,
        "step": 360
      },
      {
        "epoch": 5.013846153846154,
        "grad_norm": 6.597207546234131,
        "learning_rate": 9.846153846153848e-06,
        "loss": 0.5747,
        "step": 370
      },
      {
        "epoch": 5.016923076923077,
        "grad_norm": 15.660137176513672,
        "learning_rate": 9.811965811965812e-06,
        "loss": 0.5223,
        "step": 380
      },
      {
        "epoch": 5.02,
        "grad_norm": 13.29094123840332,
        "learning_rate": 9.777777777777779e-06,
        "loss": 0.438,
        "step": 390
      },
      {
        "epoch": 5.02,
        "eval_accuracy": 0.5434782608695652,
        "eval_loss": 1.2116292715072632,
        "eval_runtime": 10.5835,
        "eval_samples_per_second": 4.346,
        "eval_steps_per_second": 1.134,
        "step": 390
      },
      {
        "epoch": 6.003076923076923,
        "grad_norm": 10.139371871948242,
        "learning_rate": 9.743589743589744e-06,
        "loss": 0.7131,
        "step": 400
      },
      {
        "epoch": 6.006153846153846,
        "grad_norm": 15.849384307861328,
        "learning_rate": 9.70940170940171e-06,
        "loss": 0.5023,
        "step": 410
      },
      {
        "epoch": 6.009230769230769,
        "grad_norm": 5.638754844665527,
        "learning_rate": 9.675213675213677e-06,
        "loss": 0.4474,
        "step": 420
      },
      {
        "epoch": 6.012307692307692,
        "grad_norm": 25.2630672454834,
        "learning_rate": 9.641025641025642e-06,
        "loss": 0.5027,
        "step": 430
      },
      {
        "epoch": 6.015384615384615,
        "grad_norm": 13.194704055786133,
        "learning_rate": 9.606837606837607e-06,
        "loss": 0.4413,
        "step": 440
      },
      {
        "epoch": 6.018461538461539,
        "grad_norm": 24.00779151916504,
        "learning_rate": 9.572649572649575e-06,
        "loss": 0.4653,
        "step": 450
      },
      {
        "epoch": 6.02,
        "eval_accuracy": 0.5652173913043478,
        "eval_loss": 0.6001840829849243,
        "eval_runtime": 10.1529,
        "eval_samples_per_second": 4.531,
        "eval_steps_per_second": 1.182,
        "step": 455
      },
      {
        "epoch": 7.001538461538462,
        "grad_norm": 8.63178539276123,
        "learning_rate": 9.53846153846154e-06,
        "loss": 0.573,
        "step": 460
      },
      {
        "epoch": 7.004615384615384,
        "grad_norm": 26.25208282470703,
        "learning_rate": 9.504273504273504e-06,
        "loss": 0.5129,
        "step": 470
      },
      {
        "epoch": 7.007692307692308,
        "grad_norm": 6.66032600402832,
        "learning_rate": 9.470085470085471e-06,
        "loss": 0.7248,
        "step": 480
      },
      {
        "epoch": 7.010769230769231,
        "grad_norm": 3.6293530464172363,
        "learning_rate": 9.435897435897436e-06,
        "loss": 0.333,
        "step": 490
      },
      {
        "epoch": 7.013846153846154,
        "grad_norm": 6.718216419219971,
        "learning_rate": 9.401709401709402e-06,
        "loss": 0.6385,
        "step": 500
      },
      {
        "epoch": 7.016923076923077,
        "grad_norm": 15.813761711120605,
        "learning_rate": 9.367521367521369e-06,
        "loss": 0.6076,
        "step": 510
      },
      {
        "epoch": 7.02,
        "grad_norm": 9.687027931213379,
        "learning_rate": 9.333333333333334e-06,
        "loss": 0.5876,
        "step": 520
      },
      {
        "epoch": 7.02,
        "eval_accuracy": 0.8043478260869565,
        "eval_loss": 0.48940709233283997,
        "eval_runtime": 10.0977,
        "eval_samples_per_second": 4.555,
        "eval_steps_per_second": 1.188,
        "step": 520
      },
      {
        "epoch": 8.003076923076923,
        "grad_norm": 2.885645627975464,
        "learning_rate": 9.2991452991453e-06,
        "loss": 0.6155,
        "step": 530
      },
      {
        "epoch": 8.006153846153847,
        "grad_norm": 41.08794021606445,
        "learning_rate": 9.264957264957267e-06,
        "loss": 0.4796,
        "step": 540
      },
      {
        "epoch": 8.009230769230768,
        "grad_norm": 1.9967862367630005,
        "learning_rate": 9.230769230769232e-06,
        "loss": 0.333,
        "step": 550
      },
      {
        "epoch": 8.012307692307692,
        "grad_norm": 4.82252311706543,
        "learning_rate": 9.196581196581196e-06,
        "loss": 0.6022,
        "step": 560
      },
      {
        "epoch": 8.015384615384615,
        "grad_norm": 3.3050286769866943,
        "learning_rate": 9.162393162393163e-06,
        "loss": 0.4147,
        "step": 570
      },
      {
        "epoch": 8.018461538461539,
        "grad_norm": 0.10243837535381317,
        "learning_rate": 9.128205128205129e-06,
        "loss": 0.3801,
        "step": 580
      },
      {
        "epoch": 8.02,
        "eval_accuracy": 0.5434782608695652,
        "eval_loss": 0.8324462175369263,
        "eval_runtime": 10.0187,
        "eval_samples_per_second": 4.591,
        "eval_steps_per_second": 1.198,
        "step": 585
      },
      {
        "epoch": 9.00153846153846,
        "grad_norm": 4.7592668533325195,
        "learning_rate": 9.094017094017094e-06,
        "loss": 1.0305,
        "step": 590
      },
      {
        "epoch": 9.004615384615384,
        "grad_norm": 8.245183944702148,
        "learning_rate": 9.059829059829061e-06,
        "loss": 0.5193,
        "step": 600
      },
      {
        "epoch": 9.007692307692308,
        "grad_norm": 24.77902603149414,
        "learning_rate": 9.025641025641027e-06,
        "loss": 0.4919,
        "step": 610
      },
      {
        "epoch": 9.010769230769231,
        "grad_norm": 29.81329345703125,
        "learning_rate": 8.991452991452992e-06,
        "loss": 0.6428,
        "step": 620
      },
      {
        "epoch": 9.013846153846155,
        "grad_norm": 8.413053512573242,
        "learning_rate": 8.957264957264959e-06,
        "loss": 0.3073,
        "step": 630
      },
      {
        "epoch": 9.016923076923076,
        "grad_norm": 8.200304985046387,
        "learning_rate": 8.923076923076925e-06,
        "loss": 0.6606,
        "step": 640
      },
      {
        "epoch": 9.02,
        "grad_norm": 22.091062545776367,
        "learning_rate": 8.888888888888888e-06,
        "loss": 0.4474,
        "step": 650
      },
      {
        "epoch": 9.02,
        "eval_accuracy": 0.5652173913043478,
        "eval_loss": 1.158095359802246,
        "eval_runtime": 10.3694,
        "eval_samples_per_second": 4.436,
        "eval_steps_per_second": 1.157,
        "step": 650
      },
      {
        "epoch": 10.003076923076923,
        "grad_norm": 4.3752288818359375,
        "learning_rate": 8.854700854700855e-06,
        "loss": 1.0286,
        "step": 660
      },
      {
        "epoch": 10.006153846153847,
        "grad_norm": 5.74050235748291,
        "learning_rate": 8.820512820512821e-06,
        "loss": 0.4695,
        "step": 670
      },
      {
        "epoch": 10.009230769230768,
        "grad_norm": 4.538609504699707,
        "learning_rate": 8.786324786324786e-06,
        "loss": 0.3401,
        "step": 680
      },
      {
        "epoch": 10.012307692307692,
        "grad_norm": 6.925298690795898,
        "learning_rate": 8.752136752136753e-06,
        "loss": 0.616,
        "step": 690
      },
      {
        "epoch": 10.015384615384615,
        "grad_norm": 5.259721755981445,
        "learning_rate": 8.717948717948719e-06,
        "loss": 0.3077,
        "step": 700
      },
      {
        "epoch": 10.018461538461539,
        "grad_norm": 18.261491775512695,
        "learning_rate": 8.683760683760684e-06,
        "loss": 0.694,
        "step": 710
      },
      {
        "epoch": 10.02,
        "eval_accuracy": 0.717391304347826,
        "eval_loss": 0.5354219675064087,
        "eval_runtime": 10.1729,
        "eval_samples_per_second": 4.522,
        "eval_steps_per_second": 1.18,
        "step": 715
      },
      {
        "epoch": 11.00153846153846,
        "grad_norm": 6.640218257904053,
        "learning_rate": 8.649572649572651e-06,
        "loss": 0.5996,
        "step": 720
      },
      {
        "epoch": 11.004615384615384,
        "grad_norm": 3.7313828468322754,
        "learning_rate": 8.615384615384617e-06,
        "loss": 0.2938,
        "step": 730
      },
      {
        "epoch": 11.007692307692308,
        "grad_norm": 2.5741147994995117,
        "learning_rate": 8.58119658119658e-06,
        "loss": 0.559,
        "step": 740
      },
      {
        "epoch": 11.010769230769231,
        "grad_norm": 4.06730318069458,
        "learning_rate": 8.547008547008548e-06,
        "loss": 0.3129,
        "step": 750
      },
      {
        "epoch": 11.013846153846155,
        "grad_norm": 14.723116874694824,
        "learning_rate": 8.512820512820513e-06,
        "loss": 0.5298,
        "step": 760
      },
      {
        "epoch": 11.016923076923076,
        "grad_norm": 11.420027732849121,
        "learning_rate": 8.478632478632479e-06,
        "loss": 0.4968,
        "step": 770
      },
      {
        "epoch": 11.02,
        "grad_norm": 12.085827827453613,
        "learning_rate": 8.444444444444446e-06,
        "loss": 0.4773,
        "step": 780
      },
      {
        "epoch": 11.02,
        "eval_accuracy": 0.6956521739130435,
        "eval_loss": 0.6181442141532898,
        "eval_runtime": 10.1957,
        "eval_samples_per_second": 4.512,
        "eval_steps_per_second": 1.177,
        "step": 780
      },
      {
        "epoch": 12.003076923076923,
        "grad_norm": 27.175369262695312,
        "learning_rate": 8.410256410256411e-06,
        "loss": 0.37,
        "step": 790
      },
      {
        "epoch": 12.006153846153847,
        "grad_norm": 6.941166400909424,
        "learning_rate": 8.376068376068377e-06,
        "loss": 0.5221,
        "step": 800
      },
      {
        "epoch": 12.009230769230768,
        "grad_norm": 17.430849075317383,
        "learning_rate": 8.341880341880344e-06,
        "loss": 0.3823,
        "step": 810
      },
      {
        "epoch": 12.012307692307692,
        "grad_norm": 0.49046531319618225,
        "learning_rate": 8.307692307692309e-06,
        "loss": 0.5028,
        "step": 820
      },
      {
        "epoch": 12.015384615384615,
        "grad_norm": 15.024535179138184,
        "learning_rate": 8.273504273504273e-06,
        "loss": 0.5115,
        "step": 830
      },
      {
        "epoch": 12.018461538461539,
        "grad_norm": 14.742912292480469,
        "learning_rate": 8.23931623931624e-06,
        "loss": 0.6208,
        "step": 840
      },
      {
        "epoch": 12.02,
        "eval_accuracy": 0.7608695652173914,
        "eval_loss": 0.567747950553894,
        "eval_runtime": 10.3753,
        "eval_samples_per_second": 4.434,
        "eval_steps_per_second": 1.157,
        "step": 845
      },
      {
        "epoch": 13.00153846153846,
        "grad_norm": 15.91428279876709,
        "learning_rate": 8.205128205128205e-06,
        "loss": 0.4844,
        "step": 850
      },
      {
        "epoch": 13.004615384615384,
        "grad_norm": 0.08476120978593826,
        "learning_rate": 8.17094017094017e-06,
        "loss": 0.2531,
        "step": 860
      },
      {
        "epoch": 13.007692307692308,
        "grad_norm": 59.409690856933594,
        "learning_rate": 8.136752136752138e-06,
        "loss": 0.6488,
        "step": 870
      },
      {
        "epoch": 13.010769230769231,
        "grad_norm": 3.901467800140381,
        "learning_rate": 8.102564102564103e-06,
        "loss": 0.8015,
        "step": 880
      },
      {
        "epoch": 13.013846153846155,
        "grad_norm": 6.678008556365967,
        "learning_rate": 8.068376068376069e-06,
        "loss": 0.5357,
        "step": 890
      },
      {
        "epoch": 13.016923076923076,
        "grad_norm": 7.264644622802734,
        "learning_rate": 8.034188034188036e-06,
        "loss": 0.4998,
        "step": 900
      },
      {
        "epoch": 13.02,
        "grad_norm": 22.3366756439209,
        "learning_rate": 8.000000000000001e-06,
        "loss": 0.344,
        "step": 910
      },
      {
        "epoch": 13.02,
        "eval_accuracy": 0.6086956521739131,
        "eval_loss": 0.7451769113540649,
        "eval_runtime": 10.4561,
        "eval_samples_per_second": 4.399,
        "eval_steps_per_second": 1.148,
        "step": 910
      },
      {
        "epoch": 14.003076923076923,
        "grad_norm": 17.2800350189209,
        "learning_rate": 7.965811965811967e-06,
        "loss": 0.3782,
        "step": 920
      },
      {
        "epoch": 14.006153846153847,
        "grad_norm": 2.800379753112793,
        "learning_rate": 7.931623931623932e-06,
        "loss": 0.7975,
        "step": 930
      },
      {
        "epoch": 14.009230769230768,
        "grad_norm": 14.407361030578613,
        "learning_rate": 7.897435897435898e-06,
        "loss": 0.4081,
        "step": 940
      },
      {
        "epoch": 14.012307692307692,
        "grad_norm": 82.17062377929688,
        "learning_rate": 7.863247863247863e-06,
        "loss": 0.3227,
        "step": 950
      },
      {
        "epoch": 14.015384615384615,
        "grad_norm": 20.13490867614746,
        "learning_rate": 7.82905982905983e-06,
        "loss": 0.3738,
        "step": 960
      },
      {
        "epoch": 14.018461538461539,
        "grad_norm": 30.13092041015625,
        "learning_rate": 7.794871794871796e-06,
        "loss": 0.254,
        "step": 970
      },
      {
        "epoch": 14.02,
        "eval_accuracy": 0.7391304347826086,
        "eval_loss": 0.6362163424491882,
        "eval_runtime": 10.2582,
        "eval_samples_per_second": 4.484,
        "eval_steps_per_second": 1.17,
        "step": 975
      },
      {
        "epoch": 15.00153846153846,
        "grad_norm": 0.13770069181919098,
        "learning_rate": 7.760683760683761e-06,
        "loss": 0.4847,
        "step": 980
      },
      {
        "epoch": 15.004615384615384,
        "grad_norm": 17.021846771240234,
        "learning_rate": 7.726495726495728e-06,
        "loss": 0.2376,
        "step": 990
      },
      {
        "epoch": 15.007692307692308,
        "grad_norm": 16.193130493164062,
        "learning_rate": 7.692307692307694e-06,
        "loss": 0.5756,
        "step": 1000
      },
      {
        "epoch": 15.010769230769231,
        "grad_norm": 17.448122024536133,
        "learning_rate": 7.658119658119659e-06,
        "loss": 0.5613,
        "step": 1010
      },
      {
        "epoch": 15.013846153846155,
        "grad_norm": 27.816417694091797,
        "learning_rate": 7.6239316239316244e-06,
        "loss": 0.3936,
        "step": 1020
      },
      {
        "epoch": 15.016923076923076,
        "grad_norm": 9.32864761352539,
        "learning_rate": 7.58974358974359e-06,
        "loss": 0.5214,
        "step": 1030
      },
      {
        "epoch": 15.02,
        "grad_norm": 10.375736236572266,
        "learning_rate": 7.555555555555556e-06,
        "loss": 0.4578,
        "step": 1040
      },
      {
        "epoch": 15.02,
        "eval_accuracy": 0.6956521739130435,
        "eval_loss": 0.8304260969161987,
        "eval_runtime": 10.454,
        "eval_samples_per_second": 4.4,
        "eval_steps_per_second": 1.148,
        "step": 1040
      },
      {
        "epoch": 16.00307692307692,
        "grad_norm": 32.06553268432617,
        "learning_rate": 7.521367521367522e-06,
        "loss": 0.4237,
        "step": 1050
      },
      {
        "epoch": 16.006153846153847,
        "grad_norm": 3.9894378185272217,
        "learning_rate": 7.487179487179488e-06,
        "loss": 0.3599,
        "step": 1060
      },
      {
        "epoch": 16.00923076923077,
        "grad_norm": 0.25729474425315857,
        "learning_rate": 7.452991452991454e-06,
        "loss": 0.3732,
        "step": 1070
      },
      {
        "epoch": 16.012307692307694,
        "grad_norm": 16.2904109954834,
        "learning_rate": 7.4188034188034196e-06,
        "loss": 0.3938,
        "step": 1080
      },
      {
        "epoch": 16.015384615384615,
        "grad_norm": 10.762846946716309,
        "learning_rate": 7.384615384615386e-06,
        "loss": 0.312,
        "step": 1090
      },
      {
        "epoch": 16.018461538461537,
        "grad_norm": 20.59312629699707,
        "learning_rate": 7.350427350427351e-06,
        "loss": 0.3954,
        "step": 1100
      },
      {
        "epoch": 16.02,
        "eval_accuracy": 0.7608695652173914,
        "eval_loss": 0.6048559546470642,
        "eval_runtime": 10.7124,
        "eval_samples_per_second": 4.294,
        "eval_steps_per_second": 1.12,
        "step": 1105
      },
      {
        "epoch": 17.001538461538463,
        "grad_norm": 11.256546020507812,
        "learning_rate": 7.316239316239317e-06,
        "loss": 0.4876,
        "step": 1110
      },
      {
        "epoch": 17.004615384615384,
        "grad_norm": 42.64373016357422,
        "learning_rate": 7.282051282051282e-06,
        "loss": 0.1782,
        "step": 1120
      },
      {
        "epoch": 17.00769230769231,
        "grad_norm": 34.2502326965332,
        "learning_rate": 7.247863247863248e-06,
        "loss": 0.4987,
        "step": 1130
      },
      {
        "epoch": 17.01076923076923,
        "grad_norm": 9.817243576049805,
        "learning_rate": 7.213675213675214e-06,
        "loss": 0.2962,
        "step": 1140
      },
      {
        "epoch": 17.013846153846153,
        "grad_norm": 39.85562515258789,
        "learning_rate": 7.17948717948718e-06,
        "loss": 0.4634,
        "step": 1150
      },
      {
        "epoch": 17.016923076923078,
        "grad_norm": 32.21989059448242,
        "learning_rate": 7.145299145299146e-06,
        "loss": 0.2256,
        "step": 1160
      },
      {
        "epoch": 17.02,
        "grad_norm": 32.11091232299805,
        "learning_rate": 7.111111111111112e-06,
        "loss": 0.248,
        "step": 1170
      },
      {
        "epoch": 17.02,
        "eval_accuracy": 0.6739130434782609,
        "eval_loss": 0.9506217241287231,
        "eval_runtime": 10.8727,
        "eval_samples_per_second": 4.231,
        "eval_steps_per_second": 1.104,
        "step": 1170
      },
      {
        "epoch": 18.00307692307692,
        "grad_norm": 48.11212158203125,
        "learning_rate": 7.076923076923078e-06,
        "loss": 0.7942,
        "step": 1180
      },
      {
        "epoch": 18.006153846153847,
        "grad_norm": 39.191593170166016,
        "learning_rate": 7.0427350427350435e-06,
        "loss": 0.4622,
        "step": 1190
      },
      {
        "epoch": 18.00923076923077,
        "grad_norm": 27.27468490600586,
        "learning_rate": 7.008547008547009e-06,
        "loss": 0.4094,
        "step": 1200
      },
      {
        "epoch": 18.012307692307694,
        "grad_norm": 13.529548645019531,
        "learning_rate": 6.974358974358974e-06,
        "loss": 0.137,
        "step": 1210
      },
      {
        "epoch": 18.015384615384615,
        "grad_norm": 2.5885872840881348,
        "learning_rate": 6.940170940170941e-06,
        "loss": 0.3264,
        "step": 1220
      },
      {
        "epoch": 18.018461538461537,
        "grad_norm": 51.945762634277344,
        "learning_rate": 6.905982905982906e-06,
        "loss": 0.1334,
        "step": 1230
      },
      {
        "epoch": 18.02,
        "eval_accuracy": 0.6739130434782609,
        "eval_loss": 1.1876178979873657,
        "eval_runtime": 10.6484,
        "eval_samples_per_second": 4.32,
        "eval_steps_per_second": 1.127,
        "step": 1235
      },
      {
        "epoch": 19.001538461538463,
        "grad_norm": 0.604591965675354,
        "learning_rate": 6.871794871794872e-06,
        "loss": 0.1617,
        "step": 1240
      },
      {
        "epoch": 19.004615384615384,
        "grad_norm": 65.89838409423828,
        "learning_rate": 6.837606837606839e-06,
        "loss": 0.5149,
        "step": 1250
      },
      {
        "epoch": 19.00769230769231,
        "grad_norm": 103.18575286865234,
        "learning_rate": 6.803418803418804e-06,
        "loss": 0.299,
        "step": 1260
      },
      {
        "epoch": 19.01076923076923,
        "grad_norm": 0.5941734313964844,
        "learning_rate": 6.76923076923077e-06,
        "loss": 0.5991,
        "step": 1270
      },
      {
        "epoch": 19.013846153846153,
        "grad_norm": 22.712234497070312,
        "learning_rate": 6.735042735042736e-06,
        "loss": 0.3312,
        "step": 1280
      },
      {
        "epoch": 19.016923076923078,
        "grad_norm": 0.26104092597961426,
        "learning_rate": 6.700854700854701e-06,
        "loss": 0.3076,
        "step": 1290
      },
      {
        "epoch": 19.02,
        "grad_norm": 58.68645477294922,
        "learning_rate": 6.666666666666667e-06,
        "loss": 0.534,
        "step": 1300
      },
      {
        "epoch": 19.02,
        "eval_accuracy": 0.7391304347826086,
        "eval_loss": 0.6295628547668457,
        "eval_runtime": 11.1938,
        "eval_samples_per_second": 4.109,
        "eval_steps_per_second": 1.072,
        "step": 1300
      },
      {
        "epoch": 20.00307692307692,
        "grad_norm": 0.13804925978183746,
        "learning_rate": 6.632478632478633e-06,
        "loss": 0.4236,
        "step": 1310
      },
      {
        "epoch": 20.006153846153847,
        "grad_norm": 0.2732415497303009,
        "learning_rate": 6.598290598290598e-06,
        "loss": 0.2337,
        "step": 1320
      },
      {
        "epoch": 20.00923076923077,
        "grad_norm": 0.22181116044521332,
        "learning_rate": 6.564102564102565e-06,
        "loss": 0.2577,
        "step": 1330
      },
      {
        "epoch": 20.012307692307694,
        "grad_norm": 20.500429153442383,
        "learning_rate": 6.529914529914531e-06,
        "loss": 0.3988,
        "step": 1340
      },
      {
        "epoch": 20.015384615384615,
        "grad_norm": 0.25182831287384033,
        "learning_rate": 6.495726495726496e-06,
        "loss": 0.3886,
        "step": 1350
      },
      {
        "epoch": 20.018461538461537,
        "grad_norm": 29.723634719848633,
        "learning_rate": 6.461538461538463e-06,
        "loss": 0.3556,
        "step": 1360
      },
      {
        "epoch": 20.02,
        "eval_accuracy": 0.6956521739130435,
        "eval_loss": 1.3006523847579956,
        "eval_runtime": 10.3551,
        "eval_samples_per_second": 4.442,
        "eval_steps_per_second": 1.159,
        "step": 1365
      },
      {
        "epoch": 21.001538461538463,
        "grad_norm": 30.7800350189209,
        "learning_rate": 6.427350427350428e-06,
        "loss": 0.6987,
        "step": 1370
      },
      {
        "epoch": 21.004615384615384,
        "grad_norm": 13.209580421447754,
        "learning_rate": 6.3931623931623935e-06,
        "loss": 0.1747,
        "step": 1380
      },
      {
        "epoch": 21.00769230769231,
        "grad_norm": 74.4799575805664,
        "learning_rate": 6.358974358974359e-06,
        "loss": 0.2258,
        "step": 1390
      },
      {
        "epoch": 21.01076923076923,
        "grad_norm": 87.08421325683594,
        "learning_rate": 6.324786324786325e-06,
        "loss": 0.3794,
        "step": 1400
      },
      {
        "epoch": 21.013846153846153,
        "grad_norm": 35.24946594238281,
        "learning_rate": 6.290598290598291e-06,
        "loss": 0.5507,
        "step": 1410
      },
      {
        "epoch": 21.016923076923078,
        "grad_norm": 0.3907718360424042,
        "learning_rate": 6.256410256410257e-06,
        "loss": 0.3903,
        "step": 1420
      },
      {
        "epoch": 21.02,
        "grad_norm": 67.88304901123047,
        "learning_rate": 6.222222222222223e-06,
        "loss": 0.5439,
        "step": 1430
      },
      {
        "epoch": 21.02,
        "eval_accuracy": 0.6739130434782609,
        "eval_loss": 1.5065968036651611,
        "eval_runtime": 10.2377,
        "eval_samples_per_second": 4.493,
        "eval_steps_per_second": 1.172,
        "step": 1430
      },
      {
        "epoch": 22.00307692307692,
        "grad_norm": 0.019926143810153008,
        "learning_rate": 6.188034188034189e-06,
        "loss": 0.4587,
        "step": 1440
      },
      {
        "epoch": 22.006153846153847,
        "grad_norm": 61.04457092285156,
        "learning_rate": 6.153846153846155e-06,
        "loss": 0.0936,
        "step": 1450
      },
      {
        "epoch": 22.00923076923077,
        "grad_norm": 69.38724517822266,
        "learning_rate": 6.11965811965812e-06,
        "loss": 0.394,
        "step": 1460
      },
      {
        "epoch": 22.012307692307694,
        "grad_norm": 58.5753173828125,
        "learning_rate": 6.085470085470086e-06,
        "loss": 0.2639,
        "step": 1470
      },
      {
        "epoch": 22.015384615384615,
        "grad_norm": 0.5316572785377502,
        "learning_rate": 6.051282051282051e-06,
        "loss": 0.476,
        "step": 1480
      },
      {
        "epoch": 22.018461538461537,
        "grad_norm": 29.042409896850586,
        "learning_rate": 6.0170940170940174e-06,
        "loss": 0.4107,
        "step": 1490
      },
      {
        "epoch": 22.02,
        "eval_accuracy": 0.8043478260869565,
        "eval_loss": 0.9273331761360168,
        "eval_runtime": 10.2871,
        "eval_samples_per_second": 4.472,
        "eval_steps_per_second": 1.167,
        "step": 1495
      },
      {
        "epoch": 23.001538461538463,
        "grad_norm": 0.7934656143188477,
        "learning_rate": 5.982905982905983e-06,
        "loss": 0.4271,
        "step": 1500
      },
      {
        "epoch": 23.004615384615384,
        "grad_norm": 15.743200302124023,
        "learning_rate": 5.948717948717949e-06,
        "loss": 0.2889,
        "step": 1510
      },
      {
        "epoch": 23.00769230769231,
        "grad_norm": 0.05252040550112724,
        "learning_rate": 5.914529914529915e-06,
        "loss": 0.1282,
        "step": 1520
      },
      {
        "epoch": 23.01076923076923,
        "grad_norm": 27.15765953063965,
        "learning_rate": 5.880341880341881e-06,
        "loss": 0.455,
        "step": 1530
      },
      {
        "epoch": 23.013846153846153,
        "grad_norm": 154.0335235595703,
        "learning_rate": 5.846153846153847e-06,
        "loss": 0.5051,
        "step": 1540
      },
      {
        "epoch": 23.016923076923078,
        "grad_norm": 1.5758213996887207,
        "learning_rate": 5.8119658119658126e-06,
        "loss": 0.4148,
        "step": 1550
      },
      {
        "epoch": 23.02,
        "grad_norm": 46.614864349365234,
        "learning_rate": 5.777777777777778e-06,
        "loss": 0.61,
        "step": 1560
      },
      {
        "epoch": 23.02,
        "eval_accuracy": 0.717391304347826,
        "eval_loss": 1.00082528591156,
        "eval_runtime": 10.3956,
        "eval_samples_per_second": 4.425,
        "eval_steps_per_second": 1.154,
        "step": 1560
      },
      {
        "epoch": 24.00307692307692,
        "grad_norm": 1.2954469919204712,
        "learning_rate": 5.743589743589743e-06,
        "loss": 0.3196,
        "step": 1570
      },
      {
        "epoch": 24.006153846153847,
        "grad_norm": 0.4691097140312195,
        "learning_rate": 5.70940170940171e-06,
        "loss": 0.1125,
        "step": 1580
      },
      {
        "epoch": 24.00923076923077,
        "grad_norm": 66.216552734375,
        "learning_rate": 5.675213675213675e-06,
        "loss": 0.3891,
        "step": 1590
      },
      {
        "epoch": 24.012307692307694,
        "grad_norm": 0.1557655781507492,
        "learning_rate": 5.641025641025641e-06,
        "loss": 0.03,
        "step": 1600
      },
      {
        "epoch": 24.015384615384615,
        "grad_norm": 2.08573842048645,
        "learning_rate": 5.606837606837608e-06,
        "loss": 0.3544,
        "step": 1610
      },
      {
        "epoch": 24.018461538461537,
        "grad_norm": 2.804753303527832,
        "learning_rate": 5.572649572649573e-06,
        "loss": 0.6482,
        "step": 1620
      },
      {
        "epoch": 24.02,
        "eval_accuracy": 0.7608695652173914,
        "eval_loss": 0.7548108100891113,
        "eval_runtime": 10.8151,
        "eval_samples_per_second": 4.253,
        "eval_steps_per_second": 1.11,
        "step": 1625
      },
      {
        "epoch": 25.001538461538463,
        "grad_norm": 83.31999969482422,
        "learning_rate": 5.538461538461539e-06,
        "loss": 0.2905,
        "step": 1630
      },
      {
        "epoch": 25.004615384615384,
        "grad_norm": 23.389434814453125,
        "learning_rate": 5.504273504273505e-06,
        "loss": 0.438,
        "step": 1640
      },
      {
        "epoch": 25.00769230769231,
        "grad_norm": 34.89247512817383,
        "learning_rate": 5.470085470085471e-06,
        "loss": 0.0759,
        "step": 1650
      },
      {
        "epoch": 25.01076923076923,
        "grad_norm": 0.49820569157600403,
        "learning_rate": 5.435897435897436e-06,
        "loss": 0.1849,
        "step": 1660
      },
      {
        "epoch": 25.013846153846153,
        "grad_norm": 0.12784357368946075,
        "learning_rate": 5.401709401709402e-06,
        "loss": 0.259,
        "step": 1670
      },
      {
        "epoch": 25.016923076923078,
        "grad_norm": 69.064208984375,
        "learning_rate": 5.367521367521367e-06,
        "loss": 0.3631,
        "step": 1680
      },
      {
        "epoch": 25.02,
        "grad_norm": 0.5186673402786255,
        "learning_rate": 5.333333333333334e-06,
        "loss": 0.199,
        "step": 1690
      },
      {
        "epoch": 25.02,
        "eval_accuracy": 0.782608695652174,
        "eval_loss": 0.7916727662086487,
        "eval_runtime": 10.4377,
        "eval_samples_per_second": 4.407,
        "eval_steps_per_second": 1.15,
        "step": 1690
      },
      {
        "epoch": 26.00307692307692,
        "grad_norm": 0.019262276589870453,
        "learning_rate": 5.2991452991453e-06,
        "loss": 0.2937,
        "step": 1700
      },
      {
        "epoch": 26.006153846153847,
        "grad_norm": 0.6572211384773254,
        "learning_rate": 5.264957264957265e-06,
        "loss": 0.1816,
        "step": 1710
      },
      {
        "epoch": 26.00923076923077,
        "grad_norm": 52.68994903564453,
        "learning_rate": 5.230769230769232e-06,
        "loss": 0.4716,
        "step": 1720
      },
      {
        "epoch": 26.012307692307694,
        "grad_norm": 0.405513733625412,
        "learning_rate": 5.196581196581197e-06,
        "loss": 0.4085,
        "step": 1730
      },
      {
        "epoch": 26.015384615384615,
        "grad_norm": 16.66892433166504,
        "learning_rate": 5.162393162393163e-06,
        "loss": 0.4931,
        "step": 1740
      },
      {
        "epoch": 26.018461538461537,
        "grad_norm": 0.020136963576078415,
        "learning_rate": 5.128205128205128e-06,
        "loss": 0.1185,
        "step": 1750
      },
      {
        "epoch": 26.02,
        "eval_accuracy": 0.782608695652174,
        "eval_loss": 0.7529156804084778,
        "eval_runtime": 10.9054,
        "eval_samples_per_second": 4.218,
        "eval_steps_per_second": 1.1,
        "step": 1755
      },
      {
        "epoch": 27.001538461538463,
        "grad_norm": 29.928010940551758,
        "learning_rate": 5.094017094017094e-06,
        "loss": 0.2469,
        "step": 1760
      },
      {
        "epoch": 27.004615384615384,
        "grad_norm": 19.83055877685547,
        "learning_rate": 5.05982905982906e-06,
        "loss": 0.2104,
        "step": 1770
      },
      {
        "epoch": 27.00769230769231,
        "grad_norm": 0.28511443734169006,
        "learning_rate": 5.025641025641026e-06,
        "loss": 0.2268,
        "step": 1780
      },
      {
        "epoch": 27.01076923076923,
        "grad_norm": 181.45603942871094,
        "learning_rate": 4.991452991452992e-06,
        "loss": 0.1389,
        "step": 1790
      },
      {
        "epoch": 27.013846153846153,
        "grad_norm": 64.7826156616211,
        "learning_rate": 4.957264957264958e-06,
        "loss": 0.1211,
        "step": 1800
      },
      {
        "epoch": 27.016923076923078,
        "grad_norm": 4.411494255065918,
        "learning_rate": 4.923076923076924e-06,
        "loss": 0.1387,
        "step": 1810
      },
      {
        "epoch": 27.02,
        "grad_norm": 11.255193710327148,
        "learning_rate": 4.888888888888889e-06,
        "loss": 0.3886,
        "step": 1820
      },
      {
        "epoch": 27.02,
        "eval_accuracy": 0.7608695652173914,
        "eval_loss": 0.8627105355262756,
        "eval_runtime": 10.4057,
        "eval_samples_per_second": 4.421,
        "eval_steps_per_second": 1.153,
        "step": 1820
      },
      {
        "epoch": 28.00307692307692,
        "grad_norm": 92.7336196899414,
        "learning_rate": 4.854700854700855e-06,
        "loss": 0.0313,
        "step": 1830
      },
      {
        "epoch": 28.006153846153847,
        "grad_norm": 21.0554256439209,
        "learning_rate": 4.820512820512821e-06,
        "loss": 0.0353,
        "step": 1840
      },
      {
        "epoch": 28.00923076923077,
        "grad_norm": 141.20700073242188,
        "learning_rate": 4.786324786324787e-06,
        "loss": 0.5025,
        "step": 1850
      },
      {
        "epoch": 28.012307692307694,
        "grad_norm": 0.3742455542087555,
        "learning_rate": 4.752136752136752e-06,
        "loss": 0.1276,
        "step": 1860
      },
      {
        "epoch": 28.015384615384615,
        "grad_norm": 0.04171011224389076,
        "learning_rate": 4.717948717948718e-06,
        "loss": 0.0059,
        "step": 1870
      },
      {
        "epoch": 28.018461538461537,
        "grad_norm": 2.4367685317993164,
        "learning_rate": 4.6837606837606844e-06,
        "loss": 0.0123,
        "step": 1880
      },
      {
        "epoch": 28.02,
        "eval_accuracy": 0.717391304347826,
        "eval_loss": 1.3886178731918335,
        "eval_runtime": 10.496,
        "eval_samples_per_second": 4.383,
        "eval_steps_per_second": 1.143,
        "step": 1885
      },
      {
        "epoch": 29.001538461538463,
        "grad_norm": 223.50958251953125,
        "learning_rate": 4.64957264957265e-06,
        "loss": 0.1912,
        "step": 1890
      },
      {
        "epoch": 29.004615384615384,
        "grad_norm": 0.6573296189308167,
        "learning_rate": 4.615384615384616e-06,
        "loss": 0.2606,
        "step": 1900
      },
      {
        "epoch": 29.00769230769231,
        "grad_norm": 33.679935455322266,
        "learning_rate": 4.581196581196582e-06,
        "loss": 0.1504,
        "step": 1910
      },
      {
        "epoch": 29.01076923076923,
        "grad_norm": 36.40077209472656,
        "learning_rate": 4.547008547008547e-06,
        "loss": 0.1491,
        "step": 1920
      },
      {
        "epoch": 29.013846153846153,
        "grad_norm": 34.2191047668457,
        "learning_rate": 4.512820512820513e-06,
        "loss": 0.2882,
        "step": 1930
      },
      {
        "epoch": 29.016923076923078,
        "grad_norm": 0.30137479305267334,
        "learning_rate": 4.4786324786324796e-06,
        "loss": 0.2504,
        "step": 1940
      },
      {
        "epoch": 29.02,
        "grad_norm": 0.040065910667181015,
        "learning_rate": 4.444444444444444e-06,
        "loss": 0.5328,
        "step": 1950
      },
      {
        "epoch": 29.02,
        "eval_accuracy": 0.6956521739130435,
        "eval_loss": 1.2802809476852417,
        "eval_runtime": 10.2626,
        "eval_samples_per_second": 4.482,
        "eval_steps_per_second": 1.169,
        "step": 1950
      },
      {
        "epoch": 30.00307692307692,
        "grad_norm": 0.04048765078186989,
        "learning_rate": 4.4102564102564104e-06,
        "loss": 0.0121,
        "step": 1960
      },
      {
        "epoch": 30.006153846153847,
        "grad_norm": 198.26219177246094,
        "learning_rate": 4.376068376068377e-06,
        "loss": 0.2916,
        "step": 1970
      },
      {
        "epoch": 30.00923076923077,
        "grad_norm": 0.01023260410875082,
        "learning_rate": 4.341880341880342e-06,
        "loss": 0.1678,
        "step": 1980
      },
      {
        "epoch": 30.012307692307694,
        "grad_norm": 0.014997897669672966,
        "learning_rate": 4.307692307692308e-06,
        "loss": 0.3871,
        "step": 1990
      },
      {
        "epoch": 30.015384615384615,
        "grad_norm": 0.48902782797813416,
        "learning_rate": 4.273504273504274e-06,
        "loss": 0.1938,
        "step": 2000
      },
      {
        "epoch": 30.018461538461537,
        "grad_norm": 0.6367323994636536,
        "learning_rate": 4.239316239316239e-06,
        "loss": 0.2961,
        "step": 2010
      },
      {
        "epoch": 30.02,
        "eval_accuracy": 0.717391304347826,
        "eval_loss": 1.439742922782898,
        "eval_runtime": 10.3024,
        "eval_samples_per_second": 4.465,
        "eval_steps_per_second": 1.165,
        "step": 2015
      },
      {
        "epoch": 31.001538461538463,
        "grad_norm": 1.0408298969268799,
        "learning_rate": 4.2051282051282055e-06,
        "loss": 0.0824,
        "step": 2020
      },
      {
        "epoch": 31.004615384615384,
        "grad_norm": 7.177045822143555,
        "learning_rate": 4.170940170940172e-06,
        "loss": 0.1443,
        "step": 2030
      },
      {
        "epoch": 31.00769230769231,
        "grad_norm": 0.5223032832145691,
        "learning_rate": 4.136752136752136e-06,
        "loss": 0.5326,
        "step": 2040
      },
      {
        "epoch": 31.01076923076923,
        "grad_norm": 0.12590403854846954,
        "learning_rate": 4.102564102564103e-06,
        "loss": 0.2971,
        "step": 2050
      },
      {
        "epoch": 31.013846153846153,
        "grad_norm": 0.43242764472961426,
        "learning_rate": 4.068376068376069e-06,
        "loss": 0.2301,
        "step": 2060
      },
      {
        "epoch": 31.016923076923078,
        "grad_norm": 0.03505908325314522,
        "learning_rate": 4.034188034188034e-06,
        "loss": 0.019,
        "step": 2070
      },
      {
        "epoch": 31.02,
        "grad_norm": 5.052464962005615,
        "learning_rate": 4.000000000000001e-06,
        "loss": 0.1192,
        "step": 2080
      },
      {
        "epoch": 31.02,
        "eval_accuracy": 0.6304347826086957,
        "eval_loss": 2.2563250064849854,
        "eval_runtime": 10.3617,
        "eval_samples_per_second": 4.439,
        "eval_steps_per_second": 1.158,
        "step": 2080
      },
      {
        "epoch": 32.003076923076925,
        "grad_norm": 138.01513671875,
        "learning_rate": 3.965811965811966e-06,
        "loss": 0.3484,
        "step": 2090
      },
      {
        "epoch": 32.00615384615384,
        "grad_norm": 0.5178282260894775,
        "learning_rate": 3.9316239316239315e-06,
        "loss": 0.0038,
        "step": 2100
      },
      {
        "epoch": 32.00923076923077,
        "grad_norm": 81.31842041015625,
        "learning_rate": 3.897435897435898e-06,
        "loss": 0.1732,
        "step": 2110
      },
      {
        "epoch": 32.012307692307694,
        "grad_norm": 0.07390841841697693,
        "learning_rate": 3.863247863247864e-06,
        "loss": 0.2652,
        "step": 2120
      },
      {
        "epoch": 32.01538461538462,
        "grad_norm": 71.98724365234375,
        "learning_rate": 3.8290598290598295e-06,
        "loss": 0.4317,
        "step": 2130
      },
      {
        "epoch": 32.01846153846154,
        "grad_norm": 18.120813369750977,
        "learning_rate": 3.794871794871795e-06,
        "loss": 0.145,
        "step": 2140
      },
      {
        "epoch": 32.02,
        "eval_accuracy": 0.7608695652173914,
        "eval_loss": 1.0464895963668823,
        "eval_runtime": 10.3727,
        "eval_samples_per_second": 4.435,
        "eval_steps_per_second": 1.157,
        "step": 2145
      },
      {
        "epoch": 33.00153846153846,
        "grad_norm": 0.1161544919013977,
        "learning_rate": 3.760683760683761e-06,
        "loss": 0.0155,
        "step": 2150
      },
      {
        "epoch": 33.004615384615384,
        "grad_norm": 0.4286612868309021,
        "learning_rate": 3.726495726495727e-06,
        "loss": 0.3714,
        "step": 2160
      },
      {
        "epoch": 33.00769230769231,
        "grad_norm": 0.14987812936306,
        "learning_rate": 3.692307692307693e-06,
        "loss": 0.2685,
        "step": 2170
      },
      {
        "epoch": 33.01076923076923,
        "grad_norm": 0.08653810620307922,
        "learning_rate": 3.6581196581196584e-06,
        "loss": 0.0845,
        "step": 2180
      },
      {
        "epoch": 33.01384615384615,
        "grad_norm": 143.28871154785156,
        "learning_rate": 3.623931623931624e-06,
        "loss": 0.2904,
        "step": 2190
      },
      {
        "epoch": 33.01692307692308,
        "grad_norm": 0.008105006068944931,
        "learning_rate": 3.58974358974359e-06,
        "loss": 0.016,
        "step": 2200
      },
      {
        "epoch": 33.02,
        "grad_norm": 0.05293996259570122,
        "learning_rate": 3.555555555555556e-06,
        "loss": 0.0924,
        "step": 2210
      },
      {
        "epoch": 33.02,
        "eval_accuracy": 0.782608695652174,
        "eval_loss": 0.9858593940734863,
        "eval_runtime": 10.3676,
        "eval_samples_per_second": 4.437,
        "eval_steps_per_second": 1.157,
        "step": 2210
      },
      {
        "epoch": 34.003076923076925,
        "grad_norm": 35.07648468017578,
        "learning_rate": 3.5213675213675218e-06,
        "loss": 0.0892,
        "step": 2220
      },
      {
        "epoch": 34.00615384615384,
        "grad_norm": 0.11222043633460999,
        "learning_rate": 3.487179487179487e-06,
        "loss": 0.1803,
        "step": 2230
      },
      {
        "epoch": 34.00923076923077,
        "grad_norm": 0.21430917084217072,
        "learning_rate": 3.452991452991453e-06,
        "loss": 0.1777,
        "step": 2240
      },
      {
        "epoch": 34.012307692307694,
        "grad_norm": 16.7849063873291,
        "learning_rate": 3.4188034188034193e-06,
        "loss": 0.1209,
        "step": 2250
      },
      {
        "epoch": 34.01538461538462,
        "grad_norm": 64.78038787841797,
        "learning_rate": 3.384615384615385e-06,
        "loss": 0.1789,
        "step": 2260
      },
      {
        "epoch": 34.01846153846154,
        "grad_norm": 0.4827536642551422,
        "learning_rate": 3.3504273504273506e-06,
        "loss": 0.1016,
        "step": 2270
      },
      {
        "epoch": 34.02,
        "eval_accuracy": 0.782608695652174,
        "eval_loss": 1.075831413269043,
        "eval_runtime": 10.3606,
        "eval_samples_per_second": 4.44,
        "eval_steps_per_second": 1.158,
        "step": 2275
      },
      {
        "epoch": 35.00153846153846,
        "grad_norm": 19.670827865600586,
        "learning_rate": 3.3162393162393165e-06,
        "loss": 0.0053,
        "step": 2280
      },
      {
        "epoch": 35.004615384615384,
        "grad_norm": 0.02641270123422146,
        "learning_rate": 3.2820512820512823e-06,
        "loss": 0.078,
        "step": 2290
      },
      {
        "epoch": 35.00769230769231,
        "grad_norm": 13.935211181640625,
        "learning_rate": 3.247863247863248e-06,
        "loss": 0.163,
        "step": 2300
      },
      {
        "epoch": 35.01076923076923,
        "grad_norm": 80.81470489501953,
        "learning_rate": 3.213675213675214e-06,
        "loss": 0.083,
        "step": 2310
      },
      {
        "epoch": 35.01384615384615,
        "grad_norm": 0.01466295588761568,
        "learning_rate": 3.1794871794871795e-06,
        "loss": 0.183,
        "step": 2320
      },
      {
        "epoch": 35.01692307692308,
        "grad_norm": 0.007615539710968733,
        "learning_rate": 3.1452991452991453e-06,
        "loss": 0.5267,
        "step": 2330
      },
      {
        "epoch": 35.02,
        "grad_norm": 0.019310418516397476,
        "learning_rate": 3.1111111111111116e-06,
        "loss": 0.1894,
        "step": 2340
      },
      {
        "epoch": 35.02,
        "eval_accuracy": 0.7608695652173914,
        "eval_loss": 1.208774209022522,
        "eval_runtime": 10.3816,
        "eval_samples_per_second": 4.431,
        "eval_steps_per_second": 1.156,
        "step": 2340
      },
      {
        "epoch": 36.003076923076925,
        "grad_norm": 0.17081318795681,
        "learning_rate": 3.0769230769230774e-06,
        "loss": 0.1565,
        "step": 2350
      },
      {
        "epoch": 36.00615384615384,
        "grad_norm": 3.074644088745117,
        "learning_rate": 3.042735042735043e-06,
        "loss": 0.2069,
        "step": 2360
      },
      {
        "epoch": 36.00923076923077,
        "grad_norm": 0.039210639894008636,
        "learning_rate": 3.0085470085470087e-06,
        "loss": 0.0321,
        "step": 2370
      },
      {
        "epoch": 36.012307692307694,
        "grad_norm": 195.3864288330078,
        "learning_rate": 2.9743589743589746e-06,
        "loss": 0.2972,
        "step": 2380
      },
      {
        "epoch": 36.01538461538462,
        "grad_norm": 0.26288241147994995,
        "learning_rate": 2.9401709401709404e-06,
        "loss": 0.0027,
        "step": 2390
      },
      {
        "epoch": 36.01846153846154,
        "grad_norm": 1.5095382928848267,
        "learning_rate": 2.9059829059829063e-06,
        "loss": 0.2657,
        "step": 2400
      },
      {
        "epoch": 36.02,
        "eval_accuracy": 0.7391304347826086,
        "eval_loss": 1.540911078453064,
        "eval_runtime": 10.5028,
        "eval_samples_per_second": 4.38,
        "eval_steps_per_second": 1.143,
        "step": 2405
      },
      {
        "epoch": 37.00153846153846,
        "grad_norm": 0.006212185136973858,
        "learning_rate": 2.8717948717948717e-06,
        "loss": 0.0008,
        "step": 2410
      },
      {
        "epoch": 37.004615384615384,
        "grad_norm": 1.8746821880340576,
        "learning_rate": 2.8376068376068376e-06,
        "loss": 0.2584,
        "step": 2420
      },
      {
        "epoch": 37.00769230769231,
        "grad_norm": 0.13942450284957886,
        "learning_rate": 2.803418803418804e-06,
        "loss": 0.0947,
        "step": 2430
      },
      {
        "epoch": 37.01076923076923,
        "grad_norm": 0.018762849271297455,
        "learning_rate": 2.7692307692307697e-06,
        "loss": 0.3546,
        "step": 2440
      },
      {
        "epoch": 37.01384615384615,
        "grad_norm": 0.018523642793297768,
        "learning_rate": 2.7350427350427355e-06,
        "loss": 0.1265,
        "step": 2450
      },
      {
        "epoch": 37.01692307692308,
        "grad_norm": 0.03315011411905289,
        "learning_rate": 2.700854700854701e-06,
        "loss": 0.1602,
        "step": 2460
      },
      {
        "epoch": 37.02,
        "grad_norm": 0.05689428746700287,
        "learning_rate": 2.666666666666667e-06,
        "loss": 0.1235,
        "step": 2470
      },
      {
        "epoch": 37.02,
        "eval_accuracy": 0.7608695652173914,
        "eval_loss": 1.2735745906829834,
        "eval_runtime": 10.41,
        "eval_samples_per_second": 4.419,
        "eval_steps_per_second": 1.153,
        "step": 2470
      },
      {
        "epoch": 38.003076923076925,
        "grad_norm": 0.03720976412296295,
        "learning_rate": 2.6324786324786327e-06,
        "loss": 0.0013,
        "step": 2480
      },
      {
        "epoch": 38.00615384615384,
        "grad_norm": 0.026006508618593216,
        "learning_rate": 2.5982905982905985e-06,
        "loss": 0.0833,
        "step": 2490
      },
      {
        "epoch": 38.00923076923077,
        "grad_norm": 0.11773164570331573,
        "learning_rate": 2.564102564102564e-06,
        "loss": 0.0828,
        "step": 2500
      },
      {
        "epoch": 38.012307692307694,
        "grad_norm": 0.06585920602083206,
        "learning_rate": 2.52991452991453e-06,
        "loss": 0.2328,
        "step": 2510
      },
      {
        "epoch": 38.01538461538462,
        "grad_norm": 64.87152862548828,
        "learning_rate": 2.495726495726496e-06,
        "loss": 0.019,
        "step": 2520
      },
      {
        "epoch": 38.01846153846154,
        "grad_norm": 30.827510833740234,
        "learning_rate": 2.461538461538462e-06,
        "loss": 0.1539,
        "step": 2530
      },
      {
        "epoch": 38.02,
        "eval_accuracy": 0.7608695652173914,
        "eval_loss": 1.2607876062393188,
        "eval_runtime": 10.3711,
        "eval_samples_per_second": 4.435,
        "eval_steps_per_second": 1.157,
        "step": 2535
      },
      {
        "epoch": 39.00153846153846,
        "grad_norm": 0.02473655343055725,
        "learning_rate": 2.4273504273504274e-06,
        "loss": 0.001,
        "step": 2540
      },
      {
        "epoch": 39.004615384615384,
        "grad_norm": 0.024143971502780914,
        "learning_rate": 2.3931623931623937e-06,
        "loss": 0.4801,
        "step": 2550
      },
      {
        "epoch": 39.00769230769231,
        "grad_norm": 0.03832996264100075,
        "learning_rate": 2.358974358974359e-06,
        "loss": 0.332,
        "step": 2560
      },
      {
        "epoch": 39.01076923076923,
        "grad_norm": 0.019543638452887535,
        "learning_rate": 2.324786324786325e-06,
        "loss": 0.0006,
        "step": 2570
      },
      {
        "epoch": 39.01384615384615,
        "grad_norm": 0.01422965619713068,
        "learning_rate": 2.290598290598291e-06,
        "loss": 0.3965,
        "step": 2580
      },
      {
        "epoch": 39.01692307692308,
        "grad_norm": 0.04189188778400421,
        "learning_rate": 2.2564102564102566e-06,
        "loss": 0.0699,
        "step": 2590
      },
      {
        "epoch": 39.02,
        "grad_norm": 1.5431499481201172,
        "learning_rate": 2.222222222222222e-06,
        "loss": 0.03,
        "step": 2600
      },
      {
        "epoch": 39.02,
        "eval_accuracy": 0.7608695652173914,
        "eval_loss": 1.2057995796203613,
        "eval_runtime": 10.1887,
        "eval_samples_per_second": 4.515,
        "eval_steps_per_second": 1.178,
        "step": 2600
      },
      {
        "epoch": 40.003076923076925,
        "grad_norm": 0.06963592022657394,
        "learning_rate": 2.1880341880341884e-06,
        "loss": 0.1184,
        "step": 2610
      },
      {
        "epoch": 40.00615384615384,
        "grad_norm": 0.03634906932711601,
        "learning_rate": 2.153846153846154e-06,
        "loss": 0.0023,
        "step": 2620
      },
      {
        "epoch": 40.00923076923077,
        "grad_norm": 0.026238996535539627,
        "learning_rate": 2.1196581196581196e-06,
        "loss": 0.2015,
        "step": 2630
      },
      {
        "epoch": 40.012307692307694,
        "grad_norm": 0.017962316051125526,
        "learning_rate": 2.085470085470086e-06,
        "loss": 0.2145,
        "step": 2640
      },
      {
        "epoch": 40.01538461538462,
        "grad_norm": 0.0143351498991251,
        "learning_rate": 2.0512820512820513e-06,
        "loss": 0.1958,
        "step": 2650
      },
      {
        "epoch": 40.01846153846154,
        "grad_norm": 0.08577225357294083,
        "learning_rate": 2.017094017094017e-06,
        "loss": 0.1447,
        "step": 2660
      },
      {
        "epoch": 40.02,
        "eval_accuracy": 0.7608695652173914,
        "eval_loss": 1.10723876953125,
        "eval_runtime": 10.6531,
        "eval_samples_per_second": 4.318,
        "eval_steps_per_second": 1.126,
        "step": 2665
      },
      {
        "epoch": 41.00153846153846,
        "grad_norm": 0.006715043913573027,
        "learning_rate": 1.982905982905983e-06,
        "loss": 0.1283,
        "step": 2670
      },
      {
        "epoch": 41.004615384615384,
        "grad_norm": 0.07786259055137634,
        "learning_rate": 1.948717948717949e-06,
        "loss": 0.0072,
        "step": 2680
      },
      {
        "epoch": 41.00769230769231,
        "grad_norm": 1.9331977367401123,
        "learning_rate": 1.9145299145299148e-06,
        "loss": 0.1126,
        "step": 2690
      },
      {
        "epoch": 41.01076923076923,
        "grad_norm": 176.3519744873047,
        "learning_rate": 1.8803418803418804e-06,
        "loss": 0.1764,
        "step": 2700
      },
      {
        "epoch": 41.01384615384615,
        "grad_norm": 1.9080088138580322,
        "learning_rate": 1.8461538461538465e-06,
        "loss": 0.099,
        "step": 2710
      },
      {
        "epoch": 41.01692307692308,
        "grad_norm": 0.02918870560824871,
        "learning_rate": 1.811965811965812e-06,
        "loss": 0.0805,
        "step": 2720
      },
      {
        "epoch": 41.02,
        "grad_norm": 227.27313232421875,
        "learning_rate": 1.777777777777778e-06,
        "loss": 0.0888,
        "step": 2730
      },
      {
        "epoch": 41.02,
        "eval_accuracy": 0.782608695652174,
        "eval_loss": 1.1454174518585205,
        "eval_runtime": 10.4748,
        "eval_samples_per_second": 4.392,
        "eval_steps_per_second": 1.146,
        "step": 2730
      },
      {
        "epoch": 42.003076923076925,
        "grad_norm": 0.015502427704632282,
        "learning_rate": 1.7435897435897436e-06,
        "loss": 0.0901,
        "step": 2740
      },
      {
        "epoch": 42.00615384615384,
        "grad_norm": 0.02255420945584774,
        "learning_rate": 1.7094017094017097e-06,
        "loss": 0.2097,
        "step": 2750
      },
      {
        "epoch": 42.00923076923077,
        "grad_norm": 186.2987060546875,
        "learning_rate": 1.6752136752136753e-06,
        "loss": 0.0813,
        "step": 2760
      },
      {
        "epoch": 42.012307692307694,
        "grad_norm": 1.8150595426559448,
        "learning_rate": 1.6410256410256412e-06,
        "loss": 0.1309,
        "step": 2770
      },
      {
        "epoch": 42.01538461538462,
        "grad_norm": 0.03459843620657921,
        "learning_rate": 1.606837606837607e-06,
        "loss": 0.1189,
        "step": 2780
      },
      {
        "epoch": 42.01846153846154,
        "grad_norm": 0.0022768706548959017,
        "learning_rate": 1.5726495726495727e-06,
        "loss": 0.0016,
        "step": 2790
      },
      {
        "epoch": 42.02,
        "eval_accuracy": 0.782608695652174,
        "eval_loss": 1.1194366216659546,
        "eval_runtime": 10.782,
        "eval_samples_per_second": 4.266,
        "eval_steps_per_second": 1.113,
        "step": 2795
      },
      {
        "epoch": 43.00153846153846,
        "grad_norm": 0.005485150497406721,
        "learning_rate": 1.5384615384615387e-06,
        "loss": 0.0008,
        "step": 2800
      },
      {
        "epoch": 43.004615384615384,
        "grad_norm": 0.7399391531944275,
        "learning_rate": 1.5042735042735044e-06,
        "loss": 0.1494,
        "step": 2810
      },
      {
        "epoch": 43.00769230769231,
        "grad_norm": 0.02695586159825325,
        "learning_rate": 1.4700854700854702e-06,
        "loss": 0.0015,
        "step": 2820
      },
      {
        "epoch": 43.01076923076923,
        "grad_norm": 0.013859664089977741,
        "learning_rate": 1.4358974358974359e-06,
        "loss": 0.0703,
        "step": 2830
      },
      {
        "epoch": 43.01384615384615,
        "grad_norm": 0.02371545322239399,
        "learning_rate": 1.401709401709402e-06,
        "loss": 0.1537,
        "step": 2840
      },
      {
        "epoch": 43.01692307692308,
        "grad_norm": 63.8938102722168,
        "learning_rate": 1.3675213675213678e-06,
        "loss": 0.1798,
        "step": 2850
      },
      {
        "epoch": 43.02,
        "grad_norm": 0.08463682234287262,
        "learning_rate": 1.3333333333333334e-06,
        "loss": 0.1489,
        "step": 2860
      },
      {
        "epoch": 43.02,
        "eval_accuracy": 0.7608695652173914,
        "eval_loss": 1.2169851064682007,
        "eval_runtime": 10.3239,
        "eval_samples_per_second": 4.456,
        "eval_steps_per_second": 1.162,
        "step": 2860
      },
      {
        "epoch": 44.003076923076925,
        "grad_norm": 105.62195587158203,
        "learning_rate": 1.2991452991452993e-06,
        "loss": 0.036,
        "step": 2870
      },
      {
        "epoch": 44.00615384615384,
        "grad_norm": 191.20787048339844,
        "learning_rate": 1.264957264957265e-06,
        "loss": 0.0918,
        "step": 2880
      },
      {
        "epoch": 44.00923076923077,
        "grad_norm": 0.17883938550949097,
        "learning_rate": 1.230769230769231e-06,
        "loss": 0.1685,
        "step": 2890
      },
      {
        "epoch": 44.012307692307694,
        "grad_norm": 0.017934106290340424,
        "learning_rate": 1.1965811965811968e-06,
        "loss": 0.0005,
        "step": 2900
      },
      {
        "epoch": 44.01538461538462,
        "grad_norm": 0.1711297333240509,
        "learning_rate": 1.1623931623931625e-06,
        "loss": 0.3324,
        "step": 2910
      },
      {
        "epoch": 44.01846153846154,
        "grad_norm": 0.05496324226260185,
        "learning_rate": 1.1282051282051283e-06,
        "loss": 0.0004,
        "step": 2920
      },
      {
        "epoch": 44.02,
        "eval_accuracy": 0.7608695652173914,
        "eval_loss": 1.1893503665924072,
        "eval_runtime": 10.4088,
        "eval_samples_per_second": 4.419,
        "eval_steps_per_second": 1.153,
        "step": 2925
      },
      {
        "epoch": 45.00153846153846,
        "grad_norm": 0.017374416813254356,
        "learning_rate": 1.0940170940170942e-06,
        "loss": 0.3326,
        "step": 2930
      },
      {
        "epoch": 45.004615384615384,
        "grad_norm": 0.01586207188665867,
        "learning_rate": 1.0598290598290598e-06,
        "loss": 0.1512,
        "step": 2940
      },
      {
        "epoch": 45.00769230769231,
        "grad_norm": 0.7407118082046509,
        "learning_rate": 1.0256410256410257e-06,
        "loss": 0.1375,
        "step": 2950
      },
      {
        "epoch": 45.01076923076923,
        "grad_norm": 0.00448651472106576,
        "learning_rate": 9.914529914529915e-07,
        "loss": 0.0243,
        "step": 2960
      },
      {
        "epoch": 45.01384615384615,
        "grad_norm": 0.17225775122642517,
        "learning_rate": 9.572649572649574e-07,
        "loss": 0.1769,
        "step": 2970
      },
      {
        "epoch": 45.01692307692308,
        "grad_norm": 0.10238823294639587,
        "learning_rate": 9.230769230769232e-07,
        "loss": 0.0217,
        "step": 2980
      },
      {
        "epoch": 45.02,
        "grad_norm": 0.019600247964262962,
        "learning_rate": 8.88888888888889e-07,
        "loss": 0.0004,
        "step": 2990
      },
      {
        "epoch": 45.02,
        "eval_accuracy": 0.7391304347826086,
        "eval_loss": 1.332923173904419,
        "eval_runtime": 10.5547,
        "eval_samples_per_second": 4.358,
        "eval_steps_per_second": 1.137,
        "step": 2990
      },
      {
        "epoch": 46.003076923076925,
        "grad_norm": 0.018734971061348915,
        "learning_rate": 8.547008547008548e-07,
        "loss": 0.0202,
        "step": 3000
      },
      {
        "epoch": 46.00615384615384,
        "grad_norm": 0.0044385273940861225,
        "learning_rate": 8.205128205128206e-07,
        "loss": 0.0009,
        "step": 3010
      },
      {
        "epoch": 46.00923076923077,
        "grad_norm": 0.06210096552968025,
        "learning_rate": 7.863247863247863e-07,
        "loss": 0.2226,
        "step": 3020
      },
      {
        "epoch": 46.012307692307694,
        "grad_norm": 0.004177386872470379,
        "learning_rate": 7.521367521367522e-07,
        "loss": 0.0477,
        "step": 3030
      },
      {
        "epoch": 46.01538461538462,
        "grad_norm": 0.013359845615923405,
        "learning_rate": 7.179487179487179e-07,
        "loss": 0.0004,
        "step": 3040
      },
      {
        "epoch": 46.01846153846154,
        "grad_norm": 0.011247428134083748,
        "learning_rate": 6.837606837606839e-07,
        "loss": 0.0014,
        "step": 3050
      },
      {
        "epoch": 46.02,
        "eval_accuracy": 0.7608695652173914,
        "eval_loss": 1.1886558532714844,
        "eval_runtime": 10.4416,
        "eval_samples_per_second": 4.405,
        "eval_steps_per_second": 1.149,
        "step": 3055
      },
      {
        "epoch": 47.00153846153846,
        "grad_norm": 0.0468427948653698,
        "learning_rate": 6.495726495726496e-07,
        "loss": 0.1823,
        "step": 3060
      },
      {
        "epoch": 47.004615384615384,
        "grad_norm": 0.029943831264972687,
        "learning_rate": 6.153846153846155e-07,
        "loss": 0.0144,
        "step": 3070
      },
      {
        "epoch": 47.00769230769231,
        "grad_norm": 0.012487895786762238,
        "learning_rate": 5.811965811965812e-07,
        "loss": 0.1565,
        "step": 3080
      },
      {
        "epoch": 47.01076923076923,
        "grad_norm": 0.0036076002288609743,
        "learning_rate": 5.470085470085471e-07,
        "loss": 0.0188,
        "step": 3090
      },
      {
        "epoch": 47.01384615384615,
        "grad_norm": 0.043261803686618805,
        "learning_rate": 5.128205128205128e-07,
        "loss": 0.1021,
        "step": 3100
      },
      {
        "epoch": 47.01692307692308,
        "grad_norm": 0.31245148181915283,
        "learning_rate": 4.786324786324787e-07,
        "loss": 0.0004,
        "step": 3110
      },
      {
        "epoch": 47.02,
        "grad_norm": 0.009071093052625656,
        "learning_rate": 4.444444444444445e-07,
        "loss": 0.1675,
        "step": 3120
      },
      {
        "epoch": 47.02,
        "eval_accuracy": 0.7391304347826086,
        "eval_loss": 1.265150785446167,
        "eval_runtime": 10.6277,
        "eval_samples_per_second": 4.328,
        "eval_steps_per_second": 1.129,
        "step": 3120
      },
      {
        "epoch": 48.003076923076925,
        "grad_norm": 0.9440973997116089,
        "learning_rate": 4.102564102564103e-07,
        "loss": 0.128,
        "step": 3130
      },
      {
        "epoch": 48.00615384615384,
        "grad_norm": 0.008040065877139568,
        "learning_rate": 3.760683760683761e-07,
        "loss": 0.0086,
        "step": 3140
      },
      {
        "epoch": 48.00923076923077,
        "grad_norm": 0.3104320764541626,
        "learning_rate": 3.4188034188034194e-07,
        "loss": 0.0071,
        "step": 3150
      },
      {
        "epoch": 48.012307692307694,
        "grad_norm": 0.012552163563668728,
        "learning_rate": 3.0769230769230774e-07,
        "loss": 0.0006,
        "step": 3160
      },
      {
        "epoch": 48.01538461538462,
        "grad_norm": 0.014428188093006611,
        "learning_rate": 2.7350427350427354e-07,
        "loss": 0.0616,
        "step": 3170
      },
      {
        "epoch": 48.01846153846154,
        "grad_norm": 0.03152965009212494,
        "learning_rate": 2.3931623931623934e-07,
        "loss": 0.012,
        "step": 3180
      },
      {
        "epoch": 48.02,
        "eval_accuracy": 0.7391304347826086,
        "eval_loss": 1.3228085041046143,
        "eval_runtime": 10.4237,
        "eval_samples_per_second": 4.413,
        "eval_steps_per_second": 1.151,
        "step": 3185
      },
      {
        "epoch": 49.00153846153846,
        "grad_norm": 0.5161852240562439,
        "learning_rate": 2.0512820512820514e-07,
        "loss": 0.0014,
        "step": 3190
      },
      {
        "epoch": 49.004615384615384,
        "grad_norm": 0.02032632566988468,
        "learning_rate": 1.7094017094017097e-07,
        "loss": 0.0003,
        "step": 3200
      },
      {
        "epoch": 49.00769230769231,
        "grad_norm": 0.014277896843850613,
        "learning_rate": 1.3675213675213677e-07,
        "loss": 0.0047,
        "step": 3210
      },
      {
        "epoch": 49.01076923076923,
        "grad_norm": 0.3848668038845062,
        "learning_rate": 1.0256410256410257e-07,
        "loss": 0.001,
        "step": 3220
      },
      {
        "epoch": 49.01384615384615,
        "grad_norm": 0.38336512446403503,
        "learning_rate": 6.837606837606839e-08,
        "loss": 0.0148,
        "step": 3230
      },
      {
        "epoch": 49.01692307692308,
        "grad_norm": 0.012916259467601776,
        "learning_rate": 3.418803418803419e-08,
        "loss": 0.288,
        "step": 3240
      },
      {
        "epoch": 49.02,
        "grad_norm": 0.013959445059299469,
        "learning_rate": 0.0,
        "loss": 0.0475,
        "step": 3250
      },
      {
        "epoch": 49.02,
        "eval_accuracy": 0.7391304347826086,
        "eval_loss": 1.3506942987442017,
        "eval_runtime": 11.7124,
        "eval_samples_per_second": 3.927,
        "eval_steps_per_second": 1.025,
        "step": 3250
      },
      {
        "epoch": 49.02,
        "step": 3250,
        "total_flos": 5.708334209885798e+19,
        "train_loss": 0.3139108933124405,
        "train_runtime": 6672.0929,
        "train_samples_per_second": 1.948,
        "train_steps_per_second": 0.487
      },
      {
        "epoch": 49.02,
        "eval_accuracy": 0.8043478260869565,
        "eval_loss": 0.48940712213516235,
        "eval_runtime": 10.4364,
        "eval_samples_per_second": 4.408,
        "eval_steps_per_second": 1.15,
        "step": 3250
      },
      {
        "epoch": 49.02,
        "eval_accuracy": 0.8043478260869565,
        "eval_loss": 0.4894070327281952,
        "eval_runtime": 10.4359,
        "eval_samples_per_second": 4.408,
        "eval_steps_per_second": 1.15,
        "step": 3250
      }
    ],
    "logging_steps": 10,
    "max_steps": 3250,
    "num_input_tokens_seen": 0,
    "num_train_epochs": 9223372036854775807,
    "save_steps": 500,
    "stateful_callbacks": {
      "TrainerControl": {
        "args": {
          "should_epoch_stop": false,
          "should_evaluate": false,
          "should_log": false,
          "should_save": true,
          "should_training_stop": true
        },
        "attributes": {}
      }
    },
    "total_flos": 5.708334209885798e+19,
    "train_batch_size": 4,
    "trial_name": null,
    "trial_params": null
  }