{
  "best_metric": 0.09645664691925049,
  "best_model_checkpoint": "vit-base-patch16-224-in21k-finetuned_v2024-7-24-frost/checkpoint-500",
  "epoch": 30.0,
  "eval_steps": 100,
  "global_step": 1920,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.15625,
      "grad_norm": 0.3123115003108978,
      "learning_rate": 1.0416666666666668e-05,
      "loss": 0.0743,
      "step": 10
    },
    {
      "epoch": 0.3125,
      "grad_norm": 0.10650705546140671,
      "learning_rate": 2.0833333333333336e-05,
      "loss": 0.0699,
      "step": 20
    },
    {
      "epoch": 0.46875,
      "grad_norm": 0.40636828541755676,
      "learning_rate": 3.125e-05,
      "loss": 0.0732,
      "step": 30
    },
    {
      "epoch": 0.625,
      "grad_norm": 0.42912840843200684,
      "learning_rate": 4.166666666666667e-05,
      "loss": 0.075,
      "step": 40
    },
    {
      "epoch": 0.78125,
      "grad_norm": 0.3166373372077942,
      "learning_rate": 5.208333333333334e-05,
      "loss": 0.0695,
      "step": 50
    },
    {
      "epoch": 0.9375,
      "grad_norm": 0.8551476001739502,
      "learning_rate": 6.25e-05,
      "loss": 0.0883,
      "step": 60
    },
    {
      "epoch": 1.09375,
      "grad_norm": 0.18066875636577606,
      "learning_rate": 7.291666666666667e-05,
      "loss": 0.0699,
      "step": 70
    },
    {
      "epoch": 1.25,
      "grad_norm": 0.28325945138931274,
      "learning_rate": 8.333333333333334e-05,
      "loss": 0.0627,
      "step": 80
    },
    {
      "epoch": 1.40625,
      "grad_norm": 0.3701513409614563,
      "learning_rate": 9.375e-05,
      "loss": 0.0866,
      "step": 90
    },
    {
      "epoch": 1.5625,
      "grad_norm": 0.35587912797927856,
      "learning_rate": 0.00010416666666666667,
      "loss": 0.0728,
      "step": 100
    },
    {
      "epoch": 1.5625,
      "eval_accuracy": 0.984070796460177,
      "eval_f1": 0.9606986899563319,
      "eval_loss": 0.06593623757362366,
      "eval_precision": 0.9691629955947136,
      "eval_recall": 0.9523809523809523,
      "eval_runtime": 0.9039,
      "eval_samples_per_second": 125.009,
      "eval_steps_per_second": 16.594,
      "step": 100
    },
    {
      "epoch": 1.71875,
      "grad_norm": 0.23759329319000244,
      "learning_rate": 0.00011458333333333333,
      "loss": 0.0653,
      "step": 110
    },
    {
      "epoch": 1.875,
      "grad_norm": 0.6092272996902466,
      "learning_rate": 0.000125,
      "loss": 0.1015,
      "step": 120
    },
    {
      "epoch": 2.03125,
      "grad_norm": 0.15887708961963654,
      "learning_rate": 0.0001354166666666667,
      "loss": 0.09,
      "step": 130
    },
    {
      "epoch": 2.1875,
      "grad_norm": 0.3399417996406555,
      "learning_rate": 0.00014583333333333335,
      "loss": 0.0847,
      "step": 140
    },
    {
      "epoch": 2.34375,
      "grad_norm": 0.2599344253540039,
      "learning_rate": 0.00015625,
      "loss": 0.0722,
      "step": 150
    },
    {
      "epoch": 2.5,
      "grad_norm": 0.20714014768600464,
      "learning_rate": 0.0001666666666666667,
      "loss": 0.0915,
      "step": 160
    },
    {
      "epoch": 2.65625,
      "grad_norm": 0.7900287508964539,
      "learning_rate": 0.00017708333333333335,
      "loss": 0.1008,
      "step": 170
    },
    {
      "epoch": 2.8125,
      "grad_norm": 0.23315797746181488,
      "learning_rate": 0.0001875,
      "loss": 0.1142,
      "step": 180
    },
    {
      "epoch": 2.96875,
      "grad_norm": 1.258319616317749,
      "learning_rate": 0.0001979166666666667,
      "loss": 0.1027,
      "step": 190
    },
    {
      "epoch": 3.125,
      "grad_norm": 1.0168662071228027,
      "learning_rate": 0.0001990740740740741,
      "loss": 0.0871,
      "step": 200
    },
    {
      "epoch": 3.125,
      "eval_accuracy": 0.9566371681415929,
      "eval_f1": 0.8941684665226782,
      "eval_loss": 0.12436065077781677,
      "eval_precision": 0.8922413793103449,
      "eval_recall": 0.8961038961038961,
      "eval_runtime": 0.8601,
      "eval_samples_per_second": 131.383,
      "eval_steps_per_second": 17.44,
      "step": 200
    },
    {
      "epoch": 3.28125,
      "grad_norm": 0.38566353917121887,
      "learning_rate": 0.0001979166666666667,
      "loss": 0.1166,
      "step": 210
    },
    {
      "epoch": 3.4375,
      "grad_norm": 0.4687894284725189,
      "learning_rate": 0.00019675925925925926,
      "loss": 0.108,
      "step": 220
    },
    {
      "epoch": 3.59375,
      "grad_norm": 0.5190223455429077,
      "learning_rate": 0.00019560185185185186,
      "loss": 0.0901,
      "step": 230
    },
    {
      "epoch": 3.75,
      "grad_norm": 0.5094243288040161,
      "learning_rate": 0.00019444444444444446,
      "loss": 0.1144,
      "step": 240
    },
    {
      "epoch": 3.90625,
      "grad_norm": 0.5921277403831482,
      "learning_rate": 0.00019328703703703706,
      "loss": 0.1196,
      "step": 250
    },
    {
      "epoch": 4.0625,
      "grad_norm": 0.23840609192848206,
      "learning_rate": 0.00019212962962962963,
      "loss": 0.1122,
      "step": 260
    },
    {
      "epoch": 4.21875,
      "grad_norm": 0.9276812672615051,
      "learning_rate": 0.00019097222222222223,
      "loss": 0.1147,
      "step": 270
    },
    {
      "epoch": 4.375,
      "grad_norm": 0.7325614094734192,
      "learning_rate": 0.00018981481481481483,
      "loss": 0.1075,
      "step": 280
    },
    {
      "epoch": 4.53125,
      "grad_norm": 0.5574468374252319,
      "learning_rate": 0.00018865740740740743,
      "loss": 0.0958,
      "step": 290
    },
    {
      "epoch": 4.6875,
      "grad_norm": 0.3893429934978485,
      "learning_rate": 0.0001875,
      "loss": 0.0999,
      "step": 300
    },
    {
      "epoch": 4.6875,
      "eval_accuracy": 0.963716814159292,
      "eval_f1": 0.9125799573560768,
      "eval_loss": 0.10427873581647873,
      "eval_precision": 0.8991596638655462,
      "eval_recall": 0.9264069264069265,
      "eval_runtime": 0.8493,
      "eval_samples_per_second": 133.049,
      "eval_steps_per_second": 17.661,
      "step": 300
    },
    {
      "epoch": 4.84375,
      "grad_norm": 0.49028488993644714,
      "learning_rate": 0.0001863425925925926,
      "loss": 0.1087,
      "step": 310
    },
    {
      "epoch": 5.0,
      "grad_norm": 0.6510241627693176,
      "learning_rate": 0.0001851851851851852,
      "loss": 0.0949,
      "step": 320
    },
    {
      "epoch": 5.15625,
      "grad_norm": 0.3244408667087555,
      "learning_rate": 0.00018402777777777778,
      "loss": 0.0957,
      "step": 330
    },
    {
      "epoch": 5.3125,
      "grad_norm": 0.32894158363342285,
      "learning_rate": 0.00018287037037037038,
      "loss": 0.0761,
      "step": 340
    },
    {
      "epoch": 5.46875,
      "grad_norm": 0.4168912470340729,
      "learning_rate": 0.00018171296296296297,
      "loss": 0.1014,
      "step": 350
    },
    {
      "epoch": 5.625,
      "grad_norm": 0.30746978521347046,
      "learning_rate": 0.00018055555555555557,
      "loss": 0.0789,
      "step": 360
    },
    {
      "epoch": 5.78125,
      "grad_norm": 0.3337535262107849,
      "learning_rate": 0.00017939814814814815,
      "loss": 0.0891,
      "step": 370
    },
    {
      "epoch": 5.9375,
      "grad_norm": 0.2659320533275604,
      "learning_rate": 0.00017824074074074075,
      "loss": 0.0798,
      "step": 380
    },
    {
      "epoch": 6.09375,
      "grad_norm": 0.28791913390159607,
      "learning_rate": 0.00017708333333333335,
      "loss": 0.0961,
      "step": 390
    },
    {
      "epoch": 6.25,
      "grad_norm": 0.41803187131881714,
      "learning_rate": 0.00017592592592592595,
      "loss": 0.0743,
      "step": 400
    },
    {
      "epoch": 6.25,
      "eval_accuracy": 0.9610619469026549,
      "eval_f1": 0.9043478260869565,
      "eval_loss": 0.10431604832410812,
      "eval_precision": 0.9082969432314411,
      "eval_recall": 0.9004329004329005,
      "eval_runtime": 1.3126,
      "eval_samples_per_second": 86.086,
      "eval_steps_per_second": 11.427,
      "step": 400
    },
    {
      "epoch": 6.40625,
      "grad_norm": 0.398034930229187,
      "learning_rate": 0.00017476851851851852,
      "loss": 0.0798,
      "step": 410
    },
    {
      "epoch": 6.5625,
      "grad_norm": 0.533364474773407,
      "learning_rate": 0.00017361111111111112,
      "loss": 0.0808,
      "step": 420
    },
    {
      "epoch": 6.71875,
      "grad_norm": 0.6189862489700317,
      "learning_rate": 0.00017245370370370372,
      "loss": 0.091,
      "step": 430
    },
    {
      "epoch": 6.875,
      "grad_norm": 0.31593209505081177,
      "learning_rate": 0.00017129629629629632,
      "loss": 0.0729,
      "step": 440
    },
    {
      "epoch": 7.03125,
      "grad_norm": 0.3167741000652313,
      "learning_rate": 0.0001701388888888889,
      "loss": 0.0796,
      "step": 450
    },
    {
      "epoch": 7.1875,
      "grad_norm": 0.30901169776916504,
      "learning_rate": 0.0001689814814814815,
      "loss": 0.0867,
      "step": 460
    },
    {
      "epoch": 7.34375,
      "grad_norm": 0.13378705084323883,
      "learning_rate": 0.0001678240740740741,
      "loss": 0.0701,
      "step": 470
    },
    {
      "epoch": 7.5,
      "grad_norm": 0.15507709980010986,
      "learning_rate": 0.0001666666666666667,
      "loss": 0.0789,
      "step": 480
    },
    {
      "epoch": 7.65625,
      "grad_norm": 0.21113860607147217,
      "learning_rate": 0.00016550925925925926,
      "loss": 0.0647,
      "step": 490
    },
    {
      "epoch": 7.8125,
      "grad_norm": 0.15848499536514282,
      "learning_rate": 0.00016435185185185186,
      "loss": 0.0655,
      "step": 500
    },
    {
      "epoch": 7.8125,
      "eval_accuracy": 0.963716814159292,
      "eval_f1": 0.9118279569892475,
      "eval_loss": 0.09645664691925049,
      "eval_precision": 0.905982905982906,
      "eval_recall": 0.9177489177489178,
      "eval_runtime": 2.2304,
      "eval_samples_per_second": 50.664,
      "eval_steps_per_second": 6.725,
      "step": 500
    },
    {
      "epoch": 7.96875,
      "grad_norm": 0.19086486101150513,
      "learning_rate": 0.00016319444444444446,
      "loss": 0.0502,
      "step": 510
    },
    {
      "epoch": 8.125,
      "grad_norm": 0.4851354956626892,
      "learning_rate": 0.00016203703703703706,
      "loss": 0.0646,
      "step": 520
    },
    {
      "epoch": 8.28125,
      "grad_norm": 0.43803560733795166,
      "learning_rate": 0.00016087962962962963,
      "loss": 0.0668,
      "step": 530
    },
    {
      "epoch": 8.4375,
      "grad_norm": 0.26552197337150574,
      "learning_rate": 0.00015972222222222223,
      "loss": 0.0549,
      "step": 540
    },
    {
      "epoch": 8.59375,
      "grad_norm": 0.18909405171871185,
      "learning_rate": 0.00015856481481481483,
      "loss": 0.07,
      "step": 550
    },
    {
      "epoch": 8.75,
      "grad_norm": 0.2485276311635971,
      "learning_rate": 0.00015740740740740743,
      "loss": 0.0525,
      "step": 560
    },
    {
      "epoch": 8.90625,
      "grad_norm": 0.4424391984939575,
      "learning_rate": 0.00015625,
      "loss": 0.0775,
      "step": 570
    },
    {
      "epoch": 9.0625,
      "grad_norm": 0.2584344446659088,
      "learning_rate": 0.0001550925925925926,
      "loss": 0.064,
      "step": 580
    },
    {
      "epoch": 9.21875,
      "grad_norm": 0.6115286946296692,
      "learning_rate": 0.0001539351851851852,
      "loss": 0.0599,
      "step": 590
    },
    {
      "epoch": 9.375,
      "grad_norm": 0.5155323147773743,
      "learning_rate": 0.00015277777777777777,
      "loss": 0.0559,
      "step": 600
    },
    {
      "epoch": 9.375,
      "eval_accuracy": 0.9619469026548673,
      "eval_f1": 0.9087048832271762,
      "eval_loss": 0.10384609550237656,
      "eval_precision": 0.8916666666666667,
      "eval_recall": 0.9264069264069265,
      "eval_runtime": 2.2545,
      "eval_samples_per_second": 50.123,
      "eval_steps_per_second": 6.653,
      "step": 600
    },
    {
      "epoch": 9.53125,
      "grad_norm": 0.40844494104385376,
      "learning_rate": 0.00015162037037037037,
      "loss": 0.072,
      "step": 610
    },
    {
      "epoch": 9.6875,
      "grad_norm": 0.2091340720653534,
      "learning_rate": 0.00015046296296296297,
      "loss": 0.0513,
      "step": 620
    },
    {
      "epoch": 9.84375,
      "grad_norm": 0.22117160260677338,
      "learning_rate": 0.00014930555555555557,
      "loss": 0.0623,
      "step": 630
    },
    {
      "epoch": 10.0,
      "grad_norm": 0.31236401200294495,
      "learning_rate": 0.00014814814814814815,
      "loss": 0.0534,
      "step": 640
    },
    {
      "epoch": 10.15625,
      "grad_norm": 0.31281912326812744,
      "learning_rate": 0.00014699074074074075,
      "loss": 0.0644,
      "step": 650
    },
    {
      "epoch": 10.3125,
      "grad_norm": 0.5201927423477173,
      "learning_rate": 0.00014583333333333335,
      "loss": 0.057,
      "step": 660
    },
    {
      "epoch": 10.46875,
      "grad_norm": 0.2596763074398041,
      "learning_rate": 0.00014467592592592594,
      "loss": 0.0542,
      "step": 670
    },
    {
      "epoch": 10.625,
      "grad_norm": 0.3063810467720032,
      "learning_rate": 0.00014351851851851852,
      "loss": 0.0389,
      "step": 680
    },
    {
      "epoch": 10.78125,
      "grad_norm": 0.48713332414627075,
      "learning_rate": 0.00014236111111111112,
      "loss": 0.0742,
      "step": 690
    },
    {
      "epoch": 10.9375,
      "grad_norm": 0.21316884458065033,
      "learning_rate": 0.00014120370370370372,
      "loss": 0.0517,
      "step": 700
    },
    {
      "epoch": 10.9375,
      "eval_accuracy": 0.9584070796460177,
      "eval_f1": 0.8997867803837952,
      "eval_loss": 0.09719711542129517,
      "eval_precision": 0.8865546218487395,
      "eval_recall": 0.9134199134199135,
      "eval_runtime": 0.8764,
      "eval_samples_per_second": 128.937,
      "eval_steps_per_second": 17.115,
      "step": 700
    },
    {
      "epoch": 11.09375,
      "grad_norm": 0.9216361045837402,
      "learning_rate": 0.00014004629629629632,
      "loss": 0.0623,
      "step": 710
    },
    {
      "epoch": 11.25,
      "grad_norm": 0.31130528450012207,
      "learning_rate": 0.0001388888888888889,
      "loss": 0.0641,
      "step": 720
    },
    {
      "epoch": 11.40625,
      "grad_norm": 1.03948974609375,
      "learning_rate": 0.0001377314814814815,
      "loss": 0.0594,
      "step": 730
    },
    {
      "epoch": 11.5625,
      "grad_norm": 0.12757237255573273,
      "learning_rate": 0.0001365740740740741,
      "loss": 0.0572,
      "step": 740
    },
    {
      "epoch": 11.71875,
      "grad_norm": 0.25488346815109253,
      "learning_rate": 0.0001354166666666667,
      "loss": 0.0533,
      "step": 750
    },
    {
      "epoch": 11.875,
      "grad_norm": 0.2517576813697815,
      "learning_rate": 0.00013425925925925926,
      "loss": 0.0557,
      "step": 760
    },
    {
      "epoch": 12.03125,
      "grad_norm": 0.14332328736782074,
      "learning_rate": 0.00013310185185185186,
      "loss": 0.0433,
      "step": 770
    },
    {
      "epoch": 12.1875,
      "grad_norm": 0.7062014937400818,
      "learning_rate": 0.00013194444444444446,
      "loss": 0.0569,
      "step": 780
    },
    {
      "epoch": 12.34375,
      "grad_norm": 0.727057158946991,
      "learning_rate": 0.00013078703703703706,
      "loss": 0.0443,
      "step": 790
    },
    {
      "epoch": 12.5,
      "grad_norm": 0.17331984639167786,
      "learning_rate": 0.00012962962962962963,
      "loss": 0.0407,
      "step": 800
    },
    {
      "epoch": 12.5,
      "eval_accuracy": 0.963716814159292,
      "eval_f1": 0.9110629067245118,
      "eval_loss": 0.11198227852582932,
      "eval_precision": 0.9130434782608695,
      "eval_recall": 0.9090909090909091,
      "eval_runtime": 0.9019,
      "eval_samples_per_second": 125.294,
      "eval_steps_per_second": 16.632,
      "step": 800
    },
    {
      "epoch": 12.65625,
      "grad_norm": 0.104576975107193,
      "learning_rate": 0.00012847222222222223,
      "loss": 0.0465,
      "step": 810
    },
    {
      "epoch": 12.8125,
      "grad_norm": 0.6552168726921082,
      "learning_rate": 0.00012731481481481483,
      "loss": 0.0536,
      "step": 820
    },
    {
      "epoch": 12.96875,
      "grad_norm": 0.39452189207077026,
      "learning_rate": 0.00012615740740740743,
      "loss": 0.0514,
      "step": 830
    },
    {
      "epoch": 13.125,
      "grad_norm": 0.16756129264831543,
      "learning_rate": 0.000125,
      "loss": 0.0417,
      "step": 840
    },
    {
      "epoch": 13.28125,
      "grad_norm": 0.13866697251796722,
      "learning_rate": 0.00012384259259259258,
      "loss": 0.0419,
      "step": 850
    },
    {
      "epoch": 13.4375,
      "grad_norm": 0.9053749442100525,
      "learning_rate": 0.0001226851851851852,
      "loss": 0.0548,
      "step": 860
    },
    {
      "epoch": 13.59375,
      "grad_norm": 0.43149927258491516,
      "learning_rate": 0.00012152777777777777,
      "loss": 0.0503,
      "step": 870
    },
    {
      "epoch": 13.75,
      "grad_norm": 0.49532395601272583,
      "learning_rate": 0.00012037037037037037,
      "loss": 0.0476,
      "step": 880
    },
    {
      "epoch": 13.90625,
      "grad_norm": 0.12025842815637589,
      "learning_rate": 0.00011921296296296296,
      "loss": 0.049,
      "step": 890
    },
    {
      "epoch": 14.0625,
      "grad_norm": 0.8570975065231323,
      "learning_rate": 0.00011805555555555556,
      "loss": 0.0513,
      "step": 900
    },
    {
      "epoch": 14.0625,
      "eval_accuracy": 0.9557522123893806,
      "eval_f1": 0.8893805309734513,
      "eval_loss": 0.1092919260263443,
      "eval_precision": 0.9095022624434389,
      "eval_recall": 0.8701298701298701,
      "eval_runtime": 1.2237,
      "eval_samples_per_second": 92.344,
      "eval_steps_per_second": 12.258,
      "step": 900
    },
    {
      "epoch": 14.21875,
      "grad_norm": 0.4120664596557617,
      "learning_rate": 0.00011689814814814815,
      "loss": 0.0552,
      "step": 910
    },
    {
      "epoch": 14.375,
      "grad_norm": 0.24265483021736145,
      "learning_rate": 0.00011574074074074075,
      "loss": 0.0434,
      "step": 920
    },
    {
      "epoch": 14.53125,
      "grad_norm": 0.23618777096271515,
      "learning_rate": 0.00011458333333333333,
      "loss": 0.0482,
      "step": 930
    },
    {
      "epoch": 14.6875,
      "grad_norm": 0.1366555392742157,
      "learning_rate": 0.00011342592592592593,
      "loss": 0.045,
      "step": 940
    },
    {
      "epoch": 14.84375,
      "grad_norm": 0.1841152310371399,
      "learning_rate": 0.00011226851851851852,
      "loss": 0.0539,
      "step": 950
    },
    {
      "epoch": 15.0,
      "grad_norm": 0.6849538087844849,
      "learning_rate": 0.00011111111111111112,
      "loss": 0.0363,
      "step": 960
    },
    {
      "epoch": 15.15625,
      "grad_norm": 0.5442699790000916,
      "learning_rate": 0.0001099537037037037,
      "loss": 0.0372,
      "step": 970
    },
    {
      "epoch": 15.3125,
      "grad_norm": 0.3825988173484802,
      "learning_rate": 0.0001087962962962963,
      "loss": 0.0405,
      "step": 980
    },
    {
      "epoch": 15.46875,
      "grad_norm": 0.0459093414247036,
      "learning_rate": 0.00010763888888888889,
      "loss": 0.0386,
      "step": 990
    },
    {
      "epoch": 15.625,
      "grad_norm": 0.2602522373199463,
      "learning_rate": 0.00010648148148148149,
      "loss": 0.0378,
      "step": 1000
    },
    {
      "epoch": 15.625,
      "eval_accuracy": 0.9548672566371681,
      "eval_f1": 0.8888888888888888,
      "eval_loss": 0.11969945579767227,
      "eval_precision": 0.8947368421052632,
      "eval_recall": 0.8831168831168831,
      "eval_runtime": 0.8768,
      "eval_samples_per_second": 128.871,
      "eval_steps_per_second": 17.107,
      "step": 1000
    },
    {
      "epoch": 15.78125,
      "grad_norm": 0.07926033437252045,
      "learning_rate": 0.00010532407407407407,
      "loss": 0.0419,
      "step": 1010
    },
    {
      "epoch": 15.9375,
      "grad_norm": 0.2084084302186966,
      "learning_rate": 0.00010416666666666667,
      "loss": 0.0336,
      "step": 1020
    },
    {
      "epoch": 16.09375,
      "grad_norm": 0.11587415635585785,
      "learning_rate": 0.00010300925925925926,
      "loss": 0.0293,
      "step": 1030
    },
    {
      "epoch": 16.25,
      "grad_norm": 0.4128260314464569,
      "learning_rate": 0.00010185185185185186,
      "loss": 0.0346,
      "step": 1040
    },
    {
      "epoch": 16.40625,
      "grad_norm": 0.2051563411951065,
      "learning_rate": 0.00010069444444444445,
      "loss": 0.0404,
      "step": 1050
    },
    {
      "epoch": 16.5625,
      "grad_norm": 1.0257600545883179,
      "learning_rate": 9.953703703703704e-05,
      "loss": 0.0521,
      "step": 1060
    },
    {
      "epoch": 16.71875,
      "grad_norm": 0.13610199093818665,
      "learning_rate": 9.837962962962963e-05,
      "loss": 0.0513,
      "step": 1070
    },
    {
      "epoch": 16.875,
      "grad_norm": 0.5424107909202576,
      "learning_rate": 9.722222222222223e-05,
      "loss": 0.0662,
      "step": 1080
    },
    {
      "epoch": 17.03125,
      "grad_norm": 0.1417212188243866,
      "learning_rate": 9.606481481481482e-05,
      "loss": 0.0364,
      "step": 1090
    },
    {
      "epoch": 17.1875,
      "grad_norm": 0.15864621102809906,
      "learning_rate": 9.490740740740742e-05,
      "loss": 0.0487,
      "step": 1100
    },
    {
      "epoch": 17.1875,
      "eval_accuracy": 0.9646017699115044,
      "eval_f1": 0.9137931034482759,
      "eval_loss": 0.09552007168531418,
      "eval_precision": 0.9098712446351931,
      "eval_recall": 0.9177489177489178,
      "eval_runtime": 0.8603,
      "eval_samples_per_second": 131.352,
      "eval_steps_per_second": 17.436,
      "step": 1100
    },
    {
      "epoch": 17.34375,
      "grad_norm": 0.5382766127586365,
      "learning_rate": 9.375e-05,
      "loss": 0.0315,
      "step": 1110
    },
    {
      "epoch": 17.5,
      "grad_norm": 0.16078180074691772,
      "learning_rate": 9.25925925925926e-05,
      "loss": 0.0245,
      "step": 1120
    },
    {
      "epoch": 17.65625,
      "grad_norm": 0.05187100172042847,
      "learning_rate": 9.143518518518519e-05,
      "loss": 0.0342,
      "step": 1130
    },
    {
      "epoch": 17.8125,
      "grad_norm": 0.4016551077365875,
      "learning_rate": 9.027777777777779e-05,
      "loss": 0.0285,
      "step": 1140
    },
    {
      "epoch": 17.96875,
      "grad_norm": 0.22832362353801727,
      "learning_rate": 8.912037037037037e-05,
      "loss": 0.0379,
      "step": 1150
    },
    {
      "epoch": 18.125,
      "grad_norm": 1.3720444440841675,
      "learning_rate": 8.796296296296297e-05,
      "loss": 0.0369,
      "step": 1160
    },
    {
      "epoch": 18.28125,
      "grad_norm": 0.23146755993366241,
      "learning_rate": 8.680555555555556e-05,
      "loss": 0.0343,
      "step": 1170
    },
    {
      "epoch": 18.4375,
      "grad_norm": 0.2672041356563568,
      "learning_rate": 8.564814814814816e-05,
      "loss": 0.0353,
      "step": 1180
    },
    {
      "epoch": 18.59375,
      "grad_norm": 0.17212288081645966,
      "learning_rate": 8.449074074074074e-05,
      "loss": 0.0315,
      "step": 1190
    },
    {
      "epoch": 18.75,
      "grad_norm": 0.12989170849323273,
      "learning_rate": 8.333333333333334e-05,
      "loss": 0.0272,
      "step": 1200
    },
    {
      "epoch": 18.75,
      "eval_accuracy": 0.9566371681415929,
      "eval_f1": 0.8927789934354485,
      "eval_loss": 0.10875095427036285,
      "eval_precision": 0.9026548672566371,
      "eval_recall": 0.8831168831168831,
      "eval_runtime": 1.2152,
      "eval_samples_per_second": 92.985,
      "eval_steps_per_second": 12.343,
      "step": 1200
    },
    {
      "epoch": 18.90625,
      "grad_norm": 0.15251386165618896,
      "learning_rate": 8.217592592592593e-05,
      "loss": 0.0353,
      "step": 1210
    },
    {
      "epoch": 19.0625,
      "grad_norm": 0.1500956416130066,
      "learning_rate": 8.101851851851853e-05,
      "loss": 0.0258,
      "step": 1220
    },
    {
      "epoch": 19.21875,
      "grad_norm": 0.16236737370491028,
      "learning_rate": 7.986111111111112e-05,
      "loss": 0.0318,
      "step": 1230
    },
    {
      "epoch": 19.375,
      "grad_norm": 0.5188699960708618,
      "learning_rate": 7.870370370370372e-05,
      "loss": 0.0388,
      "step": 1240
    },
    {
      "epoch": 19.53125,
      "grad_norm": 0.14171747863292694,
      "learning_rate": 7.75462962962963e-05,
      "loss": 0.0371,
      "step": 1250
    },
    {
      "epoch": 19.6875,
      "grad_norm": 0.355496883392334,
      "learning_rate": 7.638888888888889e-05,
      "loss": 0.0278,
      "step": 1260
    },
    {
      "epoch": 19.84375,
      "grad_norm": 0.30447283387184143,
      "learning_rate": 7.523148148148149e-05,
      "loss": 0.0307,
      "step": 1270
    },
    {
      "epoch": 20.0,
      "grad_norm": 1.8779629468917847,
      "learning_rate": 7.407407407407407e-05,
      "loss": 0.0378,
      "step": 1280
    },
    {
      "epoch": 20.15625,
      "grad_norm": 0.24115116894245148,
      "learning_rate": 7.291666666666667e-05,
      "loss": 0.0234,
      "step": 1290
    },
    {
      "epoch": 20.3125,
      "grad_norm": 0.07465353608131409,
      "learning_rate": 7.175925925925926e-05,
      "loss": 0.0241,
      "step": 1300
    },
    {
      "epoch": 20.3125,
      "eval_accuracy": 0.963716814159292,
      "eval_f1": 0.9114470842332613,
      "eval_loss": 0.09792255610227585,
      "eval_precision": 0.9094827586206896,
      "eval_recall": 0.9134199134199135,
      "eval_runtime": 0.8811,
      "eval_samples_per_second": 128.253,
      "eval_steps_per_second": 17.025,
      "step": 1300
    },
    {
      "epoch": 20.46875,
      "grad_norm": 0.44642359018325806,
      "learning_rate": 7.060185185185186e-05,
      "loss": 0.0271,
      "step": 1310
    },
    {
      "epoch": 20.625,
      "grad_norm": 0.16677480936050415,
      "learning_rate": 6.944444444444444e-05,
      "loss": 0.0199,
      "step": 1320
    },
    {
      "epoch": 20.78125,
      "grad_norm": 0.05306961014866829,
      "learning_rate": 6.828703703703704e-05,
      "loss": 0.028,
      "step": 1330
    },
    {
      "epoch": 20.9375,
      "grad_norm": 0.7962948679924011,
      "learning_rate": 6.712962962962963e-05,
      "loss": 0.0238,
      "step": 1340
    },
    {
      "epoch": 21.09375,
      "grad_norm": 0.19253899157047272,
      "learning_rate": 6.597222222222223e-05,
      "loss": 0.0367,
      "step": 1350
    },
    {
      "epoch": 21.25,
      "grad_norm": 0.22666649520397186,
      "learning_rate": 6.481481481481482e-05,
      "loss": 0.021,
      "step": 1360
    },
    {
      "epoch": 21.40625,
      "grad_norm": 0.09341959655284882,
      "learning_rate": 6.365740740740742e-05,
      "loss": 0.0217,
      "step": 1370
    },
    {
      "epoch": 21.5625,
      "grad_norm": 0.40562504529953003,
      "learning_rate": 6.25e-05,
      "loss": 0.0268,
      "step": 1380
    },
    {
      "epoch": 21.71875,
      "grad_norm": 0.20743058621883392,
      "learning_rate": 6.13425925925926e-05,
      "loss": 0.0394,
      "step": 1390
    },
    {
      "epoch": 21.875,
      "grad_norm": 0.16062897443771362,
      "learning_rate": 6.018518518518519e-05,
      "loss": 0.0311,
      "step": 1400
    },
    {
      "epoch": 21.875,
      "eval_accuracy": 0.9654867256637168,
      "eval_f1": 0.9157667386609072,
      "eval_loss": 0.11342811584472656,
      "eval_precision": 0.9137931034482759,
      "eval_recall": 0.9177489177489178,
      "eval_runtime": 0.8884,
      "eval_samples_per_second": 127.2,
      "eval_steps_per_second": 16.885,
      "step": 1400
    },
    {
      "epoch": 22.03125,
      "grad_norm": 0.08394443988800049,
      "learning_rate": 5.902777777777778e-05,
      "loss": 0.0312,
      "step": 1410
    },
    {
      "epoch": 22.1875,
      "grad_norm": 0.6736553311347961,
      "learning_rate": 5.787037037037037e-05,
      "loss": 0.0383,
      "step": 1420
    },
    {
      "epoch": 22.34375,
      "grad_norm": 0.563914954662323,
      "learning_rate": 5.6712962962962965e-05,
      "loss": 0.0287,
      "step": 1430
    },
    {
      "epoch": 22.5,
      "grad_norm": 0.08304356783628464,
      "learning_rate": 5.555555555555556e-05,
      "loss": 0.026,
      "step": 1440
    },
    {
      "epoch": 22.65625,
      "grad_norm": 0.6314889788627625,
      "learning_rate": 5.439814814814815e-05,
      "loss": 0.0337,
      "step": 1450
    },
    {
      "epoch": 22.8125,
      "grad_norm": 0.1526585817337036,
      "learning_rate": 5.3240740740740744e-05,
      "loss": 0.0386,
      "step": 1460
    },
    {
      "epoch": 22.96875,
      "grad_norm": 0.4352094829082489,
      "learning_rate": 5.208333333333334e-05,
      "loss": 0.0225,
      "step": 1470
    },
    {
      "epoch": 23.125,
      "grad_norm": 0.07802680879831314,
      "learning_rate": 5.092592592592593e-05,
      "loss": 0.028,
      "step": 1480
    },
    {
      "epoch": 23.28125,
      "grad_norm": 0.06631523370742798,
      "learning_rate": 4.976851851851852e-05,
      "loss": 0.0216,
      "step": 1490
    },
    {
      "epoch": 23.4375,
      "grad_norm": 0.4568875730037689,
      "learning_rate": 4.8611111111111115e-05,
      "loss": 0.0303,
      "step": 1500
    },
    {
      "epoch": 23.4375,
      "eval_accuracy": 0.9628318584070796,
      "eval_f1": 0.9078947368421053,
      "eval_loss": 0.10922601819038391,
      "eval_precision": 0.92,
      "eval_recall": 0.8961038961038961,
      "eval_runtime": 1.1366,
      "eval_samples_per_second": 99.417,
      "eval_steps_per_second": 13.197,
      "step": 1500
    },
    {
      "epoch": 23.59375,
      "grad_norm": 0.16732257604599,
      "learning_rate": 4.745370370370371e-05,
      "loss": 0.0182,
      "step": 1510
    },
    {
      "epoch": 23.75,
      "grad_norm": 0.8489612340927124,
      "learning_rate": 4.62962962962963e-05,
      "loss": 0.0419,
      "step": 1520
    },
    {
      "epoch": 23.90625,
      "grad_norm": 0.23256537318229675,
      "learning_rate": 4.5138888888888894e-05,
      "loss": 0.0344,
      "step": 1530
    },
    {
      "epoch": 24.0625,
      "grad_norm": 0.23274816572666168,
      "learning_rate": 4.3981481481481486e-05,
      "loss": 0.0283,
      "step": 1540
    },
    {
      "epoch": 24.21875,
      "grad_norm": 0.1935439109802246,
      "learning_rate": 4.282407407407408e-05,
      "loss": 0.0293,
      "step": 1550
    },
    {
      "epoch": 24.375,
      "grad_norm": 0.4433891773223877,
      "learning_rate": 4.166666666666667e-05,
      "loss": 0.0341,
      "step": 1560
    },
    {
      "epoch": 24.53125,
      "grad_norm": 0.3329981863498688,
      "learning_rate": 4.0509259259259265e-05,
      "loss": 0.022,
      "step": 1570
    },
    {
      "epoch": 24.6875,
      "grad_norm": 0.774336040019989,
      "learning_rate": 3.935185185185186e-05,
      "loss": 0.0304,
      "step": 1580
    },
    {
      "epoch": 24.84375,
      "grad_norm": 0.45676717162132263,
      "learning_rate": 3.8194444444444444e-05,
      "loss": 0.0213,
      "step": 1590
    },
    {
      "epoch": 25.0,
      "grad_norm": 0.18224991858005524,
      "learning_rate": 3.7037037037037037e-05,
      "loss": 0.0225,
      "step": 1600
    },
    {
      "epoch": 25.0,
      "eval_accuracy": 0.9628318584070796,
      "eval_f1": 0.908296943231441,
      "eval_loss": 0.11213955283164978,
      "eval_precision": 0.9162995594713657,
      "eval_recall": 0.9004329004329005,
      "eval_runtime": 1.05,
      "eval_samples_per_second": 107.622,
      "eval_steps_per_second": 14.286,
      "step": 1600
    },
    {
      "epoch": 25.15625,
      "grad_norm": 0.45673811435699463,
      "learning_rate": 3.587962962962963e-05,
      "loss": 0.0253,
      "step": 1610
    },
    {
      "epoch": 25.3125,
      "grad_norm": 0.09005212038755417,
      "learning_rate": 3.472222222222222e-05,
      "loss": 0.0127,
      "step": 1620
    },
    {
      "epoch": 25.46875,
      "grad_norm": 0.20682398974895477,
      "learning_rate": 3.3564814814814815e-05,
      "loss": 0.0231,
      "step": 1630
    },
    {
      "epoch": 25.625,
      "grad_norm": 0.7664525508880615,
      "learning_rate": 3.240740740740741e-05,
      "loss": 0.0174,
      "step": 1640
    },
    {
      "epoch": 25.78125,
      "grad_norm": 0.20978455245494843,
      "learning_rate": 3.125e-05,
      "loss": 0.0203,
      "step": 1650
    },
    {
      "epoch": 25.9375,
      "grad_norm": 0.5540274977684021,
      "learning_rate": 3.0092592592592593e-05,
      "loss": 0.0205,
      "step": 1660
    },
    {
      "epoch": 26.09375,
      "grad_norm": 0.1240416169166565,
      "learning_rate": 2.8935185185185186e-05,
      "loss": 0.0168,
      "step": 1670
    },
    {
      "epoch": 26.25,
      "grad_norm": 0.04385749623179436,
      "learning_rate": 2.777777777777778e-05,
      "loss": 0.0162,
      "step": 1680
    },
    {
      "epoch": 26.40625,
      "grad_norm": 0.1799972951412201,
      "learning_rate": 2.6620370370370372e-05,
      "loss": 0.0196,
      "step": 1690
    },
    {
      "epoch": 26.5625,
      "grad_norm": 0.24593585729599,
      "learning_rate": 2.5462962962962965e-05,
      "loss": 0.0292,
      "step": 1700
    },
    {
      "epoch": 26.5625,
      "eval_accuracy": 0.9619469026548673,
      "eval_f1": 0.9071274298056156,
      "eval_loss": 0.11493521183729172,
      "eval_precision": 0.9051724137931034,
      "eval_recall": 0.9090909090909091,
      "eval_runtime": 0.8719,
      "eval_samples_per_second": 129.607,
      "eval_steps_per_second": 17.204,
      "step": 1700
    },
    {
      "epoch": 26.71875,
      "grad_norm": 0.2344673126935959,
      "learning_rate": 2.4305555555555558e-05,
      "loss": 0.0177,
      "step": 1710
    },
    {
      "epoch": 26.875,
      "grad_norm": 0.263621062040329,
      "learning_rate": 2.314814814814815e-05,
      "loss": 0.0288,
      "step": 1720
    },
    {
      "epoch": 27.03125,
      "grad_norm": 0.27248746156692505,
      "learning_rate": 2.1990740740740743e-05,
      "loss": 0.0158,
      "step": 1730
    },
    {
      "epoch": 27.1875,
      "grad_norm": 0.35065901279449463,
      "learning_rate": 2.0833333333333336e-05,
      "loss": 0.0198,
      "step": 1740
    },
    {
      "epoch": 27.34375,
      "grad_norm": 0.23319651186466217,
      "learning_rate": 1.967592592592593e-05,
      "loss": 0.0208,
      "step": 1750
    },
    {
      "epoch": 27.5,
      "grad_norm": 0.15196481347084045,
      "learning_rate": 1.8518518518518518e-05,
      "loss": 0.0161,
      "step": 1760
    },
    {
      "epoch": 27.65625,
      "grad_norm": 0.17273353040218353,
      "learning_rate": 1.736111111111111e-05,
      "loss": 0.0213,
      "step": 1770
    },
    {
      "epoch": 27.8125,
      "grad_norm": 0.31511059403419495,
      "learning_rate": 1.6203703703703704e-05,
      "loss": 0.012,
      "step": 1780
    },
    {
      "epoch": 27.96875,
      "grad_norm": 0.09265203773975372,
      "learning_rate": 1.5046296296296297e-05,
      "loss": 0.0218,
      "step": 1790
    },
    {
      "epoch": 28.125,
      "grad_norm": 0.2638147473335266,
      "learning_rate": 1.388888888888889e-05,
      "loss": 0.0261,
      "step": 1800
    },
    {
      "epoch": 28.125,
      "eval_accuracy": 0.9619469026548673,
      "eval_f1": 0.9079229122055675,
      "eval_loss": 0.11067904531955719,
      "eval_precision": 0.8983050847457628,
      "eval_recall": 0.9177489177489178,
      "eval_runtime": 0.9722,
      "eval_samples_per_second": 116.235,
      "eval_steps_per_second": 15.429,
      "step": 1800
    },
    {
      "epoch": 28.28125,
      "grad_norm": 0.2734526991844177,
      "learning_rate": 1.2731481481481482e-05,
      "loss": 0.0175,
      "step": 1810
    },
    {
      "epoch": 28.4375,
      "grad_norm": 0.06026133522391319,
      "learning_rate": 1.1574074074074075e-05,
      "loss": 0.0168,
      "step": 1820
    },
    {
      "epoch": 28.59375,
      "grad_norm": 0.02611556649208069,
      "learning_rate": 1.0416666666666668e-05,
      "loss": 0.0177,
      "step": 1830
    },
    {
      "epoch": 28.75,
      "grad_norm": 0.23434928059577942,
      "learning_rate": 9.259259259259259e-06,
      "loss": 0.0252,
      "step": 1840
    },
    {
      "epoch": 28.90625,
      "grad_norm": 0.07623735815286636,
      "learning_rate": 8.101851851851852e-06,
      "loss": 0.0235,
      "step": 1850
    },
    {
      "epoch": 29.0625,
      "grad_norm": 0.05061192065477371,
      "learning_rate": 6.944444444444445e-06,
      "loss": 0.0185,
      "step": 1860
    },
    {
      "epoch": 29.21875,
      "grad_norm": 0.03355779871344566,
      "learning_rate": 5.787037037037038e-06,
      "loss": 0.0157,
      "step": 1870
    },
    {
      "epoch": 29.375,
      "grad_norm": 0.023396974429488182,
      "learning_rate": 4.6296296296296296e-06,
      "loss": 0.016,
      "step": 1880
    },
    {
      "epoch": 29.53125,
      "grad_norm": 0.31754446029663086,
      "learning_rate": 3.4722222222222224e-06,
      "loss": 0.0213,
      "step": 1890
    },
    {
      "epoch": 29.6875,
      "grad_norm": 0.04646694287657738,
      "learning_rate": 2.3148148148148148e-06,
      "loss": 0.0166,
      "step": 1900
    },
    {
      "epoch": 29.6875,
      "eval_accuracy": 0.9610619469026549,
      "eval_f1": 0.9051724137931034,
      "eval_loss": 0.11101004481315613,
      "eval_precision": 0.9012875536480687,
      "eval_recall": 0.9090909090909091,
      "eval_runtime": 0.8549,
      "eval_samples_per_second": 132.182,
      "eval_steps_per_second": 17.546,
      "step": 1900
    },
    {
      "epoch": 29.84375,
      "grad_norm": 0.05433151125907898,
      "learning_rate": 1.1574074074074074e-06,
      "loss": 0.0174,
      "step": 1910
    },
    {
      "epoch": 30.0,
      "grad_norm": 0.3085578382015228,
      "learning_rate": 0.0,
      "loss": 0.0208,
      "step": 1920
    },
    {
      "epoch": 30.0,
      "step": 1920,
      "total_flos": 2.352825493649326e+18,
      "train_loss": 0.05003170374160012,
      "train_runtime": 517.118,
      "train_samples_per_second": 58.71,
      "train_steps_per_second": 3.713
    }
  ],
  "logging_steps": 10,
  "max_steps": 1920,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 30,
  "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": 2.352825493649326e+18,
  "train_batch_size": 16,
  "trial_name": null,
  "trial_params": null
}