{
  "best_metric": 0.39012426137924194,
  "best_model_checkpoint": "limb_classification_person_crop_seq/t4_2heads_1layers_1e-4lr/checkpoint-1332",
  "epoch": 20.0,
  "eval_steps": 500,
  "global_step": 2960,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.16891891891891891,
      "grad_norm": 359864.8125,
      "learning_rate": 1e-05,
      "loss": 1.3881,
      "step": 25
    },
    {
      "epoch": 0.33783783783783783,
      "grad_norm": 264328.90625,
      "learning_rate": 2e-05,
      "loss": 1.3598,
      "step": 50
    },
    {
      "epoch": 0.5067567567567568,
      "grad_norm": 156298.703125,
      "learning_rate": 3e-05,
      "loss": 1.298,
      "step": 75
    },
    {
      "epoch": 0.6756756756756757,
      "grad_norm": 567283.0,
      "learning_rate": 4e-05,
      "loss": 1.1865,
      "step": 100
    },
    {
      "epoch": 0.8445945945945946,
      "grad_norm": 126027.78125,
      "learning_rate": 5e-05,
      "loss": 1.0527,
      "step": 125
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.8347305389221557,
      "eval_loss": 0.6327010989189148,
      "eval_runtime": 59.1382,
      "eval_samples_per_second": 14.119,
      "eval_steps_per_second": 0.457,
      "step": 148
    },
    {
      "epoch": 1.0135135135135136,
      "grad_norm": 326887.40625,
      "learning_rate": 6e-05,
      "loss": 0.7731,
      "step": 150
    },
    {
      "epoch": 1.1824324324324325,
      "grad_norm": 105805.6015625,
      "learning_rate": 7e-05,
      "loss": 0.5788,
      "step": 175
    },
    {
      "epoch": 1.3513513513513513,
      "grad_norm": 233410.15625,
      "learning_rate": 8e-05,
      "loss": 0.5242,
      "step": 200
    },
    {
      "epoch": 1.5202702702702702,
      "grad_norm": 163717.046875,
      "learning_rate": 9e-05,
      "loss": 0.4505,
      "step": 225
    },
    {
      "epoch": 1.689189189189189,
      "grad_norm": 85214.3671875,
      "learning_rate": 0.0001,
      "loss": 0.4367,
      "step": 250
    },
    {
      "epoch": 1.8581081081081081,
      "grad_norm": 111542.828125,
      "learning_rate": 9.907749077490776e-05,
      "loss": 0.4279,
      "step": 275
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.9101796407185628,
      "eval_loss": 0.41295987367630005,
      "eval_runtime": 58.457,
      "eval_samples_per_second": 14.284,
      "eval_steps_per_second": 0.462,
      "step": 296
    },
    {
      "epoch": 2.027027027027027,
      "grad_norm": 462517.65625,
      "learning_rate": 9.81549815498155e-05,
      "loss": 0.4751,
      "step": 300
    },
    {
      "epoch": 2.195945945945946,
      "grad_norm": 166061.90625,
      "learning_rate": 9.723247232472326e-05,
      "loss": 0.4335,
      "step": 325
    },
    {
      "epoch": 2.364864864864865,
      "grad_norm": 107032.1953125,
      "learning_rate": 9.6309963099631e-05,
      "loss": 0.4508,
      "step": 350
    },
    {
      "epoch": 2.5337837837837838,
      "grad_norm": 165233.625,
      "learning_rate": 9.538745387453874e-05,
      "loss": 0.4059,
      "step": 375
    },
    {
      "epoch": 2.7027027027027026,
      "grad_norm": 130664.3203125,
      "learning_rate": 9.44649446494465e-05,
      "loss": 0.4437,
      "step": 400
    },
    {
      "epoch": 2.8716216216216215,
      "grad_norm": 232347.84375,
      "learning_rate": 9.354243542435425e-05,
      "loss": 0.4649,
      "step": 425
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.9173652694610779,
      "eval_loss": 0.41219642758369446,
      "eval_runtime": 58.4053,
      "eval_samples_per_second": 14.297,
      "eval_steps_per_second": 0.462,
      "step": 444
    },
    {
      "epoch": 3.0405405405405403,
      "grad_norm": 114807.3671875,
      "learning_rate": 9.2619926199262e-05,
      "loss": 0.4141,
      "step": 450
    },
    {
      "epoch": 3.2094594594594597,
      "grad_norm": 221290.296875,
      "learning_rate": 9.169741697416975e-05,
      "loss": 0.4303,
      "step": 475
    },
    {
      "epoch": 3.3783783783783785,
      "grad_norm": 155868.46875,
      "learning_rate": 9.077490774907749e-05,
      "loss": 0.4119,
      "step": 500
    },
    {
      "epoch": 3.5472972972972974,
      "grad_norm": 81905.5546875,
      "learning_rate": 8.985239852398525e-05,
      "loss": 0.4081,
      "step": 525
    },
    {
      "epoch": 3.7162162162162162,
      "grad_norm": 100550.203125,
      "learning_rate": 8.892988929889299e-05,
      "loss": 0.4393,
      "step": 550
    },
    {
      "epoch": 3.885135135135135,
      "grad_norm": 108909.9765625,
      "learning_rate": 8.800738007380073e-05,
      "loss": 0.4404,
      "step": 575
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.9077844311377246,
      "eval_loss": 0.40994617342948914,
      "eval_runtime": 58.9872,
      "eval_samples_per_second": 14.156,
      "eval_steps_per_second": 0.458,
      "step": 592
    },
    {
      "epoch": 4.054054054054054,
      "grad_norm": 134412.171875,
      "learning_rate": 8.708487084870849e-05,
      "loss": 0.4583,
      "step": 600
    },
    {
      "epoch": 4.222972972972973,
      "grad_norm": 61759.4609375,
      "learning_rate": 8.616236162361624e-05,
      "loss": 0.397,
      "step": 625
    },
    {
      "epoch": 4.391891891891892,
      "grad_norm": 110399.734375,
      "learning_rate": 8.523985239852399e-05,
      "loss": 0.408,
      "step": 650
    },
    {
      "epoch": 4.5608108108108105,
      "grad_norm": 87549.25,
      "learning_rate": 8.431734317343174e-05,
      "loss": 0.435,
      "step": 675
    },
    {
      "epoch": 4.72972972972973,
      "grad_norm": 154651.8125,
      "learning_rate": 8.339483394833948e-05,
      "loss": 0.4258,
      "step": 700
    },
    {
      "epoch": 4.898648648648649,
      "grad_norm": 96262.6015625,
      "learning_rate": 8.247232472324724e-05,
      "loss": 0.4154,
      "step": 725
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.9125748502994012,
      "eval_loss": 0.404330313205719,
      "eval_runtime": 57.7013,
      "eval_samples_per_second": 14.471,
      "eval_steps_per_second": 0.468,
      "step": 740
    },
    {
      "epoch": 5.0675675675675675,
      "grad_norm": 203876.703125,
      "learning_rate": 8.154981549815498e-05,
      "loss": 0.4084,
      "step": 750
    },
    {
      "epoch": 5.236486486486487,
      "grad_norm": 92342.8046875,
      "learning_rate": 8.062730627306274e-05,
      "loss": 0.3944,
      "step": 775
    },
    {
      "epoch": 5.405405405405405,
      "grad_norm": 63151.5625,
      "learning_rate": 7.970479704797048e-05,
      "loss": 0.4049,
      "step": 800
    },
    {
      "epoch": 5.574324324324325,
      "grad_norm": 166211.109375,
      "learning_rate": 7.878228782287823e-05,
      "loss": 0.4139,
      "step": 825
    },
    {
      "epoch": 5.743243243243243,
      "grad_norm": 83852.1171875,
      "learning_rate": 7.785977859778598e-05,
      "loss": 0.4137,
      "step": 850
    },
    {
      "epoch": 5.912162162162162,
      "grad_norm": 62941.37890625,
      "learning_rate": 7.693726937269373e-05,
      "loss": 0.4252,
      "step": 875
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.911377245508982,
      "eval_loss": 0.4019243121147156,
      "eval_runtime": 58.0921,
      "eval_samples_per_second": 14.374,
      "eval_steps_per_second": 0.465,
      "step": 888
    },
    {
      "epoch": 6.081081081081081,
      "grad_norm": 104411.078125,
      "learning_rate": 7.601476014760149e-05,
      "loss": 0.4388,
      "step": 900
    },
    {
      "epoch": 6.25,
      "grad_norm": 132399.46875,
      "learning_rate": 7.509225092250923e-05,
      "loss": 0.4222,
      "step": 925
    },
    {
      "epoch": 6.418918918918919,
      "grad_norm": 136244.46875,
      "learning_rate": 7.416974169741697e-05,
      "loss": 0.4053,
      "step": 950
    },
    {
      "epoch": 6.587837837837838,
      "grad_norm": 178182.4375,
      "learning_rate": 7.324723247232473e-05,
      "loss": 0.405,
      "step": 975
    },
    {
      "epoch": 6.756756756756757,
      "grad_norm": 113690.484375,
      "learning_rate": 7.232472324723247e-05,
      "loss": 0.389,
      "step": 1000
    },
    {
      "epoch": 6.925675675675675,
      "grad_norm": 60490.4453125,
      "learning_rate": 7.140221402214023e-05,
      "loss": 0.405,
      "step": 1025
    },
    {
      "epoch": 7.0,
      "eval_accuracy": 0.9149700598802395,
      "eval_loss": 0.4009867310523987,
      "eval_runtime": 59.8777,
      "eval_samples_per_second": 13.945,
      "eval_steps_per_second": 0.451,
      "step": 1036
    },
    {
      "epoch": 7.094594594594595,
      "grad_norm": 117951.9453125,
      "learning_rate": 7.047970479704797e-05,
      "loss": 0.4331,
      "step": 1050
    },
    {
      "epoch": 7.263513513513513,
      "grad_norm": 65605.359375,
      "learning_rate": 6.955719557195572e-05,
      "loss": 0.3982,
      "step": 1075
    },
    {
      "epoch": 7.4324324324324325,
      "grad_norm": 131225.96875,
      "learning_rate": 6.863468634686348e-05,
      "loss": 0.4179,
      "step": 1100
    },
    {
      "epoch": 7.601351351351351,
      "grad_norm": 97153.0234375,
      "learning_rate": 6.771217712177122e-05,
      "loss": 0.3782,
      "step": 1125
    },
    {
      "epoch": 7.77027027027027,
      "grad_norm": 106960.203125,
      "learning_rate": 6.678966789667896e-05,
      "loss": 0.4064,
      "step": 1150
    },
    {
      "epoch": 7.9391891891891895,
      "grad_norm": 166451.640625,
      "learning_rate": 6.586715867158672e-05,
      "loss": 0.3991,
      "step": 1175
    },
    {
      "epoch": 8.0,
      "eval_accuracy": 0.9173652694610779,
      "eval_loss": 0.4009450078010559,
      "eval_runtime": 59.2339,
      "eval_samples_per_second": 14.097,
      "eval_steps_per_second": 0.456,
      "step": 1184
    },
    {
      "epoch": 8.108108108108109,
      "grad_norm": 73638.703125,
      "learning_rate": 6.494464944649446e-05,
      "loss": 0.3735,
      "step": 1200
    },
    {
      "epoch": 8.277027027027026,
      "grad_norm": 46720.8671875,
      "learning_rate": 6.402214022140222e-05,
      "loss": 0.3804,
      "step": 1225
    },
    {
      "epoch": 8.445945945945946,
      "grad_norm": 90134.90625,
      "learning_rate": 6.309963099630997e-05,
      "loss": 0.3987,
      "step": 1250
    },
    {
      "epoch": 8.614864864864865,
      "grad_norm": 95625.1796875,
      "learning_rate": 6.217712177121771e-05,
      "loss": 0.3762,
      "step": 1275
    },
    {
      "epoch": 8.783783783783784,
      "grad_norm": 79090.5546875,
      "learning_rate": 6.125461254612547e-05,
      "loss": 0.4101,
      "step": 1300
    },
    {
      "epoch": 8.952702702702704,
      "grad_norm": 206360.375,
      "learning_rate": 6.033210332103322e-05,
      "loss": 0.4304,
      "step": 1325
    },
    {
      "epoch": 9.0,
      "eval_accuracy": 0.918562874251497,
      "eval_loss": 0.39012426137924194,
      "eval_runtime": 57.4974,
      "eval_samples_per_second": 14.522,
      "eval_steps_per_second": 0.47,
      "step": 1332
    },
    {
      "epoch": 9.121621621621621,
      "grad_norm": 104168.234375,
      "learning_rate": 5.940959409594096e-05,
      "loss": 0.3851,
      "step": 1350
    },
    {
      "epoch": 9.29054054054054,
      "grad_norm": 166935.71875,
      "learning_rate": 5.848708487084871e-05,
      "loss": 0.3884,
      "step": 1375
    },
    {
      "epoch": 9.45945945945946,
      "grad_norm": 109088.5078125,
      "learning_rate": 5.756457564575646e-05,
      "loss": 0.3976,
      "step": 1400
    },
    {
      "epoch": 9.628378378378379,
      "grad_norm": 73942.0390625,
      "learning_rate": 5.664206642066421e-05,
      "loss": 0.3725,
      "step": 1425
    },
    {
      "epoch": 9.797297297297296,
      "grad_norm": 48857.83203125,
      "learning_rate": 5.5719557195571956e-05,
      "loss": 0.4009,
      "step": 1450
    },
    {
      "epoch": 9.966216216216216,
      "grad_norm": 78605.0,
      "learning_rate": 5.479704797047971e-05,
      "loss": 0.3902,
      "step": 1475
    },
    {
      "epoch": 10.0,
      "eval_accuracy": 0.9137724550898204,
      "eval_loss": 0.3972083628177643,
      "eval_runtime": 58.4963,
      "eval_samples_per_second": 14.274,
      "eval_steps_per_second": 0.462,
      "step": 1480
    },
    {
      "epoch": 10.135135135135135,
      "grad_norm": 39732.36328125,
      "learning_rate": 5.387453874538746e-05,
      "loss": 0.3833,
      "step": 1500
    },
    {
      "epoch": 10.304054054054054,
      "grad_norm": 121495.34375,
      "learning_rate": 5.295202952029521e-05,
      "loss": 0.4035,
      "step": 1525
    },
    {
      "epoch": 10.472972972972974,
      "grad_norm": 126188.4140625,
      "learning_rate": 5.202952029520295e-05,
      "loss": 0.3563,
      "step": 1550
    },
    {
      "epoch": 10.641891891891891,
      "grad_norm": 84368.2265625,
      "learning_rate": 5.11070110701107e-05,
      "loss": 0.3837,
      "step": 1575
    },
    {
      "epoch": 10.81081081081081,
      "grad_norm": 152500.28125,
      "learning_rate": 5.018450184501845e-05,
      "loss": 0.4131,
      "step": 1600
    },
    {
      "epoch": 10.97972972972973,
      "grad_norm": 116200.03125,
      "learning_rate": 4.92619926199262e-05,
      "loss": 0.376,
      "step": 1625
    },
    {
      "epoch": 11.0,
      "eval_accuracy": 0.9173652694610779,
      "eval_loss": 0.4009062349796295,
      "eval_runtime": 58.017,
      "eval_samples_per_second": 14.392,
      "eval_steps_per_second": 0.465,
      "step": 1628
    },
    {
      "epoch": 11.14864864864865,
      "grad_norm": 62141.8359375,
      "learning_rate": 4.833948339483395e-05,
      "loss": 0.3911,
      "step": 1650
    },
    {
      "epoch": 11.317567567567568,
      "grad_norm": 72781.9375,
      "learning_rate": 4.74169741697417e-05,
      "loss": 0.3968,
      "step": 1675
    },
    {
      "epoch": 11.486486486486486,
      "grad_norm": 83891.21875,
      "learning_rate": 4.6494464944649444e-05,
      "loss": 0.3599,
      "step": 1700
    },
    {
      "epoch": 11.655405405405405,
      "grad_norm": 107400.21875,
      "learning_rate": 4.55719557195572e-05,
      "loss": 0.3756,
      "step": 1725
    },
    {
      "epoch": 11.824324324324325,
      "grad_norm": 155212.625,
      "learning_rate": 4.464944649446495e-05,
      "loss": 0.3567,
      "step": 1750
    },
    {
      "epoch": 11.993243243243244,
      "grad_norm": 124735.5859375,
      "learning_rate": 4.37269372693727e-05,
      "loss": 0.4028,
      "step": 1775
    },
    {
      "epoch": 12.0,
      "eval_accuracy": 0.918562874251497,
      "eval_loss": 0.3954586982727051,
      "eval_runtime": 57.1928,
      "eval_samples_per_second": 14.6,
      "eval_steps_per_second": 0.472,
      "step": 1776
    },
    {
      "epoch": 12.162162162162161,
      "grad_norm": 188278.84375,
      "learning_rate": 4.280442804428044e-05,
      "loss": 0.3678,
      "step": 1800
    },
    {
      "epoch": 12.33108108108108,
      "grad_norm": 222535.890625,
      "learning_rate": 4.1881918819188195e-05,
      "loss": 0.3756,
      "step": 1825
    },
    {
      "epoch": 12.5,
      "grad_norm": 50271.54296875,
      "learning_rate": 4.0959409594095944e-05,
      "loss": 0.396,
      "step": 1850
    },
    {
      "epoch": 12.66891891891892,
      "grad_norm": 96267.390625,
      "learning_rate": 4.003690036900369e-05,
      "loss": 0.4069,
      "step": 1875
    },
    {
      "epoch": 12.837837837837839,
      "grad_norm": 53173.109375,
      "learning_rate": 3.911439114391144e-05,
      "loss": 0.3803,
      "step": 1900
    },
    {
      "epoch": 13.0,
      "eval_accuracy": 0.918562874251497,
      "eval_loss": 0.39904314279556274,
      "eval_runtime": 57.8057,
      "eval_samples_per_second": 14.445,
      "eval_steps_per_second": 0.467,
      "step": 1924
    },
    {
      "epoch": 13.006756756756756,
      "grad_norm": 98902.09375,
      "learning_rate": 3.819188191881919e-05,
      "loss": 0.3368,
      "step": 1925
    },
    {
      "epoch": 13.175675675675675,
      "grad_norm": 148564.765625,
      "learning_rate": 3.726937269372694e-05,
      "loss": 0.3937,
      "step": 1950
    },
    {
      "epoch": 13.344594594594595,
      "grad_norm": 156506.828125,
      "learning_rate": 3.634686346863469e-05,
      "loss": 0.3587,
      "step": 1975
    },
    {
      "epoch": 13.513513513513514,
      "grad_norm": 120063.5546875,
      "learning_rate": 3.542435424354244e-05,
      "loss": 0.3677,
      "step": 2000
    },
    {
      "epoch": 13.682432432432432,
      "grad_norm": 74943.984375,
      "learning_rate": 3.4501845018450186e-05,
      "loss": 0.3675,
      "step": 2025
    },
    {
      "epoch": 13.85135135135135,
      "grad_norm": 49523.3828125,
      "learning_rate": 3.3579335793357934e-05,
      "loss": 0.4008,
      "step": 2050
    },
    {
      "epoch": 14.0,
      "eval_accuracy": 0.918562874251497,
      "eval_loss": 0.40573185682296753,
      "eval_runtime": 57.2611,
      "eval_samples_per_second": 14.582,
      "eval_steps_per_second": 0.472,
      "step": 2072
    },
    {
      "epoch": 14.02027027027027,
      "grad_norm": 81853.390625,
      "learning_rate": 3.265682656826568e-05,
      "loss": 0.3808,
      "step": 2075
    },
    {
      "epoch": 14.18918918918919,
      "grad_norm": 78917.8203125,
      "learning_rate": 3.173431734317343e-05,
      "loss": 0.3586,
      "step": 2100
    },
    {
      "epoch": 14.358108108108109,
      "grad_norm": 199516.0,
      "learning_rate": 3.081180811808118e-05,
      "loss": 0.3856,
      "step": 2125
    },
    {
      "epoch": 14.527027027027026,
      "grad_norm": 66714.390625,
      "learning_rate": 2.9889298892988933e-05,
      "loss": 0.374,
      "step": 2150
    },
    {
      "epoch": 14.695945945945946,
      "grad_norm": 162585.71875,
      "learning_rate": 2.8966789667896682e-05,
      "loss": 0.3821,
      "step": 2175
    },
    {
      "epoch": 14.864864864864865,
      "grad_norm": 28608.001953125,
      "learning_rate": 2.8044280442804427e-05,
      "loss": 0.3646,
      "step": 2200
    },
    {
      "epoch": 15.0,
      "eval_accuracy": 0.9173652694610779,
      "eval_loss": 0.4000730514526367,
      "eval_runtime": 56.6006,
      "eval_samples_per_second": 14.752,
      "eval_steps_per_second": 0.477,
      "step": 2220
    },
    {
      "epoch": 15.033783783783784,
      "grad_norm": 180226.921875,
      "learning_rate": 2.7121771217712176e-05,
      "loss": 0.3746,
      "step": 2225
    },
    {
      "epoch": 15.202702702702704,
      "grad_norm": 158761.0625,
      "learning_rate": 2.619926199261993e-05,
      "loss": 0.3766,
      "step": 2250
    },
    {
      "epoch": 15.371621621621621,
      "grad_norm": 91877.6171875,
      "learning_rate": 2.5276752767527677e-05,
      "loss": 0.3587,
      "step": 2275
    },
    {
      "epoch": 15.54054054054054,
      "grad_norm": 108580.234375,
      "learning_rate": 2.4354243542435426e-05,
      "loss": 0.3631,
      "step": 2300
    },
    {
      "epoch": 15.70945945945946,
      "grad_norm": 139217.03125,
      "learning_rate": 2.3431734317343175e-05,
      "loss": 0.3592,
      "step": 2325
    },
    {
      "epoch": 15.878378378378379,
      "grad_norm": 132462.828125,
      "learning_rate": 2.2509225092250924e-05,
      "loss": 0.3707,
      "step": 2350
    },
    {
      "epoch": 16.0,
      "eval_accuracy": 0.9197604790419162,
      "eval_loss": 0.39095538854599,
      "eval_runtime": 58.7037,
      "eval_samples_per_second": 14.224,
      "eval_steps_per_second": 0.46,
      "step": 2368
    },
    {
      "epoch": 16.0472972972973,
      "grad_norm": 72858.5703125,
      "learning_rate": 2.1586715867158673e-05,
      "loss": 0.371,
      "step": 2375
    },
    {
      "epoch": 16.216216216216218,
      "grad_norm": 123128.625,
      "learning_rate": 2.066420664206642e-05,
      "loss": 0.3316,
      "step": 2400
    },
    {
      "epoch": 16.385135135135137,
      "grad_norm": 64682.34375,
      "learning_rate": 1.974169741697417e-05,
      "loss": 0.3571,
      "step": 2425
    },
    {
      "epoch": 16.554054054054053,
      "grad_norm": 132089.734375,
      "learning_rate": 1.881918819188192e-05,
      "loss": 0.3831,
      "step": 2450
    },
    {
      "epoch": 16.722972972972972,
      "grad_norm": 95733.890625,
      "learning_rate": 1.7896678966789668e-05,
      "loss": 0.3448,
      "step": 2475
    },
    {
      "epoch": 16.89189189189189,
      "grad_norm": 42372.92578125,
      "learning_rate": 1.6974169741697417e-05,
      "loss": 0.3488,
      "step": 2500
    },
    {
      "epoch": 17.0,
      "eval_accuracy": 0.9197604790419162,
      "eval_loss": 0.39478474855422974,
      "eval_runtime": 57.144,
      "eval_samples_per_second": 14.612,
      "eval_steps_per_second": 0.472,
      "step": 2516
    },
    {
      "epoch": 17.06081081081081,
      "grad_norm": 108961.515625,
      "learning_rate": 1.6051660516605166e-05,
      "loss": 0.3863,
      "step": 2525
    },
    {
      "epoch": 17.22972972972973,
      "grad_norm": 64582.25390625,
      "learning_rate": 1.5129151291512916e-05,
      "loss": 0.3613,
      "step": 2550
    },
    {
      "epoch": 17.39864864864865,
      "grad_norm": 76625.1015625,
      "learning_rate": 1.4206642066420663e-05,
      "loss": 0.359,
      "step": 2575
    },
    {
      "epoch": 17.56756756756757,
      "grad_norm": 66550.7734375,
      "learning_rate": 1.3284132841328414e-05,
      "loss": 0.3543,
      "step": 2600
    },
    {
      "epoch": 17.736486486486488,
      "grad_norm": 131640.515625,
      "learning_rate": 1.2361623616236164e-05,
      "loss": 0.3256,
      "step": 2625
    },
    {
      "epoch": 17.905405405405407,
      "grad_norm": 73564.6484375,
      "learning_rate": 1.1439114391143913e-05,
      "loss": 0.3842,
      "step": 2650
    },
    {
      "epoch": 18.0,
      "eval_accuracy": 0.9173652694610779,
      "eval_loss": 0.39937448501586914,
      "eval_runtime": 57.7711,
      "eval_samples_per_second": 14.454,
      "eval_steps_per_second": 0.467,
      "step": 2664
    },
    {
      "epoch": 18.074324324324323,
      "grad_norm": 117016.4296875,
      "learning_rate": 1.0516605166051662e-05,
      "loss": 0.3622,
      "step": 2675
    },
    {
      "epoch": 18.243243243243242,
      "grad_norm": 101109.75,
      "learning_rate": 9.59409594095941e-06,
      "loss": 0.3351,
      "step": 2700
    },
    {
      "epoch": 18.41216216216216,
      "grad_norm": 53296.8828125,
      "learning_rate": 8.67158671586716e-06,
      "loss": 0.3686,
      "step": 2725
    },
    {
      "epoch": 18.58108108108108,
      "grad_norm": 50223.7109375,
      "learning_rate": 7.749077490774908e-06,
      "loss": 0.3327,
      "step": 2750
    },
    {
      "epoch": 18.75,
      "grad_norm": 174233.390625,
      "learning_rate": 6.826568265682657e-06,
      "loss": 0.331,
      "step": 2775
    },
    {
      "epoch": 18.91891891891892,
      "grad_norm": 88070.078125,
      "learning_rate": 5.904059040590406e-06,
      "loss": 0.3867,
      "step": 2800
    },
    {
      "epoch": 19.0,
      "eval_accuracy": 0.9173652694610779,
      "eval_loss": 0.39366012811660767,
      "eval_runtime": 57.4941,
      "eval_samples_per_second": 14.523,
      "eval_steps_per_second": 0.47,
      "step": 2812
    },
    {
      "epoch": 19.08783783783784,
      "grad_norm": 51424.33203125,
      "learning_rate": 4.981549815498155e-06,
      "loss": 0.3599,
      "step": 2825
    },
    {
      "epoch": 19.256756756756758,
      "grad_norm": 65029.12890625,
      "learning_rate": 4.059040590405904e-06,
      "loss": 0.3326,
      "step": 2850
    },
    {
      "epoch": 19.425675675675677,
      "grad_norm": 162222.421875,
      "learning_rate": 3.136531365313653e-06,
      "loss": 0.371,
      "step": 2875
    },
    {
      "epoch": 19.594594594594593,
      "grad_norm": 87439.921875,
      "learning_rate": 2.2140221402214023e-06,
      "loss": 0.3432,
      "step": 2900
    },
    {
      "epoch": 19.763513513513512,
      "grad_norm": 127598.703125,
      "learning_rate": 1.2915129151291513e-06,
      "loss": 0.3554,
      "step": 2925
    },
    {
      "epoch": 19.93243243243243,
      "grad_norm": 61254.16796875,
      "learning_rate": 3.690036900369004e-07,
      "loss": 0.3154,
      "step": 2950
    },
    {
      "epoch": 20.0,
      "eval_accuracy": 0.9197604790419162,
      "eval_loss": 0.3974940776824951,
      "eval_runtime": 57.4948,
      "eval_samples_per_second": 14.523,
      "eval_steps_per_second": 0.47,
      "step": 2960
    },
    {
      "epoch": 20.0,
      "step": 2960,
      "total_flos": 0.0,
      "train_loss": 0.43193954454885947,
      "train_runtime": 10921.5934,
      "train_samples_per_second": 8.66,
      "train_steps_per_second": 0.271
    }
  ],
  "logging_steps": 25,
  "max_steps": 2960,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 20,
  "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": 0.0,
  "train_batch_size": 32,
  "trial_name": null,
  "trial_params": null
}