{
  "best_metric": 0.8387096774193549,
  "best_model_checkpoint": "videomae-base-finetuned-ucf101-subset-finetuned-subset\\checkpoint-2655",
  "epoch": 59.00826210826211,
  "eval_steps": 500,
  "global_step": 3510,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.0,
      "learning_rate": 1.4245014245014246e-06,
      "loss": 1.5791,
      "step": 10
    },
    {
      "epoch": 0.01,
      "learning_rate": 2.8490028490028492e-06,
      "loss": 1.6248,
      "step": 20
    },
    {
      "epoch": 0.01,
      "learning_rate": 4.273504273504274e-06,
      "loss": 1.6509,
      "step": 30
    },
    {
      "epoch": 0.01,
      "learning_rate": 5.6980056980056985e-06,
      "loss": 1.6033,
      "step": 40
    },
    {
      "epoch": 0.01,
      "learning_rate": 7.122507122507123e-06,
      "loss": 1.58,
      "step": 50
    },
    {
      "epoch": 0.02,
      "eval_accuracy": 0.37327188940092165,
      "eval_loss": 1.5475504398345947,
      "eval_runtime": 280.9942,
      "eval_samples_per_second": 0.772,
      "eval_steps_per_second": 0.1,
      "step": 59
    },
    {
      "epoch": 1.0,
      "learning_rate": 8.547008547008548e-06,
      "loss": 1.5857,
      "step": 60
    },
    {
      "epoch": 1.0,
      "learning_rate": 9.971509971509972e-06,
      "loss": 1.5372,
      "step": 70
    },
    {
      "epoch": 1.01,
      "learning_rate": 1.1396011396011397e-05,
      "loss": 1.4931,
      "step": 80
    },
    {
      "epoch": 1.01,
      "learning_rate": 1.282051282051282e-05,
      "loss": 1.6021,
      "step": 90
    },
    {
      "epoch": 1.01,
      "learning_rate": 1.4245014245014246e-05,
      "loss": 1.5149,
      "step": 100
    },
    {
      "epoch": 1.01,
      "learning_rate": 1.566951566951567e-05,
      "loss": 1.3617,
      "step": 110
    },
    {
      "epoch": 1.02,
      "eval_accuracy": 0.4976958525345622,
      "eval_loss": 1.2123247385025024,
      "eval_runtime": 282.6781,
      "eval_samples_per_second": 0.768,
      "eval_steps_per_second": 0.099,
      "step": 118
    },
    {
      "epoch": 2.0,
      "learning_rate": 1.7094017094017095e-05,
      "loss": 1.345,
      "step": 120
    },
    {
      "epoch": 2.0,
      "learning_rate": 1.8518518518518518e-05,
      "loss": 1.4019,
      "step": 130
    },
    {
      "epoch": 2.01,
      "learning_rate": 1.9943019943019945e-05,
      "loss": 1.2799,
      "step": 140
    },
    {
      "epoch": 2.01,
      "learning_rate": 2.1367521367521368e-05,
      "loss": 1.2003,
      "step": 150
    },
    {
      "epoch": 2.01,
      "learning_rate": 2.2792022792022794e-05,
      "loss": 1.1992,
      "step": 160
    },
    {
      "epoch": 2.01,
      "learning_rate": 2.4216524216524217e-05,
      "loss": 1.1668,
      "step": 170
    },
    {
      "epoch": 2.02,
      "eval_accuracy": 0.3824884792626728,
      "eval_loss": 1.3331272602081299,
      "eval_runtime": 283.1967,
      "eval_samples_per_second": 0.766,
      "eval_steps_per_second": 0.099,
      "step": 177
    },
    {
      "epoch": 3.0,
      "learning_rate": 2.564102564102564e-05,
      "loss": 1.5354,
      "step": 180
    },
    {
      "epoch": 3.0,
      "learning_rate": 2.706552706552707e-05,
      "loss": 1.2036,
      "step": 190
    },
    {
      "epoch": 3.01,
      "learning_rate": 2.8490028490028492e-05,
      "loss": 1.252,
      "step": 200
    },
    {
      "epoch": 3.01,
      "learning_rate": 2.9914529914529915e-05,
      "loss": 1.1877,
      "step": 210
    },
    {
      "epoch": 3.01,
      "learning_rate": 3.133903133903134e-05,
      "loss": 1.3834,
      "step": 220
    },
    {
      "epoch": 3.02,
      "learning_rate": 3.2763532763532764e-05,
      "loss": 1.1751,
      "step": 230
    },
    {
      "epoch": 3.02,
      "eval_accuracy": 0.3686635944700461,
      "eval_loss": 1.2529356479644775,
      "eval_runtime": 303.7124,
      "eval_samples_per_second": 0.714,
      "eval_steps_per_second": 0.092,
      "step": 236
    },
    {
      "epoch": 4.0,
      "learning_rate": 3.418803418803419e-05,
      "loss": 1.1436,
      "step": 240
    },
    {
      "epoch": 4.0,
      "learning_rate": 3.561253561253561e-05,
      "loss": 1.2635,
      "step": 250
    },
    {
      "epoch": 4.01,
      "learning_rate": 3.7037037037037037e-05,
      "loss": 1.2653,
      "step": 260
    },
    {
      "epoch": 4.01,
      "learning_rate": 3.846153846153846e-05,
      "loss": 1.2505,
      "step": 270
    },
    {
      "epoch": 4.01,
      "learning_rate": 3.988603988603989e-05,
      "loss": 1.1114,
      "step": 280
    },
    {
      "epoch": 4.02,
      "learning_rate": 4.131054131054131e-05,
      "loss": 0.9961,
      "step": 290
    },
    {
      "epoch": 4.02,
      "eval_accuracy": 0.6451612903225806,
      "eval_loss": 1.0007940530776978,
      "eval_runtime": 291.2333,
      "eval_samples_per_second": 0.745,
      "eval_steps_per_second": 0.096,
      "step": 295
    },
    {
      "epoch": 5.0,
      "learning_rate": 4.2735042735042735e-05,
      "loss": 1.1199,
      "step": 300
    },
    {
      "epoch": 5.0,
      "learning_rate": 4.415954415954416e-05,
      "loss": 1.0766,
      "step": 310
    },
    {
      "epoch": 5.01,
      "learning_rate": 4.558404558404559e-05,
      "loss": 1.1833,
      "step": 320
    },
    {
      "epoch": 5.01,
      "learning_rate": 4.700854700854701e-05,
      "loss": 1.2997,
      "step": 330
    },
    {
      "epoch": 5.01,
      "learning_rate": 4.8433048433048433e-05,
      "loss": 1.1507,
      "step": 340
    },
    {
      "epoch": 5.02,
      "learning_rate": 4.985754985754986e-05,
      "loss": 1.0562,
      "step": 350
    },
    {
      "epoch": 5.02,
      "eval_accuracy": 0.5391705069124424,
      "eval_loss": 1.0534659624099731,
      "eval_runtime": 288.9126,
      "eval_samples_per_second": 0.751,
      "eval_steps_per_second": 0.097,
      "step": 354
    },
    {
      "epoch": 6.0,
      "learning_rate": 4.985754985754986e-05,
      "loss": 1.1336,
      "step": 360
    },
    {
      "epoch": 6.0,
      "learning_rate": 4.9699271921494144e-05,
      "loss": 1.192,
      "step": 370
    },
    {
      "epoch": 6.01,
      "learning_rate": 4.9540993985438435e-05,
      "loss": 1.1785,
      "step": 380
    },
    {
      "epoch": 6.01,
      "learning_rate": 4.938271604938271e-05,
      "loss": 1.2818,
      "step": 390
    },
    {
      "epoch": 6.01,
      "learning_rate": 4.9224438113327004e-05,
      "loss": 1.179,
      "step": 400
    },
    {
      "epoch": 6.02,
      "learning_rate": 4.906616017727129e-05,
      "loss": 1.0189,
      "step": 410
    },
    {
      "epoch": 6.02,
      "eval_accuracy": 0.6866359447004609,
      "eval_loss": 0.9321132302284241,
      "eval_runtime": 300.2588,
      "eval_samples_per_second": 0.723,
      "eval_steps_per_second": 0.093,
      "step": 413
    },
    {
      "epoch": 7.0,
      "learning_rate": 4.890788224121557e-05,
      "loss": 1.178,
      "step": 420
    },
    {
      "epoch": 7.0,
      "learning_rate": 4.8749604305159865e-05,
      "loss": 1.0559,
      "step": 430
    },
    {
      "epoch": 7.01,
      "learning_rate": 4.859132636910415e-05,
      "loss": 1.3566,
      "step": 440
    },
    {
      "epoch": 7.01,
      "learning_rate": 4.8433048433048433e-05,
      "loss": 1.2278,
      "step": 450
    },
    {
      "epoch": 7.01,
      "learning_rate": 4.8274770496992725e-05,
      "loss": 1.1068,
      "step": 460
    },
    {
      "epoch": 7.02,
      "learning_rate": 4.811649256093701e-05,
      "loss": 0.8553,
      "step": 470
    },
    {
      "epoch": 7.02,
      "eval_accuracy": 0.5852534562211982,
      "eval_loss": 0.895534098148346,
      "eval_runtime": 301.0743,
      "eval_samples_per_second": 0.721,
      "eval_steps_per_second": 0.093,
      "step": 472
    },
    {
      "epoch": 8.0,
      "learning_rate": 4.7958214624881294e-05,
      "loss": 0.8541,
      "step": 480
    },
    {
      "epoch": 8.01,
      "learning_rate": 4.779993668882558e-05,
      "loss": 1.1329,
      "step": 490
    },
    {
      "epoch": 8.01,
      "learning_rate": 4.764165875276987e-05,
      "loss": 1.1688,
      "step": 500
    },
    {
      "epoch": 8.01,
      "learning_rate": 4.7483380816714154e-05,
      "loss": 0.8642,
      "step": 510
    },
    {
      "epoch": 8.01,
      "learning_rate": 4.732510288065844e-05,
      "loss": 0.9557,
      "step": 520
    },
    {
      "epoch": 8.02,
      "learning_rate": 4.716682494460272e-05,
      "loss": 0.8961,
      "step": 530
    },
    {
      "epoch": 8.02,
      "eval_accuracy": 0.7419354838709677,
      "eval_loss": 0.7889271974563599,
      "eval_runtime": 301.0199,
      "eval_samples_per_second": 0.721,
      "eval_steps_per_second": 0.093,
      "step": 531
    },
    {
      "epoch": 9.0,
      "learning_rate": 4.700854700854701e-05,
      "loss": 1.1113,
      "step": 540
    },
    {
      "epoch": 9.01,
      "learning_rate": 4.68502690724913e-05,
      "loss": 1.0851,
      "step": 550
    },
    {
      "epoch": 9.01,
      "learning_rate": 4.669199113643558e-05,
      "loss": 1.0467,
      "step": 560
    },
    {
      "epoch": 9.01,
      "learning_rate": 4.653371320037987e-05,
      "loss": 0.8215,
      "step": 570
    },
    {
      "epoch": 9.01,
      "learning_rate": 4.637543526432416e-05,
      "loss": 0.9789,
      "step": 580
    },
    {
      "epoch": 9.02,
      "learning_rate": 4.621715732826844e-05,
      "loss": 1.1626,
      "step": 590
    },
    {
      "epoch": 9.02,
      "eval_accuracy": 0.6175115207373272,
      "eval_loss": 0.95318603515625,
      "eval_runtime": 305.3217,
      "eval_samples_per_second": 0.711,
      "eval_steps_per_second": 0.092,
      "step": 590
    },
    {
      "epoch": 10.0,
      "learning_rate": 4.605887939221273e-05,
      "loss": 0.8144,
      "step": 600
    },
    {
      "epoch": 10.01,
      "learning_rate": 4.590060145615702e-05,
      "loss": 0.7724,
      "step": 610
    },
    {
      "epoch": 10.01,
      "learning_rate": 4.5742323520101296e-05,
      "loss": 1.088,
      "step": 620
    },
    {
      "epoch": 10.01,
      "learning_rate": 4.558404558404559e-05,
      "loss": 1.074,
      "step": 630
    },
    {
      "epoch": 10.01,
      "learning_rate": 4.542576764798987e-05,
      "loss": 0.7951,
      "step": 640
    },
    {
      "epoch": 10.02,
      "eval_accuracy": 0.5529953917050692,
      "eval_loss": 1.1165187358856201,
      "eval_runtime": 303.3614,
      "eval_samples_per_second": 0.715,
      "eval_steps_per_second": 0.092,
      "step": 649
    },
    {
      "epoch": 11.0,
      "learning_rate": 4.5267489711934157e-05,
      "loss": 0.8376,
      "step": 650
    },
    {
      "epoch": 11.0,
      "learning_rate": 4.510921177587845e-05,
      "loss": 0.9325,
      "step": 660
    },
    {
      "epoch": 11.01,
      "learning_rate": 4.4950933839822725e-05,
      "loss": 0.8498,
      "step": 670
    },
    {
      "epoch": 11.01,
      "learning_rate": 4.479265590376702e-05,
      "loss": 0.7743,
      "step": 680
    },
    {
      "epoch": 11.01,
      "learning_rate": 4.463437796771131e-05,
      "loss": 1.0063,
      "step": 690
    },
    {
      "epoch": 11.01,
      "learning_rate": 4.4476100031655586e-05,
      "loss": 0.9042,
      "step": 700
    },
    {
      "epoch": 11.02,
      "eval_accuracy": 0.7649769585253456,
      "eval_loss": 0.7012106776237488,
      "eval_runtime": 307.8679,
      "eval_samples_per_second": 0.705,
      "eval_steps_per_second": 0.091,
      "step": 708
    },
    {
      "epoch": 12.0,
      "learning_rate": 4.431782209559988e-05,
      "loss": 0.8243,
      "step": 710
    },
    {
      "epoch": 12.0,
      "learning_rate": 4.415954415954416e-05,
      "loss": 0.7756,
      "step": 720
    },
    {
      "epoch": 12.01,
      "learning_rate": 4.4001266223488446e-05,
      "loss": 0.9731,
      "step": 730
    },
    {
      "epoch": 12.01,
      "learning_rate": 4.384298828743274e-05,
      "loss": 0.8979,
      "step": 740
    },
    {
      "epoch": 12.01,
      "learning_rate": 4.368471035137702e-05,
      "loss": 0.9586,
      "step": 750
    },
    {
      "epoch": 12.01,
      "learning_rate": 4.3526432415321306e-05,
      "loss": 0.8642,
      "step": 760
    },
    {
      "epoch": 12.02,
      "eval_accuracy": 0.6774193548387096,
      "eval_loss": 0.7588521838188171,
      "eval_runtime": 303.2535,
      "eval_samples_per_second": 0.716,
      "eval_steps_per_second": 0.092,
      "step": 767
    },
    {
      "epoch": 13.0,
      "learning_rate": 4.336815447926559e-05,
      "loss": 0.7979,
      "step": 770
    },
    {
      "epoch": 13.0,
      "learning_rate": 4.3209876543209875e-05,
      "loss": 0.8616,
      "step": 780
    },
    {
      "epoch": 13.01,
      "learning_rate": 4.3051598607154166e-05,
      "loss": 0.6542,
      "step": 790
    },
    {
      "epoch": 13.01,
      "learning_rate": 4.289332067109845e-05,
      "loss": 0.732,
      "step": 800
    },
    {
      "epoch": 13.01,
      "learning_rate": 4.2735042735042735e-05,
      "loss": 0.806,
      "step": 810
    },
    {
      "epoch": 13.02,
      "learning_rate": 4.2576764798987026e-05,
      "loss": 0.8017,
      "step": 820
    },
    {
      "epoch": 13.02,
      "eval_accuracy": 0.695852534562212,
      "eval_loss": 0.7485441565513611,
      "eval_runtime": 294.6482,
      "eval_samples_per_second": 0.736,
      "eval_steps_per_second": 0.095,
      "step": 826
    },
    {
      "epoch": 14.0,
      "learning_rate": 4.241848686293131e-05,
      "loss": 0.7243,
      "step": 830
    },
    {
      "epoch": 14.0,
      "learning_rate": 4.2260208926875595e-05,
      "loss": 0.7128,
      "step": 840
    },
    {
      "epoch": 14.01,
      "learning_rate": 4.210193099081988e-05,
      "loss": 0.876,
      "step": 850
    },
    {
      "epoch": 14.01,
      "learning_rate": 4.194365305476417e-05,
      "loss": 0.7118,
      "step": 860
    },
    {
      "epoch": 14.01,
      "learning_rate": 4.1785375118708455e-05,
      "loss": 0.7758,
      "step": 870
    },
    {
      "epoch": 14.02,
      "learning_rate": 4.162709718265274e-05,
      "loss": 0.7523,
      "step": 880
    },
    {
      "epoch": 14.02,
      "eval_accuracy": 0.783410138248848,
      "eval_loss": 0.5616857409477234,
      "eval_runtime": 286.0674,
      "eval_samples_per_second": 0.759,
      "eval_steps_per_second": 0.098,
      "step": 885
    },
    {
      "epoch": 15.0,
      "learning_rate": 4.1468819246597024e-05,
      "loss": 0.5866,
      "step": 890
    },
    {
      "epoch": 15.0,
      "learning_rate": 4.131054131054131e-05,
      "loss": 0.7429,
      "step": 900
    },
    {
      "epoch": 15.01,
      "learning_rate": 4.11522633744856e-05,
      "loss": 0.8891,
      "step": 910
    },
    {
      "epoch": 15.01,
      "learning_rate": 4.0993985438429884e-05,
      "loss": 0.8778,
      "step": 920
    },
    {
      "epoch": 15.01,
      "learning_rate": 4.083570750237417e-05,
      "loss": 0.673,
      "step": 930
    },
    {
      "epoch": 15.02,
      "learning_rate": 4.067742956631846e-05,
      "loss": 0.7223,
      "step": 940
    },
    {
      "epoch": 15.02,
      "eval_accuracy": 0.6129032258064516,
      "eval_loss": 1.0343618392944336,
      "eval_runtime": 282.8856,
      "eval_samples_per_second": 0.767,
      "eval_steps_per_second": 0.099,
      "step": 944
    },
    {
      "epoch": 16.0,
      "learning_rate": 4.0519151630262745e-05,
      "loss": 0.8074,
      "step": 950
    },
    {
      "epoch": 16.0,
      "learning_rate": 4.036087369420703e-05,
      "loss": 0.6108,
      "step": 960
    },
    {
      "epoch": 16.01,
      "learning_rate": 4.020259575815132e-05,
      "loss": 0.9762,
      "step": 970
    },
    {
      "epoch": 16.01,
      "learning_rate": 4.00443178220956e-05,
      "loss": 0.7148,
      "step": 980
    },
    {
      "epoch": 16.01,
      "learning_rate": 3.988603988603989e-05,
      "loss": 0.8286,
      "step": 990
    },
    {
      "epoch": 16.02,
      "learning_rate": 3.9727761949984174e-05,
      "loss": 0.6164,
      "step": 1000
    },
    {
      "epoch": 16.02,
      "eval_accuracy": 0.7419354838709677,
      "eval_loss": 0.727464497089386,
      "eval_runtime": 330.7312,
      "eval_samples_per_second": 0.656,
      "eval_steps_per_second": 0.085,
      "step": 1003
    },
    {
      "epoch": 17.0,
      "learning_rate": 3.956948401392846e-05,
      "loss": 0.7374,
      "step": 1010
    },
    {
      "epoch": 17.0,
      "learning_rate": 3.941120607787275e-05,
      "loss": 0.8524,
      "step": 1020
    },
    {
      "epoch": 17.01,
      "learning_rate": 3.925292814181703e-05,
      "loss": 0.8531,
      "step": 1030
    },
    {
      "epoch": 17.01,
      "learning_rate": 3.909465020576132e-05,
      "loss": 0.6948,
      "step": 1040
    },
    {
      "epoch": 17.01,
      "learning_rate": 3.893637226970561e-05,
      "loss": 0.6618,
      "step": 1050
    },
    {
      "epoch": 17.02,
      "learning_rate": 3.877809433364989e-05,
      "loss": 0.6892,
      "step": 1060
    },
    {
      "epoch": 17.02,
      "eval_accuracy": 0.7788018433179723,
      "eval_loss": 0.616241991519928,
      "eval_runtime": 331.0915,
      "eval_samples_per_second": 0.655,
      "eval_steps_per_second": 0.085,
      "step": 1062
    },
    {
      "epoch": 18.0,
      "learning_rate": 3.861981639759418e-05,
      "loss": 0.7827,
      "step": 1070
    },
    {
      "epoch": 18.01,
      "learning_rate": 3.846153846153846e-05,
      "loss": 0.6962,
      "step": 1080
    },
    {
      "epoch": 18.01,
      "learning_rate": 3.830326052548275e-05,
      "loss": 0.7623,
      "step": 1090
    },
    {
      "epoch": 18.01,
      "learning_rate": 3.814498258942704e-05,
      "loss": 0.6629,
      "step": 1100
    },
    {
      "epoch": 18.01,
      "learning_rate": 3.798670465337132e-05,
      "loss": 0.8101,
      "step": 1110
    },
    {
      "epoch": 18.02,
      "learning_rate": 3.782842671731561e-05,
      "loss": 0.7865,
      "step": 1120
    },
    {
      "epoch": 18.02,
      "eval_accuracy": 0.6589861751152074,
      "eval_loss": 0.7966123819351196,
      "eval_runtime": 336.073,
      "eval_samples_per_second": 0.646,
      "eval_steps_per_second": 0.083,
      "step": 1121
    },
    {
      "epoch": 19.0,
      "learning_rate": 3.767014878125989e-05,
      "loss": 0.5888,
      "step": 1130
    },
    {
      "epoch": 19.01,
      "learning_rate": 3.7511870845204176e-05,
      "loss": 0.8017,
      "step": 1140
    },
    {
      "epoch": 19.01,
      "learning_rate": 3.735359290914847e-05,
      "loss": 0.6144,
      "step": 1150
    },
    {
      "epoch": 19.01,
      "learning_rate": 3.719531497309275e-05,
      "loss": 0.715,
      "step": 1160
    },
    {
      "epoch": 19.01,
      "learning_rate": 3.7037037037037037e-05,
      "loss": 0.6533,
      "step": 1170
    },
    {
      "epoch": 19.02,
      "learning_rate": 3.687875910098133e-05,
      "loss": 0.6387,
      "step": 1180
    },
    {
      "epoch": 19.02,
      "eval_accuracy": 0.6774193548387096,
      "eval_loss": 0.7435915470123291,
      "eval_runtime": 310.5543,
      "eval_samples_per_second": 0.699,
      "eval_steps_per_second": 0.09,
      "step": 1180
    },
    {
      "epoch": 20.0,
      "learning_rate": 3.672048116492561e-05,
      "loss": 0.6266,
      "step": 1190
    },
    {
      "epoch": 20.01,
      "learning_rate": 3.65622032288699e-05,
      "loss": 0.519,
      "step": 1200
    },
    {
      "epoch": 20.01,
      "learning_rate": 3.640392529281418e-05,
      "loss": 0.5029,
      "step": 1210
    },
    {
      "epoch": 20.01,
      "learning_rate": 3.624564735675847e-05,
      "loss": 1.0188,
      "step": 1220
    },
    {
      "epoch": 20.01,
      "learning_rate": 3.608736942070276e-05,
      "loss": 0.6181,
      "step": 1230
    },
    {
      "epoch": 20.02,
      "eval_accuracy": 0.7373271889400922,
      "eval_loss": 0.7136892676353455,
      "eval_runtime": 305.6632,
      "eval_samples_per_second": 0.71,
      "eval_steps_per_second": 0.092,
      "step": 1239
    },
    {
      "epoch": 21.0,
      "learning_rate": 3.592909148464704e-05,
      "loss": 0.8196,
      "step": 1240
    },
    {
      "epoch": 21.0,
      "learning_rate": 3.5770813548591326e-05,
      "loss": 0.7009,
      "step": 1250
    },
    {
      "epoch": 21.01,
      "learning_rate": 3.561253561253561e-05,
      "loss": 0.7675,
      "step": 1260
    },
    {
      "epoch": 21.01,
      "learning_rate": 3.54542576764799e-05,
      "loss": 0.6991,
      "step": 1270
    },
    {
      "epoch": 21.01,
      "learning_rate": 3.5295979740424186e-05,
      "loss": 0.5857,
      "step": 1280
    },
    {
      "epoch": 21.01,
      "learning_rate": 3.513770180436847e-05,
      "loss": 0.6085,
      "step": 1290
    },
    {
      "epoch": 21.02,
      "eval_accuracy": 0.6682027649769585,
      "eval_loss": 0.9580748081207275,
      "eval_runtime": 300.6045,
      "eval_samples_per_second": 0.722,
      "eval_steps_per_second": 0.093,
      "step": 1298
    },
    {
      "epoch": 22.0,
      "learning_rate": 3.497942386831276e-05,
      "loss": 0.738,
      "step": 1300
    },
    {
      "epoch": 22.0,
      "learning_rate": 3.4821145932257046e-05,
      "loss": 0.6597,
      "step": 1310
    },
    {
      "epoch": 22.01,
      "learning_rate": 3.466286799620133e-05,
      "loss": 0.7283,
      "step": 1320
    },
    {
      "epoch": 22.01,
      "learning_rate": 3.450459006014562e-05,
      "loss": 0.8223,
      "step": 1330
    },
    {
      "epoch": 22.01,
      "learning_rate": 3.43463121240899e-05,
      "loss": 0.5629,
      "step": 1340
    },
    {
      "epoch": 22.01,
      "learning_rate": 3.418803418803419e-05,
      "loss": 0.7109,
      "step": 1350
    },
    {
      "epoch": 22.02,
      "eval_accuracy": 0.7096774193548387,
      "eval_loss": 0.7746023535728455,
      "eval_runtime": 300.4789,
      "eval_samples_per_second": 0.722,
      "eval_steps_per_second": 0.093,
      "step": 1357
    },
    {
      "epoch": 23.0,
      "learning_rate": 3.4029756251978475e-05,
      "loss": 0.8563,
      "step": 1360
    },
    {
      "epoch": 23.0,
      "learning_rate": 3.387147831592276e-05,
      "loss": 0.6669,
      "step": 1370
    },
    {
      "epoch": 23.01,
      "learning_rate": 3.371320037986705e-05,
      "loss": 0.786,
      "step": 1380
    },
    {
      "epoch": 23.01,
      "learning_rate": 3.355492244381133e-05,
      "loss": 0.5724,
      "step": 1390
    },
    {
      "epoch": 23.01,
      "learning_rate": 3.339664450775562e-05,
      "loss": 0.928,
      "step": 1400
    },
    {
      "epoch": 23.02,
      "learning_rate": 3.323836657169991e-05,
      "loss": 0.7686,
      "step": 1410
    },
    {
      "epoch": 23.02,
      "eval_accuracy": 0.7096774193548387,
      "eval_loss": 0.7969343066215515,
      "eval_runtime": 303.3779,
      "eval_samples_per_second": 0.715,
      "eval_steps_per_second": 0.092,
      "step": 1416
    },
    {
      "epoch": 24.0,
      "learning_rate": 3.308008863564419e-05,
      "loss": 0.5478,
      "step": 1420
    },
    {
      "epoch": 24.0,
      "learning_rate": 3.292181069958848e-05,
      "loss": 0.7107,
      "step": 1430
    },
    {
      "epoch": 24.01,
      "learning_rate": 3.2763532763532764e-05,
      "loss": 0.5742,
      "step": 1440
    },
    {
      "epoch": 24.01,
      "learning_rate": 3.260525482747705e-05,
      "loss": 0.5811,
      "step": 1450
    },
    {
      "epoch": 24.01,
      "learning_rate": 3.244697689142134e-05,
      "loss": 0.676,
      "step": 1460
    },
    {
      "epoch": 24.02,
      "learning_rate": 3.2288698955365625e-05,
      "loss": 0.5995,
      "step": 1470
    },
    {
      "epoch": 24.02,
      "eval_accuracy": 0.6129032258064516,
      "eval_loss": 1.0075043439865112,
      "eval_runtime": 295.2487,
      "eval_samples_per_second": 0.735,
      "eval_steps_per_second": 0.095,
      "step": 1475
    },
    {
      "epoch": 25.0,
      "learning_rate": 3.213042101930991e-05,
      "loss": 0.6382,
      "step": 1480
    },
    {
      "epoch": 25.0,
      "learning_rate": 3.1972143083254193e-05,
      "loss": 0.5384,
      "step": 1490
    },
    {
      "epoch": 25.01,
      "learning_rate": 3.181386514719848e-05,
      "loss": 0.7876,
      "step": 1500
    },
    {
      "epoch": 25.01,
      "learning_rate": 3.165558721114277e-05,
      "loss": 0.6693,
      "step": 1510
    },
    {
      "epoch": 25.01,
      "learning_rate": 3.1497309275087054e-05,
      "loss": 0.5688,
      "step": 1520
    },
    {
      "epoch": 25.02,
      "learning_rate": 3.133903133903134e-05,
      "loss": 0.5854,
      "step": 1530
    },
    {
      "epoch": 25.02,
      "eval_accuracy": 0.7419354838709677,
      "eval_loss": 0.7388768196105957,
      "eval_runtime": 335.3249,
      "eval_samples_per_second": 0.647,
      "eval_steps_per_second": 0.084,
      "step": 1534
    },
    {
      "epoch": 26.0,
      "learning_rate": 3.118075340297563e-05,
      "loss": 0.4072,
      "step": 1540
    },
    {
      "epoch": 26.0,
      "learning_rate": 3.1022475466919914e-05,
      "loss": 0.6526,
      "step": 1550
    },
    {
      "epoch": 26.01,
      "learning_rate": 3.08641975308642e-05,
      "loss": 0.6973,
      "step": 1560
    },
    {
      "epoch": 26.01,
      "learning_rate": 3.070591959480848e-05,
      "loss": 0.8606,
      "step": 1570
    },
    {
      "epoch": 26.01,
      "learning_rate": 3.0547641658752774e-05,
      "loss": 0.6417,
      "step": 1580
    },
    {
      "epoch": 26.02,
      "learning_rate": 3.0389363722697055e-05,
      "loss": 0.575,
      "step": 1590
    },
    {
      "epoch": 26.02,
      "eval_accuracy": 0.7142857142857143,
      "eval_loss": 0.7198013067245483,
      "eval_runtime": 323.3382,
      "eval_samples_per_second": 0.671,
      "eval_steps_per_second": 0.087,
      "step": 1593
    },
    {
      "epoch": 27.0,
      "learning_rate": 3.0231085786641343e-05,
      "loss": 0.5309,
      "step": 1600
    },
    {
      "epoch": 27.0,
      "learning_rate": 3.0072807850585634e-05,
      "loss": 0.6287,
      "step": 1610
    },
    {
      "epoch": 27.01,
      "learning_rate": 2.9914529914529915e-05,
      "loss": 0.7234,
      "step": 1620
    },
    {
      "epoch": 27.01,
      "learning_rate": 2.9756251978474203e-05,
      "loss": 0.6081,
      "step": 1630
    },
    {
      "epoch": 27.01,
      "learning_rate": 2.9597974042418487e-05,
      "loss": 0.7265,
      "step": 1640
    },
    {
      "epoch": 27.02,
      "learning_rate": 2.9439696106362775e-05,
      "loss": 0.7478,
      "step": 1650
    },
    {
      "epoch": 27.02,
      "eval_accuracy": 0.7741935483870968,
      "eval_loss": 0.6097679734230042,
      "eval_runtime": 329.3249,
      "eval_samples_per_second": 0.659,
      "eval_steps_per_second": 0.085,
      "step": 1652
    },
    {
      "epoch": 28.0,
      "learning_rate": 2.9281418170307063e-05,
      "loss": 0.5044,
      "step": 1660
    },
    {
      "epoch": 28.01,
      "learning_rate": 2.9123140234251344e-05,
      "loss": 0.5792,
      "step": 1670
    },
    {
      "epoch": 28.01,
      "learning_rate": 2.8964862298195632e-05,
      "loss": 0.5314,
      "step": 1680
    },
    {
      "epoch": 28.01,
      "learning_rate": 2.880658436213992e-05,
      "loss": 0.4202,
      "step": 1690
    },
    {
      "epoch": 28.01,
      "learning_rate": 2.8648306426084204e-05,
      "loss": 0.4768,
      "step": 1700
    },
    {
      "epoch": 28.02,
      "learning_rate": 2.8490028490028492e-05,
      "loss": 0.7204,
      "step": 1710
    },
    {
      "epoch": 28.02,
      "eval_accuracy": 0.7972350230414746,
      "eval_loss": 0.6458683013916016,
      "eval_runtime": 340.899,
      "eval_samples_per_second": 0.637,
      "eval_steps_per_second": 0.082,
      "step": 1711
    },
    {
      "epoch": 29.0,
      "learning_rate": 2.833175055397278e-05,
      "loss": 0.5258,
      "step": 1720
    },
    {
      "epoch": 29.01,
      "learning_rate": 2.817347261791706e-05,
      "loss": 0.7548,
      "step": 1730
    },
    {
      "epoch": 29.01,
      "learning_rate": 2.8015194681861352e-05,
      "loss": 0.6974,
      "step": 1740
    },
    {
      "epoch": 29.01,
      "learning_rate": 2.7856916745805633e-05,
      "loss": 0.5245,
      "step": 1750
    },
    {
      "epoch": 29.01,
      "learning_rate": 2.769863880974992e-05,
      "loss": 0.6382,
      "step": 1760
    },
    {
      "epoch": 29.02,
      "learning_rate": 2.754036087369421e-05,
      "loss": 0.4325,
      "step": 1770
    },
    {
      "epoch": 29.02,
      "eval_accuracy": 0.7511520737327189,
      "eval_loss": 0.726778507232666,
      "eval_runtime": 317.8406,
      "eval_samples_per_second": 0.683,
      "eval_steps_per_second": 0.088,
      "step": 1770
    },
    {
      "epoch": 30.0,
      "learning_rate": 2.7382082937638494e-05,
      "loss": 0.4622,
      "step": 1780
    },
    {
      "epoch": 30.01,
      "learning_rate": 2.722380500158278e-05,
      "loss": 0.6414,
      "step": 1790
    },
    {
      "epoch": 30.01,
      "learning_rate": 2.706552706552707e-05,
      "loss": 0.72,
      "step": 1800
    },
    {
      "epoch": 30.01,
      "learning_rate": 2.690724912947135e-05,
      "loss": 0.5362,
      "step": 1810
    },
    {
      "epoch": 30.01,
      "learning_rate": 2.6748971193415638e-05,
      "loss": 0.593,
      "step": 1820
    },
    {
      "epoch": 30.02,
      "eval_accuracy": 0.7880184331797235,
      "eval_loss": 0.5901350975036621,
      "eval_runtime": 316.6009,
      "eval_samples_per_second": 0.685,
      "eval_steps_per_second": 0.088,
      "step": 1829
    },
    {
      "epoch": 31.0,
      "learning_rate": 2.6590693257359926e-05,
      "loss": 0.4719,
      "step": 1830
    },
    {
      "epoch": 31.0,
      "learning_rate": 2.643241532130421e-05,
      "loss": 0.461,
      "step": 1840
    },
    {
      "epoch": 31.01,
      "learning_rate": 2.62741373852485e-05,
      "loss": 0.4394,
      "step": 1850
    },
    {
      "epoch": 31.01,
      "learning_rate": 2.611585944919278e-05,
      "loss": 0.7129,
      "step": 1860
    },
    {
      "epoch": 31.01,
      "learning_rate": 2.595758151313707e-05,
      "loss": 0.6618,
      "step": 1870
    },
    {
      "epoch": 31.01,
      "learning_rate": 2.579930357708136e-05,
      "loss": 0.6432,
      "step": 1880
    },
    {
      "epoch": 31.02,
      "eval_accuracy": 0.7880184331797235,
      "eval_loss": 0.5924447774887085,
      "eval_runtime": 310.8109,
      "eval_samples_per_second": 0.698,
      "eval_steps_per_second": 0.09,
      "step": 1888
    },
    {
      "epoch": 32.0,
      "learning_rate": 2.564102564102564e-05,
      "loss": 0.4229,
      "step": 1890
    },
    {
      "epoch": 32.0,
      "learning_rate": 2.5482747704969927e-05,
      "loss": 0.7339,
      "step": 1900
    },
    {
      "epoch": 32.01,
      "learning_rate": 2.5324469768914215e-05,
      "loss": 0.6894,
      "step": 1910
    },
    {
      "epoch": 32.01,
      "learning_rate": 2.51661918328585e-05,
      "loss": 0.5088,
      "step": 1920
    },
    {
      "epoch": 32.01,
      "learning_rate": 2.5007913896802788e-05,
      "loss": 0.4812,
      "step": 1930
    },
    {
      "epoch": 32.01,
      "learning_rate": 2.4849635960747072e-05,
      "loss": 0.4821,
      "step": 1940
    },
    {
      "epoch": 32.02,
      "eval_accuracy": 0.815668202764977,
      "eval_loss": 0.5730257630348206,
      "eval_runtime": 290.7811,
      "eval_samples_per_second": 0.746,
      "eval_steps_per_second": 0.096,
      "step": 1947
    },
    {
      "epoch": 33.0,
      "learning_rate": 2.4691358024691357e-05,
      "loss": 0.4808,
      "step": 1950
    },
    {
      "epoch": 33.0,
      "learning_rate": 2.4533080088635644e-05,
      "loss": 0.4051,
      "step": 1960
    },
    {
      "epoch": 33.01,
      "learning_rate": 2.4374802152579932e-05,
      "loss": 0.559,
      "step": 1970
    },
    {
      "epoch": 33.01,
      "learning_rate": 2.4216524216524217e-05,
      "loss": 0.4306,
      "step": 1980
    },
    {
      "epoch": 33.01,
      "learning_rate": 2.4058246280468505e-05,
      "loss": 0.7199,
      "step": 1990
    },
    {
      "epoch": 33.02,
      "learning_rate": 2.389996834441279e-05,
      "loss": 0.9189,
      "step": 2000
    },
    {
      "epoch": 33.02,
      "eval_accuracy": 0.783410138248848,
      "eval_loss": 0.6241814494132996,
      "eval_runtime": 285.8334,
      "eval_samples_per_second": 0.759,
      "eval_steps_per_second": 0.098,
      "step": 2006
    },
    {
      "epoch": 34.0,
      "learning_rate": 2.3741690408357077e-05,
      "loss": 0.5293,
      "step": 2010
    },
    {
      "epoch": 34.0,
      "learning_rate": 2.358341247230136e-05,
      "loss": 0.5029,
      "step": 2020
    },
    {
      "epoch": 34.01,
      "learning_rate": 2.342513453624565e-05,
      "loss": 0.7058,
      "step": 2030
    },
    {
      "epoch": 34.01,
      "learning_rate": 2.3266856600189934e-05,
      "loss": 0.4917,
      "step": 2040
    },
    {
      "epoch": 34.01,
      "learning_rate": 2.310857866413422e-05,
      "loss": 0.6182,
      "step": 2050
    },
    {
      "epoch": 34.02,
      "learning_rate": 2.295030072807851e-05,
      "loss": 0.6179,
      "step": 2060
    },
    {
      "epoch": 34.02,
      "eval_accuracy": 0.8018433179723502,
      "eval_loss": 0.584676206111908,
      "eval_runtime": 329.7075,
      "eval_samples_per_second": 0.658,
      "eval_steps_per_second": 0.085,
      "step": 2065
    },
    {
      "epoch": 35.0,
      "learning_rate": 2.2792022792022794e-05,
      "loss": 0.4189,
      "step": 2070
    },
    {
      "epoch": 35.0,
      "learning_rate": 2.2633744855967078e-05,
      "loss": 0.3077,
      "step": 2080
    },
    {
      "epoch": 35.01,
      "learning_rate": 2.2475466919911363e-05,
      "loss": 0.4646,
      "step": 2090
    },
    {
      "epoch": 35.01,
      "learning_rate": 2.2317188983855654e-05,
      "loss": 0.4642,
      "step": 2100
    },
    {
      "epoch": 35.01,
      "learning_rate": 2.215891104779994e-05,
      "loss": 0.658,
      "step": 2110
    },
    {
      "epoch": 35.02,
      "learning_rate": 2.2000633111744223e-05,
      "loss": 0.5767,
      "step": 2120
    },
    {
      "epoch": 35.02,
      "eval_accuracy": 0.8248847926267281,
      "eval_loss": 0.59653639793396,
      "eval_runtime": 363.307,
      "eval_samples_per_second": 0.597,
      "eval_steps_per_second": 0.077,
      "step": 2124
    },
    {
      "epoch": 36.0,
      "learning_rate": 2.184235517568851e-05,
      "loss": 0.5256,
      "step": 2130
    },
    {
      "epoch": 36.0,
      "learning_rate": 2.1684077239632795e-05,
      "loss": 0.7286,
      "step": 2140
    },
    {
      "epoch": 36.01,
      "learning_rate": 2.1525799303577083e-05,
      "loss": 0.5342,
      "step": 2150
    },
    {
      "epoch": 36.01,
      "learning_rate": 2.1367521367521368e-05,
      "loss": 0.5483,
      "step": 2160
    },
    {
      "epoch": 36.01,
      "learning_rate": 2.1209243431465655e-05,
      "loss": 0.4426,
      "step": 2170
    },
    {
      "epoch": 36.02,
      "learning_rate": 2.105096549540994e-05,
      "loss": 0.5298,
      "step": 2180
    },
    {
      "epoch": 36.02,
      "eval_accuracy": 0.7235023041474654,
      "eval_loss": 0.7917733788490295,
      "eval_runtime": 318.2413,
      "eval_samples_per_second": 0.682,
      "eval_steps_per_second": 0.088,
      "step": 2183
    },
    {
      "epoch": 37.0,
      "learning_rate": 2.0892687559354228e-05,
      "loss": 0.4776,
      "step": 2190
    },
    {
      "epoch": 37.0,
      "learning_rate": 2.0734409623298512e-05,
      "loss": 0.4896,
      "step": 2200
    },
    {
      "epoch": 37.01,
      "learning_rate": 2.05761316872428e-05,
      "loss": 0.4383,
      "step": 2210
    },
    {
      "epoch": 37.01,
      "learning_rate": 2.0417853751187084e-05,
      "loss": 0.5552,
      "step": 2220
    },
    {
      "epoch": 37.01,
      "learning_rate": 2.0259575815131372e-05,
      "loss": 0.3671,
      "step": 2230
    },
    {
      "epoch": 37.02,
      "learning_rate": 2.010129787907566e-05,
      "loss": 0.5651,
      "step": 2240
    },
    {
      "epoch": 37.02,
      "eval_accuracy": 0.7327188940092166,
      "eval_loss": 0.8338220119476318,
      "eval_runtime": 318.76,
      "eval_samples_per_second": 0.681,
      "eval_steps_per_second": 0.088,
      "step": 2242
    },
    {
      "epoch": 38.0,
      "learning_rate": 1.9943019943019945e-05,
      "loss": 0.5144,
      "step": 2250
    },
    {
      "epoch": 38.01,
      "learning_rate": 1.978474200696423e-05,
      "loss": 0.3722,
      "step": 2260
    },
    {
      "epoch": 38.01,
      "learning_rate": 1.9626464070908514e-05,
      "loss": 0.5191,
      "step": 2270
    },
    {
      "epoch": 38.01,
      "learning_rate": 1.9468186134852805e-05,
      "loss": 0.5935,
      "step": 2280
    },
    {
      "epoch": 38.01,
      "learning_rate": 1.930990819879709e-05,
      "loss": 0.5033,
      "step": 2290
    },
    {
      "epoch": 38.02,
      "learning_rate": 1.9151630262741374e-05,
      "loss": 0.9236,
      "step": 2300
    },
    {
      "epoch": 38.02,
      "eval_accuracy": 0.7142857142857143,
      "eval_loss": 0.8370778560638428,
      "eval_runtime": 304.2351,
      "eval_samples_per_second": 0.713,
      "eval_steps_per_second": 0.092,
      "step": 2301
    },
    {
      "epoch": 39.0,
      "learning_rate": 1.899335232668566e-05,
      "loss": 0.3032,
      "step": 2310
    },
    {
      "epoch": 39.01,
      "learning_rate": 1.8835074390629946e-05,
      "loss": 0.6756,
      "step": 2320
    },
    {
      "epoch": 39.01,
      "learning_rate": 1.8676796454574234e-05,
      "loss": 0.3696,
      "step": 2330
    },
    {
      "epoch": 39.01,
      "learning_rate": 1.8518518518518518e-05,
      "loss": 0.5895,
      "step": 2340
    },
    {
      "epoch": 39.01,
      "learning_rate": 1.8360240582462806e-05,
      "loss": 0.6435,
      "step": 2350
    },
    {
      "epoch": 39.02,
      "learning_rate": 1.820196264640709e-05,
      "loss": 0.4854,
      "step": 2360
    },
    {
      "epoch": 39.02,
      "eval_accuracy": 0.7695852534562212,
      "eval_loss": 0.7114577889442444,
      "eval_runtime": 292.8077,
      "eval_samples_per_second": 0.741,
      "eval_steps_per_second": 0.096,
      "step": 2360
    },
    {
      "epoch": 40.0,
      "learning_rate": 1.804368471035138e-05,
      "loss": 0.534,
      "step": 2370
    },
    {
      "epoch": 40.01,
      "learning_rate": 1.7885406774295663e-05,
      "loss": 0.5408,
      "step": 2380
    },
    {
      "epoch": 40.01,
      "learning_rate": 1.772712883823995e-05,
      "loss": 0.5112,
      "step": 2390
    },
    {
      "epoch": 40.01,
      "learning_rate": 1.7568850902184235e-05,
      "loss": 0.4957,
      "step": 2400
    },
    {
      "epoch": 40.01,
      "learning_rate": 1.7410572966128523e-05,
      "loss": 0.4837,
      "step": 2410
    },
    {
      "epoch": 40.02,
      "eval_accuracy": 0.7695852534562212,
      "eval_loss": 0.6325687170028687,
      "eval_runtime": 288.9229,
      "eval_samples_per_second": 0.751,
      "eval_steps_per_second": 0.097,
      "step": 2419
    },
    {
      "epoch": 41.0,
      "learning_rate": 1.725229503007281e-05,
      "loss": 0.6155,
      "step": 2420
    },
    {
      "epoch": 41.0,
      "learning_rate": 1.7094017094017095e-05,
      "loss": 0.4444,
      "step": 2430
    },
    {
      "epoch": 41.01,
      "learning_rate": 1.693573915796138e-05,
      "loss": 0.5507,
      "step": 2440
    },
    {
      "epoch": 41.01,
      "learning_rate": 1.6777461221905664e-05,
      "loss": 0.5443,
      "step": 2450
    },
    {
      "epoch": 41.01,
      "learning_rate": 1.6619183285849956e-05,
      "loss": 0.5536,
      "step": 2460
    },
    {
      "epoch": 41.01,
      "learning_rate": 1.646090534979424e-05,
      "loss": 0.4142,
      "step": 2470
    },
    {
      "epoch": 41.02,
      "eval_accuracy": 0.8202764976958525,
      "eval_loss": 0.6266201138496399,
      "eval_runtime": 291.4173,
      "eval_samples_per_second": 0.745,
      "eval_steps_per_second": 0.096,
      "step": 2478
    },
    {
      "epoch": 42.0,
      "learning_rate": 1.6302627413738524e-05,
      "loss": 0.7275,
      "step": 2480
    },
    {
      "epoch": 42.0,
      "learning_rate": 1.6144349477682812e-05,
      "loss": 0.425,
      "step": 2490
    },
    {
      "epoch": 42.01,
      "learning_rate": 1.5986071541627097e-05,
      "loss": 0.445,
      "step": 2500
    },
    {
      "epoch": 42.01,
      "learning_rate": 1.5827793605571385e-05,
      "loss": 0.5725,
      "step": 2510
    },
    {
      "epoch": 42.01,
      "learning_rate": 1.566951566951567e-05,
      "loss": 0.619,
      "step": 2520
    },
    {
      "epoch": 42.01,
      "learning_rate": 1.5511237733459957e-05,
      "loss": 0.3309,
      "step": 2530
    },
    {
      "epoch": 42.02,
      "eval_accuracy": 0.8110599078341014,
      "eval_loss": 0.5805542469024658,
      "eval_runtime": 347.0673,
      "eval_samples_per_second": 0.625,
      "eval_steps_per_second": 0.081,
      "step": 2537
    },
    {
      "epoch": 43.0,
      "learning_rate": 1.535295979740424e-05,
      "loss": 0.5073,
      "step": 2540
    },
    {
      "epoch": 43.0,
      "learning_rate": 1.5194681861348528e-05,
      "loss": 0.5336,
      "step": 2550
    },
    {
      "epoch": 43.01,
      "learning_rate": 1.5036403925292817e-05,
      "loss": 0.4556,
      "step": 2560
    },
    {
      "epoch": 43.01,
      "learning_rate": 1.4878125989237102e-05,
      "loss": 0.5441,
      "step": 2570
    },
    {
      "epoch": 43.01,
      "learning_rate": 1.4719848053181388e-05,
      "loss": 0.4513,
      "step": 2580
    },
    {
      "epoch": 43.02,
      "learning_rate": 1.4561570117125672e-05,
      "loss": 0.3939,
      "step": 2590
    },
    {
      "epoch": 43.02,
      "eval_accuracy": 0.8248847926267281,
      "eval_loss": 0.5745505690574646,
      "eval_runtime": 351.9313,
      "eval_samples_per_second": 0.617,
      "eval_steps_per_second": 0.08,
      "step": 2596
    },
    {
      "epoch": 44.0,
      "learning_rate": 1.440329218106996e-05,
      "loss": 0.4569,
      "step": 2600
    },
    {
      "epoch": 44.0,
      "learning_rate": 1.4245014245014246e-05,
      "loss": 0.3418,
      "step": 2610
    },
    {
      "epoch": 44.01,
      "learning_rate": 1.408673630895853e-05,
      "loss": 0.5716,
      "step": 2620
    },
    {
      "epoch": 44.01,
      "learning_rate": 1.3928458372902817e-05,
      "loss": 0.4198,
      "step": 2630
    },
    {
      "epoch": 44.01,
      "learning_rate": 1.3770180436847105e-05,
      "loss": 0.5132,
      "step": 2640
    },
    {
      "epoch": 44.02,
      "learning_rate": 1.361190250079139e-05,
      "loss": 0.4993,
      "step": 2650
    },
    {
      "epoch": 44.02,
      "eval_accuracy": 0.8387096774193549,
      "eval_loss": 0.6113536357879639,
      "eval_runtime": 312.9643,
      "eval_samples_per_second": 0.693,
      "eval_steps_per_second": 0.089,
      "step": 2655
    },
    {
      "epoch": 45.0,
      "learning_rate": 1.3453624564735675e-05,
      "loss": 0.4347,
      "step": 2660
    },
    {
      "epoch": 45.0,
      "learning_rate": 1.3295346628679963e-05,
      "loss": 0.6201,
      "step": 2670
    },
    {
      "epoch": 45.01,
      "learning_rate": 1.313706869262425e-05,
      "loss": 0.4937,
      "step": 2680
    },
    {
      "epoch": 45.01,
      "learning_rate": 1.2978790756568535e-05,
      "loss": 0.4629,
      "step": 2690
    },
    {
      "epoch": 45.01,
      "learning_rate": 1.282051282051282e-05,
      "loss": 0.6889,
      "step": 2700
    },
    {
      "epoch": 45.02,
      "learning_rate": 1.2662234884457108e-05,
      "loss": 0.3785,
      "step": 2710
    },
    {
      "epoch": 45.02,
      "eval_accuracy": 0.8018433179723502,
      "eval_loss": 0.6317055821418762,
      "eval_runtime": 311.0444,
      "eval_samples_per_second": 0.698,
      "eval_steps_per_second": 0.09,
      "step": 2714
    },
    {
      "epoch": 46.0,
      "learning_rate": 1.2503956948401394e-05,
      "loss": 0.3386,
      "step": 2720
    },
    {
      "epoch": 46.0,
      "learning_rate": 1.2345679012345678e-05,
      "loss": 0.5517,
      "step": 2730
    },
    {
      "epoch": 46.01,
      "learning_rate": 1.2187401076289966e-05,
      "loss": 0.3834,
      "step": 2740
    },
    {
      "epoch": 46.01,
      "learning_rate": 1.2029123140234252e-05,
      "loss": 0.4447,
      "step": 2750
    },
    {
      "epoch": 46.01,
      "learning_rate": 1.1870845204178538e-05,
      "loss": 0.3444,
      "step": 2760
    },
    {
      "epoch": 46.02,
      "learning_rate": 1.1712567268122825e-05,
      "loss": 0.5224,
      "step": 2770
    },
    {
      "epoch": 46.02,
      "eval_accuracy": 0.7972350230414746,
      "eval_loss": 0.666746199131012,
      "eval_runtime": 298.4433,
      "eval_samples_per_second": 0.727,
      "eval_steps_per_second": 0.094,
      "step": 2773
    },
    {
      "epoch": 47.0,
      "learning_rate": 1.155428933206711e-05,
      "loss": 0.4419,
      "step": 2780
    },
    {
      "epoch": 47.0,
      "learning_rate": 1.1396011396011397e-05,
      "loss": 0.3244,
      "step": 2790
    },
    {
      "epoch": 47.01,
      "learning_rate": 1.1237733459955681e-05,
      "loss": 0.5273,
      "step": 2800
    },
    {
      "epoch": 47.01,
      "learning_rate": 1.107945552389997e-05,
      "loss": 0.3442,
      "step": 2810
    },
    {
      "epoch": 47.01,
      "learning_rate": 1.0921177587844255e-05,
      "loss": 0.4701,
      "step": 2820
    },
    {
      "epoch": 47.02,
      "learning_rate": 1.0762899651788542e-05,
      "loss": 0.5705,
      "step": 2830
    },
    {
      "epoch": 47.02,
      "eval_accuracy": 0.7926267281105991,
      "eval_loss": 0.6382005214691162,
      "eval_runtime": 298.8237,
      "eval_samples_per_second": 0.726,
      "eval_steps_per_second": 0.094,
      "step": 2832
    },
    {
      "epoch": 48.0,
      "learning_rate": 1.0604621715732828e-05,
      "loss": 0.5681,
      "step": 2840
    },
    {
      "epoch": 48.01,
      "learning_rate": 1.0446343779677114e-05,
      "loss": 0.4655,
      "step": 2850
    },
    {
      "epoch": 48.01,
      "learning_rate": 1.02880658436214e-05,
      "loss": 0.5486,
      "step": 2860
    },
    {
      "epoch": 48.01,
      "learning_rate": 1.0129787907565686e-05,
      "loss": 0.4268,
      "step": 2870
    },
    {
      "epoch": 48.01,
      "learning_rate": 9.971509971509972e-06,
      "loss": 0.5592,
      "step": 2880
    },
    {
      "epoch": 48.02,
      "learning_rate": 9.813232035454257e-06,
      "loss": 0.3342,
      "step": 2890
    },
    {
      "epoch": 48.02,
      "eval_accuracy": 0.815668202764977,
      "eval_loss": 0.5592498779296875,
      "eval_runtime": 307.6803,
      "eval_samples_per_second": 0.705,
      "eval_steps_per_second": 0.091,
      "step": 2891
    },
    {
      "epoch": 49.0,
      "learning_rate": 9.654954099398545e-06,
      "loss": 0.4835,
      "step": 2900
    },
    {
      "epoch": 49.01,
      "learning_rate": 9.49667616334283e-06,
      "loss": 0.3971,
      "step": 2910
    },
    {
      "epoch": 49.01,
      "learning_rate": 9.338398227287117e-06,
      "loss": 0.4908,
      "step": 2920
    },
    {
      "epoch": 49.01,
      "learning_rate": 9.180120291231403e-06,
      "loss": 0.4665,
      "step": 2930
    },
    {
      "epoch": 49.01,
      "learning_rate": 9.02184235517569e-06,
      "loss": 0.4262,
      "step": 2940
    },
    {
      "epoch": 49.02,
      "learning_rate": 8.863564419119975e-06,
      "loss": 0.5044,
      "step": 2950
    },
    {
      "epoch": 49.02,
      "eval_accuracy": 0.7926267281105991,
      "eval_loss": 0.574846625328064,
      "eval_runtime": 296.7758,
      "eval_samples_per_second": 0.731,
      "eval_steps_per_second": 0.094,
      "step": 2950
    },
    {
      "epoch": 50.0,
      "learning_rate": 8.705286483064262e-06,
      "loss": 0.5795,
      "step": 2960
    },
    {
      "epoch": 50.01,
      "learning_rate": 8.547008547008548e-06,
      "loss": 0.6522,
      "step": 2970
    },
    {
      "epoch": 50.01,
      "learning_rate": 8.388730610952832e-06,
      "loss": 0.4391,
      "step": 2980
    },
    {
      "epoch": 50.01,
      "learning_rate": 8.23045267489712e-06,
      "loss": 0.3763,
      "step": 2990
    },
    {
      "epoch": 50.01,
      "learning_rate": 8.072174738841406e-06,
      "loss": 0.3838,
      "step": 3000
    },
    {
      "epoch": 50.02,
      "eval_accuracy": 0.7788018433179723,
      "eval_loss": 0.6050373315811157,
      "eval_runtime": 338.1899,
      "eval_samples_per_second": 0.642,
      "eval_steps_per_second": 0.083,
      "step": 3009
    },
    {
      "epoch": 51.0,
      "learning_rate": 7.913896802785692e-06,
      "loss": 0.3898,
      "step": 3010
    },
    {
      "epoch": 51.0,
      "learning_rate": 7.755618866729978e-06,
      "loss": 0.4642,
      "step": 3020
    },
    {
      "epoch": 51.01,
      "learning_rate": 7.597340930674264e-06,
      "loss": 0.4582,
      "step": 3030
    },
    {
      "epoch": 51.01,
      "learning_rate": 7.439062994618551e-06,
      "loss": 0.4935,
      "step": 3040
    },
    {
      "epoch": 51.01,
      "learning_rate": 7.280785058562836e-06,
      "loss": 0.3878,
      "step": 3050
    },
    {
      "epoch": 51.01,
      "learning_rate": 7.122507122507123e-06,
      "loss": 0.5099,
      "step": 3060
    },
    {
      "epoch": 51.02,
      "eval_accuracy": 0.7603686635944701,
      "eval_loss": 0.6556740999221802,
      "eval_runtime": 332.6773,
      "eval_samples_per_second": 0.652,
      "eval_steps_per_second": 0.084,
      "step": 3068
    },
    {
      "epoch": 52.0,
      "learning_rate": 6.964229186451408e-06,
      "loss": 0.4264,
      "step": 3070
    },
    {
      "epoch": 52.0,
      "learning_rate": 6.805951250395695e-06,
      "loss": 0.378,
      "step": 3080
    },
    {
      "epoch": 52.01,
      "learning_rate": 6.6476733143399815e-06,
      "loss": 0.4568,
      "step": 3090
    },
    {
      "epoch": 52.01,
      "learning_rate": 6.489395378284268e-06,
      "loss": 0.4432,
      "step": 3100
    },
    {
      "epoch": 52.01,
      "learning_rate": 6.331117442228554e-06,
      "loss": 0.523,
      "step": 3110
    },
    {
      "epoch": 52.01,
      "learning_rate": 6.172839506172839e-06,
      "loss": 0.4335,
      "step": 3120
    },
    {
      "epoch": 52.02,
      "eval_accuracy": 0.7511520737327189,
      "eval_loss": 0.7118790149688721,
      "eval_runtime": 337.6952,
      "eval_samples_per_second": 0.643,
      "eval_steps_per_second": 0.083,
      "step": 3127
    },
    {
      "epoch": 53.0,
      "learning_rate": 6.014561570117126e-06,
      "loss": 0.3422,
      "step": 3130
    },
    {
      "epoch": 53.0,
      "learning_rate": 5.856283634061412e-06,
      "loss": 0.3893,
      "step": 3140
    },
    {
      "epoch": 53.01,
      "learning_rate": 5.6980056980056985e-06,
      "loss": 0.407,
      "step": 3150
    },
    {
      "epoch": 53.01,
      "learning_rate": 5.539727761949985e-06,
      "loss": 0.561,
      "step": 3160
    },
    {
      "epoch": 53.01,
      "learning_rate": 5.381449825894271e-06,
      "loss": 0.2949,
      "step": 3170
    },
    {
      "epoch": 53.02,
      "learning_rate": 5.223171889838557e-06,
      "loss": 0.4122,
      "step": 3180
    },
    {
      "epoch": 53.02,
      "eval_accuracy": 0.7788018433179723,
      "eval_loss": 0.6562320590019226,
      "eval_runtime": 360.9203,
      "eval_samples_per_second": 0.601,
      "eval_steps_per_second": 0.078,
      "step": 3186
    },
    {
      "epoch": 54.0,
      "learning_rate": 5.064893953782843e-06,
      "loss": 0.291,
      "step": 3190
    },
    {
      "epoch": 54.0,
      "learning_rate": 4.906616017727128e-06,
      "loss": 0.3532,
      "step": 3200
    },
    {
      "epoch": 54.01,
      "learning_rate": 4.748338081671415e-06,
      "loss": 0.3911,
      "step": 3210
    },
    {
      "epoch": 54.01,
      "learning_rate": 4.5900601456157015e-06,
      "loss": 0.5269,
      "step": 3220
    },
    {
      "epoch": 54.01,
      "learning_rate": 4.431782209559988e-06,
      "loss": 0.3041,
      "step": 3230
    },
    {
      "epoch": 54.02,
      "learning_rate": 4.273504273504274e-06,
      "loss": 0.4431,
      "step": 3240
    },
    {
      "epoch": 54.02,
      "eval_accuracy": 0.7649769585253456,
      "eval_loss": 0.6700997948646545,
      "eval_runtime": 319.5418,
      "eval_samples_per_second": 0.679,
      "eval_steps_per_second": 0.088,
      "step": 3245
    },
    {
      "epoch": 55.0,
      "learning_rate": 4.11522633744856e-06,
      "loss": 0.5306,
      "step": 3250
    },
    {
      "epoch": 55.0,
      "learning_rate": 3.956948401392846e-06,
      "loss": 0.5086,
      "step": 3260
    },
    {
      "epoch": 55.01,
      "learning_rate": 3.798670465337132e-06,
      "loss": 0.3345,
      "step": 3270
    },
    {
      "epoch": 55.01,
      "learning_rate": 3.640392529281418e-06,
      "loss": 0.3607,
      "step": 3280
    },
    {
      "epoch": 55.01,
      "learning_rate": 3.482114593225704e-06,
      "loss": 0.4845,
      "step": 3290
    },
    {
      "epoch": 55.02,
      "learning_rate": 3.3238366571699908e-06,
      "loss": 0.4536,
      "step": 3300
    },
    {
      "epoch": 55.02,
      "eval_accuracy": 0.7695852534562212,
      "eval_loss": 0.6659231781959534,
      "eval_runtime": 312.5964,
      "eval_samples_per_second": 0.694,
      "eval_steps_per_second": 0.09,
      "step": 3304
    },
    {
      "epoch": 56.0,
      "learning_rate": 3.165558721114277e-06,
      "loss": 0.3168,
      "step": 3310
    },
    {
      "epoch": 56.0,
      "learning_rate": 3.007280785058563e-06,
      "loss": 0.459,
      "step": 3320
    },
    {
      "epoch": 56.01,
      "learning_rate": 2.8490028490028492e-06,
      "loss": 0.4485,
      "step": 3330
    },
    {
      "epoch": 56.01,
      "learning_rate": 2.6907249129471354e-06,
      "loss": 0.4044,
      "step": 3340
    },
    {
      "epoch": 56.01,
      "learning_rate": 2.5324469768914215e-06,
      "loss": 0.4181,
      "step": 3350
    },
    {
      "epoch": 56.02,
      "learning_rate": 2.3741690408357077e-06,
      "loss": 0.3867,
      "step": 3360
    },
    {
      "epoch": 56.02,
      "eval_accuracy": 0.7788018433179723,
      "eval_loss": 0.6631873846054077,
      "eval_runtime": 302.7333,
      "eval_samples_per_second": 0.717,
      "eval_steps_per_second": 0.092,
      "step": 3363
    },
    {
      "epoch": 57.0,
      "learning_rate": 2.215891104779994e-06,
      "loss": 0.4615,
      "step": 3370
    },
    {
      "epoch": 57.0,
      "learning_rate": 2.05761316872428e-06,
      "loss": 0.459,
      "step": 3380
    },
    {
      "epoch": 57.01,
      "learning_rate": 1.899335232668566e-06,
      "loss": 0.2905,
      "step": 3390
    },
    {
      "epoch": 57.01,
      "learning_rate": 1.741057296612852e-06,
      "loss": 0.3776,
      "step": 3400
    },
    {
      "epoch": 57.01,
      "learning_rate": 1.5827793605571385e-06,
      "loss": 0.4584,
      "step": 3410
    },
    {
      "epoch": 57.02,
      "learning_rate": 1.4245014245014246e-06,
      "loss": 0.3878,
      "step": 3420
    },
    {
      "epoch": 57.02,
      "eval_accuracy": 0.7741935483870968,
      "eval_loss": 0.6911003589630127,
      "eval_runtime": 288.2901,
      "eval_samples_per_second": 0.753,
      "eval_steps_per_second": 0.097,
      "step": 3422
    },
    {
      "epoch": 58.0,
      "learning_rate": 1.2662234884457108e-06,
      "loss": 0.3642,
      "step": 3430
    },
    {
      "epoch": 58.01,
      "learning_rate": 1.107945552389997e-06,
      "loss": 0.5532,
      "step": 3440
    },
    {
      "epoch": 58.01,
      "learning_rate": 9.49667616334283e-07,
      "loss": 0.3968,
      "step": 3450
    },
    {
      "epoch": 58.01,
      "learning_rate": 7.913896802785692e-07,
      "loss": 0.4202,
      "step": 3460
    },
    {
      "epoch": 58.01,
      "learning_rate": 6.331117442228554e-07,
      "loss": 0.5291,
      "step": 3470
    },
    {
      "epoch": 58.02,
      "learning_rate": 4.748338081671415e-07,
      "loss": 0.3853,
      "step": 3480
    },
    {
      "epoch": 58.02,
      "eval_accuracy": 0.7788018433179723,
      "eval_loss": 0.6931061148643494,
      "eval_runtime": 338.901,
      "eval_samples_per_second": 0.64,
      "eval_steps_per_second": 0.083,
      "step": 3481
    },
    {
      "epoch": 59.0,
      "learning_rate": 3.165558721114277e-07,
      "loss": 0.4615,
      "step": 3490
    },
    {
      "epoch": 59.01,
      "learning_rate": 1.5827793605571385e-07,
      "loss": 0.4759,
      "step": 3500
    },
    {
      "epoch": 59.01,
      "learning_rate": 0.0,
      "loss": 0.3268,
      "step": 3510
    },
    {
      "epoch": 59.01,
      "eval_accuracy": 0.7788018433179723,
      "eval_loss": 0.6913787126541138,
      "eval_runtime": 307.5387,
      "eval_samples_per_second": 0.706,
      "eval_steps_per_second": 0.091,
      "step": 3510
    },
    {
      "epoch": 59.01,
      "step": 3510,
      "total_flos": 3.476987046291161e+19,
      "train_loss": 0.68851605188473,
      "train_runtime": 67730.7107,
      "train_samples_per_second": 0.415,
      "train_steps_per_second": 0.052
    },
    {
      "epoch": 59.01,
      "eval_accuracy": 0.8148148148148148,
      "eval_loss": 0.567668616771698,
      "eval_runtime": 331.762,
      "eval_samples_per_second": 0.651,
      "eval_steps_per_second": 0.081,
      "step": 3510
    },
    {
      "epoch": 59.01,
      "eval_accuracy": 0.8148148148148148,
      "eval_loss": 0.5676685571670532,
      "eval_runtime": 292.9189,
      "eval_samples_per_second": 0.737,
      "eval_steps_per_second": 0.092,
      "step": 3510
    },
    {
      "epoch": 59.01,
      "eval_accuracy": 0.8165137614678899,
      "eval_loss": 0.5670480728149414,
      "eval_runtime": 312.7468,
      "eval_samples_per_second": 0.697,
      "eval_steps_per_second": 0.09,
      "step": 3510
    }
  ],
  "logging_steps": 10,
  "max_steps": 3510,
  "num_train_epochs": 9223372036854775807,
  "save_steps": 500,
  "total_flos": 3.476987046291161e+19,
  "trial_name": null,
  "trial_params": null
}