{
  "best_metric": 0.6355140186915887,
  "best_model_checkpoint": "BEiT-RHS-DA\\checkpoint-360",
  "epoch": 39.111111111111114,
  "eval_steps": 500,
  "global_step": 880,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.44,
      "learning_rate": 1.1363636363636365e-05,
      "loss": 1.3359,
      "step": 10
    },
    {
      "epoch": 0.89,
      "learning_rate": 2.272727272727273e-05,
      "loss": 1.2357,
      "step": 20
    },
    {
      "epoch": 0.98,
      "eval_accuracy": 0.5887850467289719,
      "eval_loss": 0.7114280462265015,
      "eval_runtime": 2.6053,
      "eval_samples_per_second": 41.07,
      "eval_steps_per_second": 2.687,
      "step": 22
    },
    {
      "epoch": 1.33,
      "learning_rate": 3.409090909090909e-05,
      "loss": 0.9874,
      "step": 30
    },
    {
      "epoch": 1.78,
      "learning_rate": 4.545454545454546e-05,
      "loss": 0.6596,
      "step": 40
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.5981308411214953,
      "eval_loss": 0.7058635950088501,
      "eval_runtime": 2.8903,
      "eval_samples_per_second": 37.02,
      "eval_steps_per_second": 2.422,
      "step": 45
    },
    {
      "epoch": 2.22,
      "learning_rate": 4.964114832535885e-05,
      "loss": 0.3724,
      "step": 50
    },
    {
      "epoch": 2.67,
      "learning_rate": 4.904306220095694e-05,
      "loss": 0.206,
      "step": 60
    },
    {
      "epoch": 2.98,
      "eval_accuracy": 0.5981308411214953,
      "eval_loss": 1.14487624168396,
      "eval_runtime": 2.567,
      "eval_samples_per_second": 41.682,
      "eval_steps_per_second": 2.727,
      "step": 67
    },
    {
      "epoch": 3.11,
      "learning_rate": 4.844497607655503e-05,
      "loss": 0.1872,
      "step": 70
    },
    {
      "epoch": 3.56,
      "learning_rate": 4.784688995215311e-05,
      "loss": 0.1164,
      "step": 80
    },
    {
      "epoch": 4.0,
      "learning_rate": 4.72488038277512e-05,
      "loss": 0.1664,
      "step": 90
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.3925233644859813,
      "eval_loss": 2.2061808109283447,
      "eval_runtime": 2.5208,
      "eval_samples_per_second": 42.448,
      "eval_steps_per_second": 2.777,
      "step": 90
    },
    {
      "epoch": 4.44,
      "learning_rate": 4.665071770334928e-05,
      "loss": 0.1706,
      "step": 100
    },
    {
      "epoch": 4.89,
      "learning_rate": 4.605263157894737e-05,
      "loss": 0.1011,
      "step": 110
    },
    {
      "epoch": 4.98,
      "eval_accuracy": 0.4672897196261682,
      "eval_loss": 2.040870189666748,
      "eval_runtime": 2.5598,
      "eval_samples_per_second": 41.8,
      "eval_steps_per_second": 2.735,
      "step": 112
    },
    {
      "epoch": 5.33,
      "learning_rate": 4.545454545454546e-05,
      "loss": 0.2439,
      "step": 120
    },
    {
      "epoch": 5.78,
      "learning_rate": 4.485645933014354e-05,
      "loss": 0.0653,
      "step": 130
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.6261682242990654,
      "eval_loss": 1.3038437366485596,
      "eval_runtime": 2.6986,
      "eval_samples_per_second": 39.65,
      "eval_steps_per_second": 2.594,
      "step": 135
    },
    {
      "epoch": 6.22,
      "learning_rate": 4.425837320574163e-05,
      "loss": 0.2341,
      "step": 140
    },
    {
      "epoch": 6.67,
      "learning_rate": 4.366028708133971e-05,
      "loss": 0.2843,
      "step": 150
    },
    {
      "epoch": 6.98,
      "eval_accuracy": 0.5981308411214953,
      "eval_loss": 1.7210038900375366,
      "eval_runtime": 2.6158,
      "eval_samples_per_second": 40.905,
      "eval_steps_per_second": 2.676,
      "step": 157
    },
    {
      "epoch": 7.11,
      "learning_rate": 4.3062200956937806e-05,
      "loss": 0.0424,
      "step": 160
    },
    {
      "epoch": 7.56,
      "learning_rate": 4.246411483253589e-05,
      "loss": 0.0082,
      "step": 170
    },
    {
      "epoch": 8.0,
      "learning_rate": 4.1866028708133976e-05,
      "loss": 0.059,
      "step": 180
    },
    {
      "epoch": 8.0,
      "eval_accuracy": 0.4672897196261682,
      "eval_loss": 2.870556116104126,
      "eval_runtime": 2.8378,
      "eval_samples_per_second": 37.705,
      "eval_steps_per_second": 2.467,
      "step": 180
    },
    {
      "epoch": 8.44,
      "learning_rate": 4.1267942583732064e-05,
      "loss": 0.1224,
      "step": 190
    },
    {
      "epoch": 8.89,
      "learning_rate": 4.0669856459330146e-05,
      "loss": 0.1318,
      "step": 200
    },
    {
      "epoch": 8.98,
      "eval_accuracy": 0.5887850467289719,
      "eval_loss": 2.451904058456421,
      "eval_runtime": 2.6549,
      "eval_samples_per_second": 40.303,
      "eval_steps_per_second": 2.637,
      "step": 202
    },
    {
      "epoch": 9.33,
      "learning_rate": 4.0071770334928235e-05,
      "loss": 0.0267,
      "step": 210
    },
    {
      "epoch": 9.78,
      "learning_rate": 3.9473684210526316e-05,
      "loss": 0.0501,
      "step": 220
    },
    {
      "epoch": 10.0,
      "eval_accuracy": 0.5887850467289719,
      "eval_loss": 2.2037127017974854,
      "eval_runtime": 2.6808,
      "eval_samples_per_second": 39.913,
      "eval_steps_per_second": 2.611,
      "step": 225
    },
    {
      "epoch": 10.22,
      "learning_rate": 3.8875598086124405e-05,
      "loss": 0.0481,
      "step": 230
    },
    {
      "epoch": 10.67,
      "learning_rate": 3.8277511961722486e-05,
      "loss": 0.054,
      "step": 240
    },
    {
      "epoch": 10.98,
      "eval_accuracy": 0.5887850467289719,
      "eval_loss": 2.646747589111328,
      "eval_runtime": 2.8537,
      "eval_samples_per_second": 37.495,
      "eval_steps_per_second": 2.453,
      "step": 247
    },
    {
      "epoch": 11.11,
      "learning_rate": 3.7679425837320575e-05,
      "loss": 0.0495,
      "step": 250
    },
    {
      "epoch": 11.56,
      "learning_rate": 3.7081339712918663e-05,
      "loss": 0.0367,
      "step": 260
    },
    {
      "epoch": 12.0,
      "learning_rate": 3.6483253588516745e-05,
      "loss": 0.0263,
      "step": 270
    },
    {
      "epoch": 12.0,
      "eval_accuracy": 0.5981308411214953,
      "eval_loss": 2.4033188819885254,
      "eval_runtime": 2.6437,
      "eval_samples_per_second": 40.474,
      "eval_steps_per_second": 2.648,
      "step": 270
    },
    {
      "epoch": 12.44,
      "learning_rate": 3.5885167464114834e-05,
      "loss": 0.0138,
      "step": 280
    },
    {
      "epoch": 12.89,
      "learning_rate": 3.5287081339712915e-05,
      "loss": 0.0553,
      "step": 290
    },
    {
      "epoch": 12.98,
      "eval_accuracy": 0.5887850467289719,
      "eval_loss": 1.6589338779449463,
      "eval_runtime": 2.6219,
      "eval_samples_per_second": 40.811,
      "eval_steps_per_second": 2.67,
      "step": 292
    },
    {
      "epoch": 13.33,
      "learning_rate": 3.4688995215311004e-05,
      "loss": 0.0092,
      "step": 300
    },
    {
      "epoch": 13.78,
      "learning_rate": 3.409090909090909e-05,
      "loss": 0.0898,
      "step": 310
    },
    {
      "epoch": 14.0,
      "eval_accuracy": 0.5981308411214953,
      "eval_loss": 1.7657066583633423,
      "eval_runtime": 2.6872,
      "eval_samples_per_second": 39.819,
      "eval_steps_per_second": 2.605,
      "step": 315
    },
    {
      "epoch": 14.22,
      "learning_rate": 3.349282296650718e-05,
      "loss": 0.0304,
      "step": 320
    },
    {
      "epoch": 14.67,
      "learning_rate": 3.289473684210527e-05,
      "loss": 0.0324,
      "step": 330
    },
    {
      "epoch": 14.98,
      "eval_accuracy": 0.5887850467289719,
      "eval_loss": 2.8265936374664307,
      "eval_runtime": 2.6021,
      "eval_samples_per_second": 41.121,
      "eval_steps_per_second": 2.69,
      "step": 337
    },
    {
      "epoch": 15.11,
      "learning_rate": 3.229665071770335e-05,
      "loss": 0.0424,
      "step": 340
    },
    {
      "epoch": 15.56,
      "learning_rate": 3.169856459330144e-05,
      "loss": 0.0556,
      "step": 350
    },
    {
      "epoch": 16.0,
      "learning_rate": 3.110047846889952e-05,
      "loss": 0.0322,
      "step": 360
    },
    {
      "epoch": 16.0,
      "eval_accuracy": 0.6355140186915887,
      "eval_loss": 1.7194138765335083,
      "eval_runtime": 2.6447,
      "eval_samples_per_second": 40.458,
      "eval_steps_per_second": 2.647,
      "step": 360
    },
    {
      "epoch": 16.44,
      "learning_rate": 3.050239234449761e-05,
      "loss": 0.0236,
      "step": 370
    },
    {
      "epoch": 16.89,
      "learning_rate": 2.9904306220095695e-05,
      "loss": 0.03,
      "step": 380
    },
    {
      "epoch": 16.98,
      "eval_accuracy": 0.616822429906542,
      "eval_loss": 2.035153865814209,
      "eval_runtime": 2.6411,
      "eval_samples_per_second": 40.513,
      "eval_steps_per_second": 2.65,
      "step": 382
    },
    {
      "epoch": 17.33,
      "learning_rate": 2.9306220095693783e-05,
      "loss": 0.0244,
      "step": 390
    },
    {
      "epoch": 17.78,
      "learning_rate": 2.8708133971291868e-05,
      "loss": 0.0392,
      "step": 400
    },
    {
      "epoch": 18.0,
      "eval_accuracy": 0.616822429906542,
      "eval_loss": 2.412965774536133,
      "eval_runtime": 2.7163,
      "eval_samples_per_second": 39.392,
      "eval_steps_per_second": 2.577,
      "step": 405
    },
    {
      "epoch": 18.22,
      "learning_rate": 2.8110047846889953e-05,
      "loss": 0.0096,
      "step": 410
    },
    {
      "epoch": 18.67,
      "learning_rate": 2.751196172248804e-05,
      "loss": 0.0428,
      "step": 420
    },
    {
      "epoch": 18.98,
      "eval_accuracy": 0.6074766355140186,
      "eval_loss": 2.0628373622894287,
      "eval_runtime": 2.5916,
      "eval_samples_per_second": 41.288,
      "eval_steps_per_second": 2.701,
      "step": 427
    },
    {
      "epoch": 19.11,
      "learning_rate": 2.6913875598086123e-05,
      "loss": 0.0189,
      "step": 430
    },
    {
      "epoch": 19.56,
      "learning_rate": 2.6315789473684212e-05,
      "loss": 0.0119,
      "step": 440
    },
    {
      "epoch": 20.0,
      "learning_rate": 2.5717703349282297e-05,
      "loss": 0.0127,
      "step": 450
    },
    {
      "epoch": 20.0,
      "eval_accuracy": 0.5887850467289719,
      "eval_loss": 2.743129253387451,
      "eval_runtime": 2.5383,
      "eval_samples_per_second": 42.154,
      "eval_steps_per_second": 2.758,
      "step": 450
    },
    {
      "epoch": 20.44,
      "learning_rate": 2.5119617224880382e-05,
      "loss": 0.0164,
      "step": 460
    },
    {
      "epoch": 20.89,
      "learning_rate": 2.452153110047847e-05,
      "loss": 0.0187,
      "step": 470
    },
    {
      "epoch": 20.98,
      "eval_accuracy": 0.5981308411214953,
      "eval_loss": 2.700892686843872,
      "eval_runtime": 2.592,
      "eval_samples_per_second": 41.281,
      "eval_steps_per_second": 2.701,
      "step": 472
    },
    {
      "epoch": 21.33,
      "learning_rate": 2.3923444976076556e-05,
      "loss": 0.0349,
      "step": 480
    },
    {
      "epoch": 21.78,
      "learning_rate": 2.332535885167464e-05,
      "loss": 0.0469,
      "step": 490
    },
    {
      "epoch": 22.0,
      "eval_accuracy": 0.5981308411214953,
      "eval_loss": 2.578273057937622,
      "eval_runtime": 2.6372,
      "eval_samples_per_second": 40.573,
      "eval_steps_per_second": 2.654,
      "step": 495
    },
    {
      "epoch": 22.22,
      "learning_rate": 2.272727272727273e-05,
      "loss": 0.0196,
      "step": 500
    },
    {
      "epoch": 22.67,
      "learning_rate": 2.2129186602870814e-05,
      "loss": 0.0095,
      "step": 510
    },
    {
      "epoch": 22.98,
      "eval_accuracy": 0.5981308411214953,
      "eval_loss": 2.304006338119507,
      "eval_runtime": 2.5827,
      "eval_samples_per_second": 41.429,
      "eval_steps_per_second": 2.71,
      "step": 517
    },
    {
      "epoch": 23.11,
      "learning_rate": 2.1531100478468903e-05,
      "loss": 0.0428,
      "step": 520
    },
    {
      "epoch": 23.56,
      "learning_rate": 2.0933014354066988e-05,
      "loss": 0.0007,
      "step": 530
    },
    {
      "epoch": 24.0,
      "learning_rate": 2.0334928229665073e-05,
      "loss": 0.0025,
      "step": 540
    },
    {
      "epoch": 24.0,
      "eval_accuracy": 0.616822429906542,
      "eval_loss": 2.521818161010742,
      "eval_runtime": 2.612,
      "eval_samples_per_second": 40.964,
      "eval_steps_per_second": 2.68,
      "step": 540
    },
    {
      "epoch": 24.44,
      "learning_rate": 1.9736842105263158e-05,
      "loss": 0.0005,
      "step": 550
    },
    {
      "epoch": 24.89,
      "learning_rate": 1.9138755980861243e-05,
      "loss": 0.0281,
      "step": 560
    },
    {
      "epoch": 24.98,
      "eval_accuracy": 0.5981308411214953,
      "eval_loss": 3.231044292449951,
      "eval_runtime": 2.5813,
      "eval_samples_per_second": 41.451,
      "eval_steps_per_second": 2.712,
      "step": 562
    },
    {
      "epoch": 25.33,
      "learning_rate": 1.8540669856459332e-05,
      "loss": 0.0084,
      "step": 570
    },
    {
      "epoch": 25.78,
      "learning_rate": 1.7942583732057417e-05,
      "loss": 0.0004,
      "step": 580
    },
    {
      "epoch": 26.0,
      "eval_accuracy": 0.5981308411214953,
      "eval_loss": 3.273077964782715,
      "eval_runtime": 2.6145,
      "eval_samples_per_second": 40.925,
      "eval_steps_per_second": 2.677,
      "step": 585
    },
    {
      "epoch": 26.22,
      "learning_rate": 1.7344497607655502e-05,
      "loss": 0.001,
      "step": 590
    },
    {
      "epoch": 26.67,
      "learning_rate": 1.674641148325359e-05,
      "loss": 0.0109,
      "step": 600
    },
    {
      "epoch": 26.98,
      "eval_accuracy": 0.6261682242990654,
      "eval_loss": 2.480931282043457,
      "eval_runtime": 2.6486,
      "eval_samples_per_second": 40.399,
      "eval_steps_per_second": 2.643,
      "step": 607
    },
    {
      "epoch": 27.11,
      "learning_rate": 1.6148325358851675e-05,
      "loss": 0.0158,
      "step": 610
    },
    {
      "epoch": 27.56,
      "learning_rate": 1.555023923444976e-05,
      "loss": 0.0375,
      "step": 620
    },
    {
      "epoch": 28.0,
      "learning_rate": 1.4952153110047847e-05,
      "loss": 0.0191,
      "step": 630
    },
    {
      "epoch": 28.0,
      "eval_accuracy": 0.5887850467289719,
      "eval_loss": 2.782459259033203,
      "eval_runtime": 2.7194,
      "eval_samples_per_second": 39.347,
      "eval_steps_per_second": 2.574,
      "step": 630
    },
    {
      "epoch": 28.44,
      "learning_rate": 1.4354066985645934e-05,
      "loss": 0.0057,
      "step": 640
    },
    {
      "epoch": 28.89,
      "learning_rate": 1.375598086124402e-05,
      "loss": 0.0005,
      "step": 650
    },
    {
      "epoch": 28.98,
      "eval_accuracy": 0.5887850467289719,
      "eval_loss": 3.528034210205078,
      "eval_runtime": 2.584,
      "eval_samples_per_second": 41.408,
      "eval_steps_per_second": 2.709,
      "step": 652
    },
    {
      "epoch": 29.33,
      "learning_rate": 1.3157894736842106e-05,
      "loss": 0.0136,
      "step": 660
    },
    {
      "epoch": 29.78,
      "learning_rate": 1.2559808612440191e-05,
      "loss": 0.0093,
      "step": 670
    },
    {
      "epoch": 30.0,
      "eval_accuracy": 0.6074766355140186,
      "eval_loss": 2.828961133956909,
      "eval_runtime": 2.5935,
      "eval_samples_per_second": 41.257,
      "eval_steps_per_second": 2.699,
      "step": 675
    },
    {
      "epoch": 30.22,
      "learning_rate": 1.1961722488038278e-05,
      "loss": 0.0004,
      "step": 680
    },
    {
      "epoch": 30.67,
      "learning_rate": 1.1363636363636365e-05,
      "loss": 0.0224,
      "step": 690
    },
    {
      "epoch": 30.98,
      "eval_accuracy": 0.5794392523364486,
      "eval_loss": 2.9545552730560303,
      "eval_runtime": 2.7643,
      "eval_samples_per_second": 38.707,
      "eval_steps_per_second": 2.532,
      "step": 697
    },
    {
      "epoch": 31.11,
      "learning_rate": 1.0765550239234451e-05,
      "loss": 0.0325,
      "step": 700
    },
    {
      "epoch": 31.56,
      "learning_rate": 1.0167464114832537e-05,
      "loss": 0.0102,
      "step": 710
    },
    {
      "epoch": 32.0,
      "learning_rate": 9.569377990430622e-06,
      "loss": 0.0011,
      "step": 720
    },
    {
      "epoch": 32.0,
      "eval_accuracy": 0.6074766355140186,
      "eval_loss": 3.0147745609283447,
      "eval_runtime": 2.6,
      "eval_samples_per_second": 41.155,
      "eval_steps_per_second": 2.692,
      "step": 720
    },
    {
      "epoch": 32.44,
      "learning_rate": 8.971291866028708e-06,
      "loss": 0.0014,
      "step": 730
    },
    {
      "epoch": 32.89,
      "learning_rate": 8.373205741626795e-06,
      "loss": 0.003,
      "step": 740
    },
    {
      "epoch": 32.98,
      "eval_accuracy": 0.5981308411214953,
      "eval_loss": 3.2915587425231934,
      "eval_runtime": 2.5651,
      "eval_samples_per_second": 41.714,
      "eval_steps_per_second": 2.729,
      "step": 742
    },
    {
      "epoch": 33.33,
      "learning_rate": 7.77511961722488e-06,
      "loss": 0.0005,
      "step": 750
    },
    {
      "epoch": 33.78,
      "learning_rate": 7.177033492822967e-06,
      "loss": 0.0003,
      "step": 760
    },
    {
      "epoch": 34.0,
      "eval_accuracy": 0.5981308411214953,
      "eval_loss": 3.292980909347534,
      "eval_runtime": 2.5274,
      "eval_samples_per_second": 42.337,
      "eval_steps_per_second": 2.77,
      "step": 765
    },
    {
      "epoch": 34.22,
      "learning_rate": 6.578947368421053e-06,
      "loss": 0.0003,
      "step": 770
    },
    {
      "epoch": 34.67,
      "learning_rate": 5.980861244019139e-06,
      "loss": 0.0003,
      "step": 780
    },
    {
      "epoch": 34.98,
      "eval_accuracy": 0.5887850467289719,
      "eval_loss": 3.628692626953125,
      "eval_runtime": 2.6394,
      "eval_samples_per_second": 40.539,
      "eval_steps_per_second": 2.652,
      "step": 787
    },
    {
      "epoch": 35.11,
      "learning_rate": 5.382775119617226e-06,
      "loss": 0.0029,
      "step": 790
    },
    {
      "epoch": 35.56,
      "learning_rate": 4.784688995215311e-06,
      "loss": 0.0003,
      "step": 800
    },
    {
      "epoch": 36.0,
      "learning_rate": 4.186602870813398e-06,
      "loss": 0.0002,
      "step": 810
    },
    {
      "epoch": 36.0,
      "eval_accuracy": 0.5887850467289719,
      "eval_loss": 3.6918015480041504,
      "eval_runtime": 2.7357,
      "eval_samples_per_second": 39.112,
      "eval_steps_per_second": 2.559,
      "step": 810
    },
    {
      "epoch": 36.44,
      "learning_rate": 3.5885167464114835e-06,
      "loss": 0.0003,
      "step": 820
    },
    {
      "epoch": 36.89,
      "learning_rate": 2.9904306220095695e-06,
      "loss": 0.0004,
      "step": 830
    },
    {
      "epoch": 36.98,
      "eval_accuracy": 0.5887850467289719,
      "eval_loss": 3.659736156463623,
      "eval_runtime": 2.6137,
      "eval_samples_per_second": 40.938,
      "eval_steps_per_second": 2.678,
      "step": 832
    },
    {
      "epoch": 37.33,
      "learning_rate": 2.3923444976076554e-06,
      "loss": 0.0002,
      "step": 840
    },
    {
      "epoch": 37.78,
      "learning_rate": 1.7942583732057418e-06,
      "loss": 0.0003,
      "step": 850
    },
    {
      "epoch": 38.0,
      "eval_accuracy": 0.5887850467289719,
      "eval_loss": 3.65986967086792,
      "eval_runtime": 2.5356,
      "eval_samples_per_second": 42.198,
      "eval_steps_per_second": 2.761,
      "step": 855
    },
    {
      "epoch": 38.22,
      "learning_rate": 1.1961722488038277e-06,
      "loss": 0.0022,
      "step": 860
    },
    {
      "epoch": 38.67,
      "learning_rate": 5.980861244019139e-07,
      "loss": 0.0002,
      "step": 870
    },
    {
      "epoch": 38.98,
      "eval_accuracy": 0.5887850467289719,
      "eval_loss": 3.673996925354004,
      "eval_runtime": 2.6372,
      "eval_samples_per_second": 40.573,
      "eval_steps_per_second": 2.654,
      "step": 877
    },
    {
      "epoch": 39.11,
      "learning_rate": 0.0,
      "loss": 0.0002,
      "step": 880
    },
    {
      "epoch": 39.11,
      "eval_accuracy": 0.5887850467289719,
      "eval_loss": 3.674065351486206,
      "eval_runtime": 2.9557,
      "eval_samples_per_second": 36.201,
      "eval_steps_per_second": 2.368,
      "step": 880
    },
    {
      "epoch": 39.11,
      "step": 880,
      "total_flos": 4.350399898167558e+18,
      "train_loss": 0.09111572096166624,
      "train_runtime": 1452.6155,
      "train_samples_per_second": 39.542,
      "train_steps_per_second": 0.606
    }
  ],
  "logging_steps": 10,
  "max_steps": 880,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 40,
  "save_steps": 500,
  "total_flos": 4.350399898167558e+18,
  "train_batch_size": 16,
  "trial_name": null,
  "trial_params": null
}