{
  "best_metric": 0.39134106040000916,
  "best_model_checkpoint": "limb_classification_person_crop_seq/t4_4heads_1layers_1e-4lr/checkpoint-2368",
  "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": 357031.3125,
      "learning_rate": 1e-05,
      "loss": 1.3943,
      "step": 25
    },
    {
      "epoch": 0.33783783783783783,
      "grad_norm": 254554.484375,
      "learning_rate": 2e-05,
      "loss": 1.3595,
      "step": 50
    },
    {
      "epoch": 0.5067567567567568,
      "grad_norm": 133251.71875,
      "learning_rate": 3e-05,
      "loss": 1.2984,
      "step": 75
    },
    {
      "epoch": 0.6756756756756757,
      "grad_norm": 592813.0625,
      "learning_rate": 4e-05,
      "loss": 1.1902,
      "step": 100
    },
    {
      "epoch": 0.8445945945945946,
      "grad_norm": 185851.640625,
      "learning_rate": 5e-05,
      "loss": 1.0589,
      "step": 125
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.8754491017964072,
      "eval_loss": 0.5867581963539124,
      "eval_runtime": 58.1404,
      "eval_samples_per_second": 14.362,
      "eval_steps_per_second": 0.464,
      "step": 148
    },
    {
      "epoch": 1.0135135135135136,
      "grad_norm": 182520.671875,
      "learning_rate": 6e-05,
      "loss": 0.7488,
      "step": 150
    },
    {
      "epoch": 1.1824324324324325,
      "grad_norm": 137231.96875,
      "learning_rate": 7e-05,
      "loss": 0.5925,
      "step": 175
    },
    {
      "epoch": 1.3513513513513513,
      "grad_norm": 171974.703125,
      "learning_rate": 8e-05,
      "loss": 0.5257,
      "step": 200
    },
    {
      "epoch": 1.5202702702702702,
      "grad_norm": 231897.359375,
      "learning_rate": 9e-05,
      "loss": 0.4506,
      "step": 225
    },
    {
      "epoch": 1.689189189189189,
      "grad_norm": 104676.7421875,
      "learning_rate": 0.0001,
      "loss": 0.4391,
      "step": 250
    },
    {
      "epoch": 1.8581081081081081,
      "grad_norm": 205613.765625,
      "learning_rate": 9.907749077490776e-05,
      "loss": 0.429,
      "step": 275
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.8970059880239521,
      "eval_loss": 0.425736665725708,
      "eval_runtime": 57.872,
      "eval_samples_per_second": 14.428,
      "eval_steps_per_second": 0.467,
      "step": 296
    },
    {
      "epoch": 2.027027027027027,
      "grad_norm": 433545.125,
      "learning_rate": 9.81549815498155e-05,
      "loss": 0.4631,
      "step": 300
    },
    {
      "epoch": 2.195945945945946,
      "grad_norm": 126672.9609375,
      "learning_rate": 9.723247232472326e-05,
      "loss": 0.439,
      "step": 325
    },
    {
      "epoch": 2.364864864864865,
      "grad_norm": 122062.65625,
      "learning_rate": 9.6309963099631e-05,
      "loss": 0.4457,
      "step": 350
    },
    {
      "epoch": 2.5337837837837838,
      "grad_norm": 176409.59375,
      "learning_rate": 9.538745387453874e-05,
      "loss": 0.4201,
      "step": 375
    },
    {
      "epoch": 2.7027027027027026,
      "grad_norm": 91792.4453125,
      "learning_rate": 9.44649446494465e-05,
      "loss": 0.4323,
      "step": 400
    },
    {
      "epoch": 2.8716216216216215,
      "grad_norm": 254908.78125,
      "learning_rate": 9.354243542435425e-05,
      "loss": 0.4438,
      "step": 425
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.9077844311377246,
      "eval_loss": 0.4206658601760864,
      "eval_runtime": 57.1877,
      "eval_samples_per_second": 14.601,
      "eval_steps_per_second": 0.472,
      "step": 444
    },
    {
      "epoch": 3.0405405405405403,
      "grad_norm": 59820.00390625,
      "learning_rate": 9.2619926199262e-05,
      "loss": 0.433,
      "step": 450
    },
    {
      "epoch": 3.2094594594594597,
      "grad_norm": 118211.828125,
      "learning_rate": 9.169741697416975e-05,
      "loss": 0.4325,
      "step": 475
    },
    {
      "epoch": 3.3783783783783785,
      "grad_norm": 108471.75,
      "learning_rate": 9.077490774907749e-05,
      "loss": 0.4089,
      "step": 500
    },
    {
      "epoch": 3.5472972972972974,
      "grad_norm": 89766.4921875,
      "learning_rate": 8.985239852398525e-05,
      "loss": 0.3995,
      "step": 525
    },
    {
      "epoch": 3.7162162162162162,
      "grad_norm": 116075.578125,
      "learning_rate": 8.892988929889299e-05,
      "loss": 0.4442,
      "step": 550
    },
    {
      "epoch": 3.885135135135135,
      "grad_norm": 88052.2421875,
      "learning_rate": 8.800738007380073e-05,
      "loss": 0.4218,
      "step": 575
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.9005988023952096,
      "eval_loss": 0.42495548725128174,
      "eval_runtime": 57.8238,
      "eval_samples_per_second": 14.44,
      "eval_steps_per_second": 0.467,
      "step": 592
    },
    {
      "epoch": 4.054054054054054,
      "grad_norm": 92320.6484375,
      "learning_rate": 8.708487084870849e-05,
      "loss": 0.474,
      "step": 600
    },
    {
      "epoch": 4.222972972972973,
      "grad_norm": 67100.453125,
      "learning_rate": 8.616236162361624e-05,
      "loss": 0.3908,
      "step": 625
    },
    {
      "epoch": 4.391891891891892,
      "grad_norm": 137585.5,
      "learning_rate": 8.523985239852399e-05,
      "loss": 0.4137,
      "step": 650
    },
    {
      "epoch": 4.5608108108108105,
      "grad_norm": 82812.984375,
      "learning_rate": 8.431734317343174e-05,
      "loss": 0.4281,
      "step": 675
    },
    {
      "epoch": 4.72972972972973,
      "grad_norm": 120294.703125,
      "learning_rate": 8.339483394833948e-05,
      "loss": 0.4272,
      "step": 700
    },
    {
      "epoch": 4.898648648648649,
      "grad_norm": 99253.6484375,
      "learning_rate": 8.247232472324724e-05,
      "loss": 0.4232,
      "step": 725
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.9197604790419162,
      "eval_loss": 0.3950238525867462,
      "eval_runtime": 58.5088,
      "eval_samples_per_second": 14.271,
      "eval_steps_per_second": 0.461,
      "step": 740
    },
    {
      "epoch": 5.0675675675675675,
      "grad_norm": 197806.0625,
      "learning_rate": 8.154981549815498e-05,
      "loss": 0.411,
      "step": 750
    },
    {
      "epoch": 5.236486486486487,
      "grad_norm": 79927.8125,
      "learning_rate": 8.062730627306274e-05,
      "loss": 0.3884,
      "step": 775
    },
    {
      "epoch": 5.405405405405405,
      "grad_norm": 101610.2578125,
      "learning_rate": 7.970479704797048e-05,
      "loss": 0.4105,
      "step": 800
    },
    {
      "epoch": 5.574324324324325,
      "grad_norm": 158660.6875,
      "learning_rate": 7.878228782287823e-05,
      "loss": 0.4189,
      "step": 825
    },
    {
      "epoch": 5.743243243243243,
      "grad_norm": 109132.0390625,
      "learning_rate": 7.785977859778598e-05,
      "loss": 0.4072,
      "step": 850
    },
    {
      "epoch": 5.912162162162162,
      "grad_norm": 73329.078125,
      "learning_rate": 7.693726937269373e-05,
      "loss": 0.4047,
      "step": 875
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.9197604790419162,
      "eval_loss": 0.3974027633666992,
      "eval_runtime": 60.2569,
      "eval_samples_per_second": 13.857,
      "eval_steps_per_second": 0.448,
      "step": 888
    },
    {
      "epoch": 6.081081081081081,
      "grad_norm": 115863.328125,
      "learning_rate": 7.601476014760149e-05,
      "loss": 0.4244,
      "step": 900
    },
    {
      "epoch": 6.25,
      "grad_norm": 109137.578125,
      "learning_rate": 7.509225092250923e-05,
      "loss": 0.4013,
      "step": 925
    },
    {
      "epoch": 6.418918918918919,
      "grad_norm": 112666.5625,
      "learning_rate": 7.416974169741697e-05,
      "loss": 0.4167,
      "step": 950
    },
    {
      "epoch": 6.587837837837838,
      "grad_norm": 169974.65625,
      "learning_rate": 7.324723247232473e-05,
      "loss": 0.4067,
      "step": 975
    },
    {
      "epoch": 6.756756756756757,
      "grad_norm": 150158.71875,
      "learning_rate": 7.232472324723247e-05,
      "loss": 0.3791,
      "step": 1000
    },
    {
      "epoch": 6.925675675675675,
      "grad_norm": 75048.8046875,
      "learning_rate": 7.140221402214023e-05,
      "loss": 0.3966,
      "step": 1025
    },
    {
      "epoch": 7.0,
      "eval_accuracy": 0.9173652694610779,
      "eval_loss": 0.3983464241027832,
      "eval_runtime": 59.6478,
      "eval_samples_per_second": 13.999,
      "eval_steps_per_second": 0.453,
      "step": 1036
    },
    {
      "epoch": 7.094594594594595,
      "grad_norm": 119260.1796875,
      "learning_rate": 7.047970479704797e-05,
      "loss": 0.4127,
      "step": 1050
    },
    {
      "epoch": 7.263513513513513,
      "grad_norm": 34167.7734375,
      "learning_rate": 6.955719557195572e-05,
      "loss": 0.4047,
      "step": 1075
    },
    {
      "epoch": 7.4324324324324325,
      "grad_norm": 124205.6640625,
      "learning_rate": 6.863468634686348e-05,
      "loss": 0.4237,
      "step": 1100
    },
    {
      "epoch": 7.601351351351351,
      "grad_norm": 146023.78125,
      "learning_rate": 6.771217712177122e-05,
      "loss": 0.3756,
      "step": 1125
    },
    {
      "epoch": 7.77027027027027,
      "grad_norm": 115511.2109375,
      "learning_rate": 6.678966789667896e-05,
      "loss": 0.3945,
      "step": 1150
    },
    {
      "epoch": 7.9391891891891895,
      "grad_norm": 120214.0390625,
      "learning_rate": 6.586715867158672e-05,
      "loss": 0.394,
      "step": 1175
    },
    {
      "epoch": 8.0,
      "eval_accuracy": 0.9161676646706587,
      "eval_loss": 0.40396106243133545,
      "eval_runtime": 59.1269,
      "eval_samples_per_second": 14.122,
      "eval_steps_per_second": 0.457,
      "step": 1184
    },
    {
      "epoch": 8.108108108108109,
      "grad_norm": 84249.5859375,
      "learning_rate": 6.494464944649446e-05,
      "loss": 0.3855,
      "step": 1200
    },
    {
      "epoch": 8.277027027027026,
      "grad_norm": 55820.75390625,
      "learning_rate": 6.402214022140222e-05,
      "loss": 0.3844,
      "step": 1225
    },
    {
      "epoch": 8.445945945945946,
      "grad_norm": 78390.0,
      "learning_rate": 6.309963099630997e-05,
      "loss": 0.4079,
      "step": 1250
    },
    {
      "epoch": 8.614864864864865,
      "grad_norm": 117951.7421875,
      "learning_rate": 6.217712177121771e-05,
      "loss": 0.37,
      "step": 1275
    },
    {
      "epoch": 8.783783783783784,
      "grad_norm": 86200.9375,
      "learning_rate": 6.125461254612547e-05,
      "loss": 0.4198,
      "step": 1300
    },
    {
      "epoch": 8.952702702702704,
      "grad_norm": 201158.421875,
      "learning_rate": 6.033210332103322e-05,
      "loss": 0.4226,
      "step": 1325
    },
    {
      "epoch": 9.0,
      "eval_accuracy": 0.918562874251497,
      "eval_loss": 0.39536988735198975,
      "eval_runtime": 59.6348,
      "eval_samples_per_second": 14.002,
      "eval_steps_per_second": 0.453,
      "step": 1332
    },
    {
      "epoch": 9.121621621621621,
      "grad_norm": 129266.46875,
      "learning_rate": 5.940959409594096e-05,
      "loss": 0.3987,
      "step": 1350
    },
    {
      "epoch": 9.29054054054054,
      "grad_norm": 163080.140625,
      "learning_rate": 5.848708487084871e-05,
      "loss": 0.4056,
      "step": 1375
    },
    {
      "epoch": 9.45945945945946,
      "grad_norm": 60286.765625,
      "learning_rate": 5.756457564575646e-05,
      "loss": 0.3886,
      "step": 1400
    },
    {
      "epoch": 9.628378378378379,
      "grad_norm": 99267.8125,
      "learning_rate": 5.664206642066421e-05,
      "loss": 0.3858,
      "step": 1425
    },
    {
      "epoch": 9.797297297297296,
      "grad_norm": 64863.20703125,
      "learning_rate": 5.5719557195571956e-05,
      "loss": 0.4,
      "step": 1450
    },
    {
      "epoch": 9.966216216216216,
      "grad_norm": 98722.5078125,
      "learning_rate": 5.479704797047971e-05,
      "loss": 0.3786,
      "step": 1475
    },
    {
      "epoch": 10.0,
      "eval_accuracy": 0.9173652694610779,
      "eval_loss": 0.3975362181663513,
      "eval_runtime": 59.427,
      "eval_samples_per_second": 14.051,
      "eval_steps_per_second": 0.454,
      "step": 1480
    },
    {
      "epoch": 10.135135135135135,
      "grad_norm": 45542.5703125,
      "learning_rate": 5.387453874538746e-05,
      "loss": 0.3975,
      "step": 1500
    },
    {
      "epoch": 10.304054054054054,
      "grad_norm": 186803.515625,
      "learning_rate": 5.295202952029521e-05,
      "loss": 0.4041,
      "step": 1525
    },
    {
      "epoch": 10.472972972972974,
      "grad_norm": 134740.75,
      "learning_rate": 5.202952029520295e-05,
      "loss": 0.3538,
      "step": 1550
    },
    {
      "epoch": 10.641891891891891,
      "grad_norm": 96566.1796875,
      "learning_rate": 5.11070110701107e-05,
      "loss": 0.3726,
      "step": 1575
    },
    {
      "epoch": 10.81081081081081,
      "grad_norm": 147596.359375,
      "learning_rate": 5.018450184501845e-05,
      "loss": 0.4219,
      "step": 1600
    },
    {
      "epoch": 10.97972972972973,
      "grad_norm": 106429.2265625,
      "learning_rate": 4.92619926199262e-05,
      "loss": 0.3887,
      "step": 1625
    },
    {
      "epoch": 11.0,
      "eval_accuracy": 0.9149700598802395,
      "eval_loss": 0.4018101394176483,
      "eval_runtime": 59.155,
      "eval_samples_per_second": 14.115,
      "eval_steps_per_second": 0.456,
      "step": 1628
    },
    {
      "epoch": 11.14864864864865,
      "grad_norm": 107516.53125,
      "learning_rate": 4.833948339483395e-05,
      "loss": 0.4087,
      "step": 1650
    },
    {
      "epoch": 11.317567567567568,
      "grad_norm": 58793.83984375,
      "learning_rate": 4.74169741697417e-05,
      "loss": 0.3958,
      "step": 1675
    },
    {
      "epoch": 11.486486486486486,
      "grad_norm": 146357.75,
      "learning_rate": 4.6494464944649444e-05,
      "loss": 0.3637,
      "step": 1700
    },
    {
      "epoch": 11.655405405405405,
      "grad_norm": 122970.703125,
      "learning_rate": 4.55719557195572e-05,
      "loss": 0.3806,
      "step": 1725
    },
    {
      "epoch": 11.824324324324325,
      "grad_norm": 222967.296875,
      "learning_rate": 4.464944649446495e-05,
      "loss": 0.3613,
      "step": 1750
    },
    {
      "epoch": 11.993243243243244,
      "grad_norm": 78722.59375,
      "learning_rate": 4.37269372693727e-05,
      "loss": 0.3966,
      "step": 1775
    },
    {
      "epoch": 12.0,
      "eval_accuracy": 0.9173652694610779,
      "eval_loss": 0.3930155038833618,
      "eval_runtime": 59.3577,
      "eval_samples_per_second": 14.067,
      "eval_steps_per_second": 0.455,
      "step": 1776
    },
    {
      "epoch": 12.162162162162161,
      "grad_norm": 117368.453125,
      "learning_rate": 4.280442804428044e-05,
      "loss": 0.3691,
      "step": 1800
    },
    {
      "epoch": 12.33108108108108,
      "grad_norm": 200185.265625,
      "learning_rate": 4.1881918819188195e-05,
      "loss": 0.3899,
      "step": 1825
    },
    {
      "epoch": 12.5,
      "grad_norm": 71087.515625,
      "learning_rate": 4.0959409594095944e-05,
      "loss": 0.3922,
      "step": 1850
    },
    {
      "epoch": 12.66891891891892,
      "grad_norm": 109740.953125,
      "learning_rate": 4.003690036900369e-05,
      "loss": 0.4048,
      "step": 1875
    },
    {
      "epoch": 12.837837837837839,
      "grad_norm": 79667.484375,
      "learning_rate": 3.911439114391144e-05,
      "loss": 0.382,
      "step": 1900
    },
    {
      "epoch": 13.0,
      "eval_accuracy": 0.9161676646706587,
      "eval_loss": 0.40007835626602173,
      "eval_runtime": 58.9675,
      "eval_samples_per_second": 14.16,
      "eval_steps_per_second": 0.458,
      "step": 1924
    },
    {
      "epoch": 13.006756756756756,
      "grad_norm": 98287.578125,
      "learning_rate": 3.819188191881919e-05,
      "loss": 0.3412,
      "step": 1925
    },
    {
      "epoch": 13.175675675675675,
      "grad_norm": 190703.640625,
      "learning_rate": 3.726937269372694e-05,
      "loss": 0.402,
      "step": 1950
    },
    {
      "epoch": 13.344594594594595,
      "grad_norm": 98728.4453125,
      "learning_rate": 3.634686346863469e-05,
      "loss": 0.3603,
      "step": 1975
    },
    {
      "epoch": 13.513513513513514,
      "grad_norm": 80456.5859375,
      "learning_rate": 3.542435424354244e-05,
      "loss": 0.3754,
      "step": 2000
    },
    {
      "epoch": 13.682432432432432,
      "grad_norm": 118070.0703125,
      "learning_rate": 3.4501845018450186e-05,
      "loss": 0.3599,
      "step": 2025
    },
    {
      "epoch": 13.85135135135135,
      "grad_norm": 62820.54296875,
      "learning_rate": 3.3579335793357934e-05,
      "loss": 0.3805,
      "step": 2050
    },
    {
      "epoch": 14.0,
      "eval_accuracy": 0.9137724550898204,
      "eval_loss": 0.4115696847438812,
      "eval_runtime": 59.0999,
      "eval_samples_per_second": 14.129,
      "eval_steps_per_second": 0.457,
      "step": 2072
    },
    {
      "epoch": 14.02027027027027,
      "grad_norm": 98389.9296875,
      "learning_rate": 3.265682656826568e-05,
      "loss": 0.3766,
      "step": 2075
    },
    {
      "epoch": 14.18918918918919,
      "grad_norm": 67428.9921875,
      "learning_rate": 3.173431734317343e-05,
      "loss": 0.367,
      "step": 2100
    },
    {
      "epoch": 14.358108108108109,
      "grad_norm": 94258.109375,
      "learning_rate": 3.081180811808118e-05,
      "loss": 0.3901,
      "step": 2125
    },
    {
      "epoch": 14.527027027027026,
      "grad_norm": 57570.140625,
      "learning_rate": 2.9889298892988933e-05,
      "loss": 0.3653,
      "step": 2150
    },
    {
      "epoch": 14.695945945945946,
      "grad_norm": 161211.34375,
      "learning_rate": 2.8966789667896682e-05,
      "loss": 0.385,
      "step": 2175
    },
    {
      "epoch": 14.864864864864865,
      "grad_norm": 92826.1875,
      "learning_rate": 2.8044280442804427e-05,
      "loss": 0.3781,
      "step": 2200
    },
    {
      "epoch": 15.0,
      "eval_accuracy": 0.9173652694610779,
      "eval_loss": 0.40341073274612427,
      "eval_runtime": 59.0972,
      "eval_samples_per_second": 14.129,
      "eval_steps_per_second": 0.457,
      "step": 2220
    },
    {
      "epoch": 15.033783783783784,
      "grad_norm": 185792.328125,
      "learning_rate": 2.7121771217712176e-05,
      "loss": 0.3753,
      "step": 2225
    },
    {
      "epoch": 15.202702702702704,
      "grad_norm": 173501.625,
      "learning_rate": 2.619926199261993e-05,
      "loss": 0.3739,
      "step": 2250
    },
    {
      "epoch": 15.371621621621621,
      "grad_norm": 76993.28125,
      "learning_rate": 2.5276752767527677e-05,
      "loss": 0.3513,
      "step": 2275
    },
    {
      "epoch": 15.54054054054054,
      "grad_norm": 95068.1953125,
      "learning_rate": 2.4354243542435426e-05,
      "loss": 0.3554,
      "step": 2300
    },
    {
      "epoch": 15.70945945945946,
      "grad_norm": 160660.328125,
      "learning_rate": 2.3431734317343175e-05,
      "loss": 0.3633,
      "step": 2325
    },
    {
      "epoch": 15.878378378378379,
      "grad_norm": 99192.015625,
      "learning_rate": 2.2509225092250924e-05,
      "loss": 0.3796,
      "step": 2350
    },
    {
      "epoch": 16.0,
      "eval_accuracy": 0.9233532934131736,
      "eval_loss": 0.39134106040000916,
      "eval_runtime": 58.6893,
      "eval_samples_per_second": 14.227,
      "eval_steps_per_second": 0.46,
      "step": 2368
    },
    {
      "epoch": 16.0472972972973,
      "grad_norm": 67577.7890625,
      "learning_rate": 2.1586715867158673e-05,
      "loss": 0.3657,
      "step": 2375
    },
    {
      "epoch": 16.216216216216218,
      "grad_norm": 75403.8046875,
      "learning_rate": 2.066420664206642e-05,
      "loss": 0.3427,
      "step": 2400
    },
    {
      "epoch": 16.385135135135137,
      "grad_norm": 63007.6875,
      "learning_rate": 1.974169741697417e-05,
      "loss": 0.3491,
      "step": 2425
    },
    {
      "epoch": 16.554054054054053,
      "grad_norm": 191865.390625,
      "learning_rate": 1.881918819188192e-05,
      "loss": 0.3809,
      "step": 2450
    },
    {
      "epoch": 16.722972972972972,
      "grad_norm": 83243.609375,
      "learning_rate": 1.7896678966789668e-05,
      "loss": 0.3475,
      "step": 2475
    },
    {
      "epoch": 16.89189189189189,
      "grad_norm": 39799.421875,
      "learning_rate": 1.6974169741697417e-05,
      "loss": 0.3522,
      "step": 2500
    },
    {
      "epoch": 17.0,
      "eval_accuracy": 0.9209580838323354,
      "eval_loss": 0.3977585732936859,
      "eval_runtime": 59.2068,
      "eval_samples_per_second": 14.103,
      "eval_steps_per_second": 0.456,
      "step": 2516
    },
    {
      "epoch": 17.06081081081081,
      "grad_norm": 152904.921875,
      "learning_rate": 1.6051660516605166e-05,
      "loss": 0.3973,
      "step": 2525
    },
    {
      "epoch": 17.22972972972973,
      "grad_norm": 72565.2109375,
      "learning_rate": 1.5129151291512916e-05,
      "loss": 0.3491,
      "step": 2550
    },
    {
      "epoch": 17.39864864864865,
      "grad_norm": 63665.20703125,
      "learning_rate": 1.4206642066420663e-05,
      "loss": 0.3641,
      "step": 2575
    },
    {
      "epoch": 17.56756756756757,
      "grad_norm": 53316.6953125,
      "learning_rate": 1.3284132841328414e-05,
      "loss": 0.3507,
      "step": 2600
    },
    {
      "epoch": 17.736486486486488,
      "grad_norm": 115725.8671875,
      "learning_rate": 1.2361623616236164e-05,
      "loss": 0.3399,
      "step": 2625
    },
    {
      "epoch": 17.905405405405407,
      "grad_norm": 37301.25390625,
      "learning_rate": 1.1439114391143913e-05,
      "loss": 0.3799,
      "step": 2650
    },
    {
      "epoch": 18.0,
      "eval_accuracy": 0.9197604790419162,
      "eval_loss": 0.39950963854789734,
      "eval_runtime": 58.9959,
      "eval_samples_per_second": 14.154,
      "eval_steps_per_second": 0.458,
      "step": 2664
    },
    {
      "epoch": 18.074324324324323,
      "grad_norm": 119603.8203125,
      "learning_rate": 1.0516605166051662e-05,
      "loss": 0.3536,
      "step": 2675
    },
    {
      "epoch": 18.243243243243242,
      "grad_norm": 44970.4140625,
      "learning_rate": 9.59409594095941e-06,
      "loss": 0.3228,
      "step": 2700
    },
    {
      "epoch": 18.41216216216216,
      "grad_norm": 118582.2734375,
      "learning_rate": 8.67158671586716e-06,
      "loss": 0.3729,
      "step": 2725
    },
    {
      "epoch": 18.58108108108108,
      "grad_norm": 35576.6171875,
      "learning_rate": 7.749077490774908e-06,
      "loss": 0.3247,
      "step": 2750
    },
    {
      "epoch": 18.75,
      "grad_norm": 157502.953125,
      "learning_rate": 6.826568265682657e-06,
      "loss": 0.3417,
      "step": 2775
    },
    {
      "epoch": 18.91891891891892,
      "grad_norm": 55949.453125,
      "learning_rate": 5.904059040590406e-06,
      "loss": 0.3741,
      "step": 2800
    },
    {
      "epoch": 19.0,
      "eval_accuracy": 0.918562874251497,
      "eval_loss": 0.39410221576690674,
      "eval_runtime": 58.4623,
      "eval_samples_per_second": 14.283,
      "eval_steps_per_second": 0.462,
      "step": 2812
    },
    {
      "epoch": 19.08783783783784,
      "grad_norm": 104191.796875,
      "learning_rate": 4.981549815498155e-06,
      "loss": 0.3745,
      "step": 2825
    },
    {
      "epoch": 19.256756756756758,
      "grad_norm": 78876.4453125,
      "learning_rate": 4.059040590405904e-06,
      "loss": 0.3366,
      "step": 2850
    },
    {
      "epoch": 19.425675675675677,
      "grad_norm": 94403.84375,
      "learning_rate": 3.136531365313653e-06,
      "loss": 0.3692,
      "step": 2875
    },
    {
      "epoch": 19.594594594594593,
      "grad_norm": 106615.25,
      "learning_rate": 2.2140221402214023e-06,
      "loss": 0.3458,
      "step": 2900
    },
    {
      "epoch": 19.763513513513512,
      "grad_norm": 127746.890625,
      "learning_rate": 1.2915129151291513e-06,
      "loss": 0.3527,
      "step": 2925
    },
    {
      "epoch": 19.93243243243243,
      "grad_norm": 103114.9765625,
      "learning_rate": 3.690036900369004e-07,
      "loss": 0.3208,
      "step": 2950
    },
    {
      "epoch": 20.0,
      "eval_accuracy": 0.9221556886227545,
      "eval_loss": 0.3963671326637268,
      "eval_runtime": 58.8449,
      "eval_samples_per_second": 14.19,
      "eval_steps_per_second": 0.459,
      "step": 2960
    },
    {
      "epoch": 20.0,
      "step": 2960,
      "total_flos": 0.0,
      "train_loss": 0.43214384381835524,
      "train_runtime": 11030.2328,
      "train_samples_per_second": 8.575,
      "train_steps_per_second": 0.268
    }
  ],
  "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
}