{
  "best_metric": 0.26857295632362366,
  "best_model_checkpoint": "./vit-mae-flysheet/checkpoint-2716",
  "epoch": 100.0,
  "global_step": 2800,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.36,
      "learning_rate": 2.3437499999999998e-07,
      "loss": 2.2918,
      "step": 10
    },
    {
      "epoch": 0.71,
      "learning_rate": 5.691964285714285e-07,
      "loss": 2.284,
      "step": 20
    },
    {
      "epoch": 1.0,
      "eval_loss": 2.281158924102783,
      "eval_runtime": 2.6825,
      "eval_samples_per_second": 115.564,
      "eval_steps_per_second": 1.864,
      "step": 28
    },
    {
      "epoch": 1.07,
      "learning_rate": 9.040178571428571e-07,
      "loss": 2.29,
      "step": 30
    },
    {
      "epoch": 1.43,
      "learning_rate": 1.2388392857142857e-06,
      "loss": 2.1474,
      "step": 40
    },
    {
      "epoch": 1.79,
      "learning_rate": 1.573660714285714e-06,
      "loss": 2.137,
      "step": 50
    },
    {
      "epoch": 2.0,
      "eval_loss": 2.0287907123565674,
      "eval_runtime": 2.9845,
      "eval_samples_per_second": 103.869,
      "eval_steps_per_second": 1.675,
      "step": 56
    },
    {
      "epoch": 2.14,
      "learning_rate": 1.9084821428571425e-06,
      "loss": 2.0038,
      "step": 60
    },
    {
      "epoch": 2.5,
      "learning_rate": 2.2433035714285713e-06,
      "loss": 1.8154,
      "step": 70
    },
    {
      "epoch": 2.86,
      "learning_rate": 2.578125e-06,
      "loss": 1.6016,
      "step": 80
    },
    {
      "epoch": 3.0,
      "eval_loss": 1.243688702583313,
      "eval_runtime": 3.1026,
      "eval_samples_per_second": 99.916,
      "eval_steps_per_second": 1.612,
      "step": 84
    },
    {
      "epoch": 3.21,
      "learning_rate": 2.9129464285714283e-06,
      "loss": 1.2624,
      "step": 90
    },
    {
      "epoch": 3.57,
      "learning_rate": 3.2477678571428566e-06,
      "loss": 0.9693,
      "step": 100
    },
    {
      "epoch": 3.93,
      "learning_rate": 3.5825892857142853e-06,
      "loss": 0.8055,
      "step": 110
    },
    {
      "epoch": 4.0,
      "eval_loss": 0.7418646812438965,
      "eval_runtime": 3.0666,
      "eval_samples_per_second": 101.089,
      "eval_steps_per_second": 1.63,
      "step": 112
    },
    {
      "epoch": 4.29,
      "learning_rate": 3.917410714285714e-06,
      "loss": 0.7145,
      "step": 120
    },
    {
      "epoch": 4.64,
      "learning_rate": 4.252232142857143e-06,
      "loss": 0.6144,
      "step": 130
    },
    {
      "epoch": 5.0,
      "learning_rate": 4.587053571428571e-06,
      "loss": 0.5304,
      "step": 140
    },
    {
      "epoch": 5.0,
      "eval_loss": 0.5150813460350037,
      "eval_runtime": 2.9639,
      "eval_samples_per_second": 104.591,
      "eval_steps_per_second": 1.687,
      "step": 140
    },
    {
      "epoch": 5.36,
      "learning_rate": 4.921874999999999e-06,
      "loss": 0.4939,
      "step": 150
    },
    {
      "epoch": 5.71,
      "learning_rate": 5.256696428571428e-06,
      "loss": 0.4873,
      "step": 160
    },
    {
      "epoch": 6.0,
      "eval_loss": 0.4884074628353119,
      "eval_runtime": 3.1297,
      "eval_samples_per_second": 99.052,
      "eval_steps_per_second": 1.598,
      "step": 168
    },
    {
      "epoch": 6.07,
      "learning_rate": 5.591517857142857e-06,
      "loss": 0.465,
      "step": 170
    },
    {
      "epoch": 6.43,
      "learning_rate": 5.926339285714285e-06,
      "loss": 0.4548,
      "step": 180
    },
    {
      "epoch": 6.79,
      "learning_rate": 6.261160714285713e-06,
      "loss": 0.442,
      "step": 190
    },
    {
      "epoch": 7.0,
      "eval_loss": 0.44406256079673767,
      "eval_runtime": 3.1182,
      "eval_samples_per_second": 99.418,
      "eval_steps_per_second": 1.604,
      "step": 196
    },
    {
      "epoch": 7.14,
      "learning_rate": 6.595982142857143e-06,
      "loss": 0.4262,
      "step": 200
    },
    {
      "epoch": 7.5,
      "learning_rate": 6.930803571428571e-06,
      "loss": 0.4275,
      "step": 210
    },
    {
      "epoch": 7.86,
      "learning_rate": 7.2656249999999996e-06,
      "loss": 0.4039,
      "step": 220
    },
    {
      "epoch": 8.0,
      "eval_loss": 0.41594529151916504,
      "eval_runtime": 3.0271,
      "eval_samples_per_second": 102.41,
      "eval_steps_per_second": 1.652,
      "step": 224
    },
    {
      "epoch": 8.21,
      "learning_rate": 7.600446428571428e-06,
      "loss": 0.3925,
      "step": 230
    },
    {
      "epoch": 8.57,
      "learning_rate": 7.935267857142856e-06,
      "loss": 0.4092,
      "step": 240
    },
    {
      "epoch": 8.93,
      "learning_rate": 8.270089285714285e-06,
      "loss": 0.3866,
      "step": 250
    },
    {
      "epoch": 9.0,
      "eval_loss": 0.39747074246406555,
      "eval_runtime": 3.1219,
      "eval_samples_per_second": 99.3,
      "eval_steps_per_second": 1.602,
      "step": 252
    },
    {
      "epoch": 9.29,
      "learning_rate": 8.604910714285714e-06,
      "loss": 0.3675,
      "step": 260
    },
    {
      "epoch": 9.64,
      "learning_rate": 8.939732142857142e-06,
      "loss": 0.3827,
      "step": 270
    },
    {
      "epoch": 10.0,
      "learning_rate": 9.274553571428571e-06,
      "loss": 0.391,
      "step": 280
    },
    {
      "epoch": 10.0,
      "eval_loss": 0.38686203956604004,
      "eval_runtime": 2.6794,
      "eval_samples_per_second": 115.697,
      "eval_steps_per_second": 1.866,
      "step": 280
    },
    {
      "epoch": 10.36,
      "learning_rate": 9.609374999999998e-06,
      "loss": 0.3701,
      "step": 290
    },
    {
      "epoch": 10.71,
      "learning_rate": 9.944196428571429e-06,
      "loss": 0.3549,
      "step": 300
    },
    {
      "epoch": 11.0,
      "eval_loss": 0.38007286190986633,
      "eval_runtime": 2.6889,
      "eval_samples_per_second": 115.289,
      "eval_steps_per_second": 1.859,
      "step": 308
    },
    {
      "epoch": 11.07,
      "learning_rate": 1.0279017857142857e-05,
      "loss": 0.3618,
      "step": 310
    },
    {
      "epoch": 11.43,
      "learning_rate": 1.0613839285714284e-05,
      "loss": 0.357,
      "step": 320
    },
    {
      "epoch": 11.79,
      "learning_rate": 1.0948660714285715e-05,
      "loss": 0.3462,
      "step": 330
    },
    {
      "epoch": 12.0,
      "eval_loss": 0.35770875215530396,
      "eval_runtime": 2.6881,
      "eval_samples_per_second": 115.324,
      "eval_steps_per_second": 1.86,
      "step": 336
    },
    {
      "epoch": 12.14,
      "learning_rate": 1.1283482142857142e-05,
      "loss": 0.3413,
      "step": 340
    },
    {
      "epoch": 12.5,
      "learning_rate": 1.161830357142857e-05,
      "loss": 0.341,
      "step": 350
    },
    {
      "epoch": 12.86,
      "learning_rate": 1.1953124999999997e-05,
      "loss": 0.3402,
      "step": 360
    },
    {
      "epoch": 13.0,
      "eval_loss": 0.3519320487976074,
      "eval_runtime": 2.6575,
      "eval_samples_per_second": 116.65,
      "eval_steps_per_second": 1.881,
      "step": 364
    },
    {
      "epoch": 13.21,
      "learning_rate": 1.2287946428571428e-05,
      "loss": 0.3281,
      "step": 370
    },
    {
      "epoch": 13.57,
      "learning_rate": 1.2622767857142857e-05,
      "loss": 0.3423,
      "step": 380
    },
    {
      "epoch": 13.93,
      "learning_rate": 1.2957589285714284e-05,
      "loss": 0.3357,
      "step": 390
    },
    {
      "epoch": 14.0,
      "eval_loss": 0.344655305147171,
      "eval_runtime": 2.6732,
      "eval_samples_per_second": 115.968,
      "eval_steps_per_second": 1.87,
      "step": 392
    },
    {
      "epoch": 14.29,
      "learning_rate": 1.3292410714285714e-05,
      "loss": 0.3254,
      "step": 400
    },
    {
      "epoch": 14.64,
      "learning_rate": 1.3627232142857141e-05,
      "loss": 0.3231,
      "step": 410
    },
    {
      "epoch": 15.0,
      "learning_rate": 1.396205357142857e-05,
      "loss": 0.3474,
      "step": 420
    },
    {
      "epoch": 15.0,
      "eval_loss": 0.33689987659454346,
      "eval_runtime": 2.6785,
      "eval_samples_per_second": 115.736,
      "eval_steps_per_second": 1.867,
      "step": 420
    },
    {
      "epoch": 15.36,
      "learning_rate": 1.4296874999999999e-05,
      "loss": 0.3117,
      "step": 430
    },
    {
      "epoch": 15.71,
      "learning_rate": 1.4631696428571427e-05,
      "loss": 0.3254,
      "step": 440
    },
    {
      "epoch": 16.0,
      "eval_loss": 0.33861759305000305,
      "eval_runtime": 2.6231,
      "eval_samples_per_second": 118.183,
      "eval_steps_per_second": 1.906,
      "step": 448
    },
    {
      "epoch": 16.07,
      "learning_rate": 1.4966517857142856e-05,
      "loss": 0.3221,
      "step": 450
    },
    {
      "epoch": 16.43,
      "learning_rate": 1.5301339285714285e-05,
      "loss": 0.3347,
      "step": 460
    },
    {
      "epoch": 16.79,
      "learning_rate": 1.5636160714285715e-05,
      "loss": 0.3033,
      "step": 470
    },
    {
      "epoch": 17.0,
      "eval_loss": 0.3294062316417694,
      "eval_runtime": 2.6872,
      "eval_samples_per_second": 115.362,
      "eval_steps_per_second": 1.861,
      "step": 476
    },
    {
      "epoch": 17.14,
      "learning_rate": 1.5970982142857142e-05,
      "loss": 0.3135,
      "step": 480
    },
    {
      "epoch": 17.5,
      "learning_rate": 1.630580357142857e-05,
      "loss": 0.3104,
      "step": 490
    },
    {
      "epoch": 17.86,
      "learning_rate": 1.6640624999999996e-05,
      "loss": 0.3047,
      "step": 500
    },
    {
      "epoch": 18.0,
      "eval_loss": 0.32739999890327454,
      "eval_runtime": 2.6527,
      "eval_samples_per_second": 116.864,
      "eval_steps_per_second": 1.885,
      "step": 504
    },
    {
      "epoch": 18.21,
      "learning_rate": 1.6975446428571427e-05,
      "loss": 0.3088,
      "step": 510
    },
    {
      "epoch": 18.57,
      "learning_rate": 1.7310267857142857e-05,
      "loss": 0.3079,
      "step": 520
    },
    {
      "epoch": 18.93,
      "learning_rate": 1.7645089285714284e-05,
      "loss": 0.3103,
      "step": 530
    },
    {
      "epoch": 19.0,
      "eval_loss": 0.3209254741668701,
      "eval_runtime": 2.6351,
      "eval_samples_per_second": 117.643,
      "eval_steps_per_second": 1.897,
      "step": 532
    },
    {
      "epoch": 19.29,
      "learning_rate": 1.7979910714285714e-05,
      "loss": 0.3004,
      "step": 540
    },
    {
      "epoch": 19.64,
      "learning_rate": 1.831473214285714e-05,
      "loss": 0.3037,
      "step": 550
    },
    {
      "epoch": 20.0,
      "learning_rate": 1.864955357142857e-05,
      "loss": 0.3067,
      "step": 560
    },
    {
      "epoch": 20.0,
      "eval_loss": 0.3185921013355255,
      "eval_runtime": 2.6798,
      "eval_samples_per_second": 115.68,
      "eval_steps_per_second": 1.866,
      "step": 560
    },
    {
      "epoch": 20.36,
      "learning_rate": 1.8984375e-05,
      "loss": 0.3007,
      "step": 570
    },
    {
      "epoch": 20.71,
      "learning_rate": 1.9319196428571426e-05,
      "loss": 0.2959,
      "step": 580
    },
    {
      "epoch": 21.0,
      "eval_loss": 0.31901559233665466,
      "eval_runtime": 2.6847,
      "eval_samples_per_second": 115.468,
      "eval_steps_per_second": 1.862,
      "step": 588
    },
    {
      "epoch": 21.07,
      "learning_rate": 1.9654017857142856e-05,
      "loss": 0.3031,
      "step": 590
    },
    {
      "epoch": 21.43,
      "learning_rate": 1.9988839285714283e-05,
      "loss": 0.3107,
      "step": 600
    },
    {
      "epoch": 21.79,
      "learning_rate": 2.032366071428571e-05,
      "loss": 0.2899,
      "step": 610
    },
    {
      "epoch": 22.0,
      "eval_loss": 0.3146977126598358,
      "eval_runtime": 2.6772,
      "eval_samples_per_second": 115.792,
      "eval_steps_per_second": 1.868,
      "step": 616
    },
    {
      "epoch": 22.14,
      "learning_rate": 2.0658482142857144e-05,
      "loss": 0.2868,
      "step": 620
    },
    {
      "epoch": 22.5,
      "learning_rate": 2.099330357142857e-05,
      "loss": 0.3018,
      "step": 630
    },
    {
      "epoch": 22.86,
      "learning_rate": 2.1328125e-05,
      "loss": 0.2872,
      "step": 640
    },
    {
      "epoch": 23.0,
      "eval_loss": 0.30817940831184387,
      "eval_runtime": 2.6553,
      "eval_samples_per_second": 116.748,
      "eval_steps_per_second": 1.883,
      "step": 644
    },
    {
      "epoch": 23.21,
      "learning_rate": 2.1662946428571425e-05,
      "loss": 0.3033,
      "step": 650
    },
    {
      "epoch": 23.57,
      "learning_rate": 2.1997767857142856e-05,
      "loss": 0.2946,
      "step": 660
    },
    {
      "epoch": 23.93,
      "learning_rate": 2.2332589285714283e-05,
      "loss": 0.2956,
      "step": 670
    },
    {
      "epoch": 24.0,
      "eval_loss": 0.3069721758365631,
      "eval_runtime": 2.6278,
      "eval_samples_per_second": 117.971,
      "eval_steps_per_second": 1.903,
      "step": 672
    },
    {
      "epoch": 24.29,
      "learning_rate": 2.266741071428571e-05,
      "loss": 0.2754,
      "step": 680
    },
    {
      "epoch": 24.64,
      "learning_rate": 2.3002232142857144e-05,
      "loss": 0.3039,
      "step": 690
    },
    {
      "epoch": 25.0,
      "learning_rate": 2.333705357142857e-05,
      "loss": 0.2865,
      "step": 700
    },
    {
      "epoch": 25.0,
      "eval_loss": 0.30717733502388,
      "eval_runtime": 2.6953,
      "eval_samples_per_second": 115.015,
      "eval_steps_per_second": 1.855,
      "step": 700
    },
    {
      "epoch": 25.36,
      "learning_rate": 2.3671874999999998e-05,
      "loss": 0.2865,
      "step": 710
    },
    {
      "epoch": 25.71,
      "learning_rate": 2.4006696428571425e-05,
      "loss": 0.2947,
      "step": 720
    },
    {
      "epoch": 26.0,
      "eval_loss": 0.3071895241737366,
      "eval_runtime": 2.6411,
      "eval_samples_per_second": 117.376,
      "eval_steps_per_second": 1.893,
      "step": 728
    },
    {
      "epoch": 26.07,
      "learning_rate": 2.4341517857142855e-05,
      "loss": 0.2831,
      "step": 730
    },
    {
      "epoch": 26.43,
      "learning_rate": 2.4676339285714282e-05,
      "loss": 0.3008,
      "step": 740
    },
    {
      "epoch": 26.79,
      "learning_rate": 2.501116071428571e-05,
      "loss": 0.2811,
      "step": 750
    },
    {
      "epoch": 27.0,
      "eval_loss": 0.3130524158477783,
      "eval_runtime": 2.683,
      "eval_samples_per_second": 115.544,
      "eval_steps_per_second": 1.864,
      "step": 756
    },
    {
      "epoch": 27.14,
      "learning_rate": 2.5345982142857143e-05,
      "loss": 0.2814,
      "step": 760
    },
    {
      "epoch": 27.5,
      "learning_rate": 2.568080357142857e-05,
      "loss": 0.2872,
      "step": 770
    },
    {
      "epoch": 27.86,
      "learning_rate": 2.6015624999999997e-05,
      "loss": 0.2935,
      "step": 780
    },
    {
      "epoch": 28.0,
      "eval_loss": 0.3069266080856323,
      "eval_runtime": 2.6568,
      "eval_samples_per_second": 116.681,
      "eval_steps_per_second": 1.882,
      "step": 784
    },
    {
      "epoch": 28.21,
      "learning_rate": 2.6350446428571424e-05,
      "loss": 0.2968,
      "step": 790
    },
    {
      "epoch": 28.57,
      "learning_rate": 2.6685267857142855e-05,
      "loss": 0.284,
      "step": 800
    },
    {
      "epoch": 28.93,
      "learning_rate": 2.702008928571428e-05,
      "loss": 0.2814,
      "step": 810
    },
    {
      "epoch": 29.0,
      "eval_loss": 0.30426502227783203,
      "eval_runtime": 2.6601,
      "eval_samples_per_second": 116.537,
      "eval_steps_per_second": 1.88,
      "step": 812
    },
    {
      "epoch": 29.29,
      "learning_rate": 2.7354910714285712e-05,
      "loss": 0.2963,
      "step": 820
    },
    {
      "epoch": 29.64,
      "learning_rate": 2.7689732142857142e-05,
      "loss": 0.2891,
      "step": 830
    },
    {
      "epoch": 30.0,
      "learning_rate": 2.802455357142857e-05,
      "loss": 0.2753,
      "step": 840
    },
    {
      "epoch": 30.0,
      "eval_loss": 0.2983907461166382,
      "eval_runtime": 2.7173,
      "eval_samples_per_second": 114.083,
      "eval_steps_per_second": 1.84,
      "step": 840
    },
    {
      "epoch": 30.36,
      "learning_rate": 2.8359374999999996e-05,
      "loss": 0.2894,
      "step": 850
    },
    {
      "epoch": 30.71,
      "learning_rate": 2.8694196428571423e-05,
      "loss": 0.2823,
      "step": 860
    },
    {
      "epoch": 31.0,
      "eval_loss": 0.2994972765445709,
      "eval_runtime": 2.6498,
      "eval_samples_per_second": 116.99,
      "eval_steps_per_second": 1.887,
      "step": 868
    },
    {
      "epoch": 31.07,
      "learning_rate": 2.9029017857142854e-05,
      "loss": 0.2792,
      "step": 870
    },
    {
      "epoch": 31.43,
      "learning_rate": 2.9363839285714284e-05,
      "loss": 0.2755,
      "step": 880
    },
    {
      "epoch": 31.79,
      "learning_rate": 2.969866071428571e-05,
      "loss": 0.2962,
      "step": 890
    },
    {
      "epoch": 32.0,
      "eval_loss": 0.30121392011642456,
      "eval_runtime": 2.6677,
      "eval_samples_per_second": 116.206,
      "eval_steps_per_second": 1.874,
      "step": 896
    },
    {
      "epoch": 32.14,
      "learning_rate": 3.0033482142857142e-05,
      "loss": 0.2835,
      "step": 900
    },
    {
      "epoch": 32.5,
      "learning_rate": 3.036830357142857e-05,
      "loss": 0.2724,
      "step": 910
    },
    {
      "epoch": 32.86,
      "learning_rate": 3.0703124999999996e-05,
      "loss": 0.2869,
      "step": 920
    },
    {
      "epoch": 33.0,
      "eval_loss": 0.30497127771377563,
      "eval_runtime": 2.6483,
      "eval_samples_per_second": 117.058,
      "eval_steps_per_second": 1.888,
      "step": 924
    },
    {
      "epoch": 33.21,
      "learning_rate": 3.1037946428571426e-05,
      "loss": 0.2812,
      "step": 930
    },
    {
      "epoch": 33.57,
      "learning_rate": 3.137276785714286e-05,
      "loss": 0.2855,
      "step": 940
    },
    {
      "epoch": 33.93,
      "learning_rate": 3.170758928571428e-05,
      "loss": 0.2833,
      "step": 950
    },
    {
      "epoch": 34.0,
      "eval_loss": 0.29598191380500793,
      "eval_runtime": 2.6636,
      "eval_samples_per_second": 116.384,
      "eval_steps_per_second": 1.877,
      "step": 952
    },
    {
      "epoch": 34.29,
      "learning_rate": 3.204241071428571e-05,
      "loss": 0.2771,
      "step": 960
    },
    {
      "epoch": 34.64,
      "learning_rate": 3.237723214285714e-05,
      "loss": 0.2753,
      "step": 970
    },
    {
      "epoch": 35.0,
      "learning_rate": 3.271205357142857e-05,
      "loss": 0.2892,
      "step": 980
    },
    {
      "epoch": 35.0,
      "eval_loss": 0.3039288818836212,
      "eval_runtime": 2.6706,
      "eval_samples_per_second": 116.078,
      "eval_steps_per_second": 1.872,
      "step": 980
    },
    {
      "epoch": 35.36,
      "learning_rate": 3.3046874999999995e-05,
      "loss": 0.2915,
      "step": 990
    },
    {
      "epoch": 35.71,
      "learning_rate": 3.3381696428571426e-05,
      "loss": 0.2764,
      "step": 1000
    },
    {
      "epoch": 36.0,
      "eval_loss": 0.301042377948761,
      "eval_runtime": 2.6356,
      "eval_samples_per_second": 117.622,
      "eval_steps_per_second": 1.897,
      "step": 1008
    },
    {
      "epoch": 36.07,
      "learning_rate": 3.3716517857142856e-05,
      "loss": 0.2854,
      "step": 1010
    },
    {
      "epoch": 36.43,
      "learning_rate": 3.405133928571428e-05,
      "loss": 0.2874,
      "step": 1020
    },
    {
      "epoch": 36.79,
      "learning_rate": 3.438616071428571e-05,
      "loss": 0.2807,
      "step": 1030
    },
    {
      "epoch": 37.0,
      "eval_loss": 0.2998359799385071,
      "eval_runtime": 2.6115,
      "eval_samples_per_second": 118.706,
      "eval_steps_per_second": 1.915,
      "step": 1036
    },
    {
      "epoch": 37.14,
      "learning_rate": 3.472098214285714e-05,
      "loss": 0.2629,
      "step": 1040
    },
    {
      "epoch": 37.5,
      "learning_rate": 3.505580357142857e-05,
      "loss": 0.2759,
      "step": 1050
    },
    {
      "epoch": 37.86,
      "learning_rate": 3.5390624999999995e-05,
      "loss": 0.2843,
      "step": 1060
    },
    {
      "epoch": 38.0,
      "eval_loss": 0.2989421784877777,
      "eval_runtime": 2.6315,
      "eval_samples_per_second": 117.805,
      "eval_steps_per_second": 1.9,
      "step": 1064
    },
    {
      "epoch": 38.21,
      "learning_rate": 3.5725446428571425e-05,
      "loss": 0.284,
      "step": 1070
    },
    {
      "epoch": 38.57,
      "learning_rate": 3.6060267857142855e-05,
      "loss": 0.2786,
      "step": 1080
    },
    {
      "epoch": 38.93,
      "learning_rate": 3.639508928571428e-05,
      "loss": 0.2808,
      "step": 1090
    },
    {
      "epoch": 39.0,
      "eval_loss": 0.2969764173030853,
      "eval_runtime": 2.6345,
      "eval_samples_per_second": 117.671,
      "eval_steps_per_second": 1.898,
      "step": 1092
    },
    {
      "epoch": 39.29,
      "learning_rate": 3.672991071428571e-05,
      "loss": 0.2748,
      "step": 1100
    },
    {
      "epoch": 39.64,
      "learning_rate": 3.706473214285714e-05,
      "loss": 0.2751,
      "step": 1110
    },
    {
      "epoch": 40.0,
      "learning_rate": 3.739955357142857e-05,
      "loss": 0.2862,
      "step": 1120
    },
    {
      "epoch": 40.0,
      "eval_loss": 0.29402607679367065,
      "eval_runtime": 2.6346,
      "eval_samples_per_second": 117.666,
      "eval_steps_per_second": 1.898,
      "step": 1120
    },
    {
      "epoch": 40.36,
      "learning_rate": 3.7499989987931416e-05,
      "loss": 0.2951,
      "step": 1130
    },
    {
      "epoch": 40.71,
      "learning_rate": 3.7499940949253886e-05,
      "loss": 0.2601,
      "step": 1140
    },
    {
      "epoch": 41.0,
      "eval_loss": 0.2952007055282593,
      "eval_runtime": 2.6361,
      "eval_samples_per_second": 117.597,
      "eval_steps_per_second": 1.897,
      "step": 1148
    },
    {
      "epoch": 41.07,
      "learning_rate": 3.749985104512278e-05,
      "loss": 0.2796,
      "step": 1150
    },
    {
      "epoch": 41.43,
      "learning_rate": 3.7499720275734064e-05,
      "loss": 0.2759,
      "step": 1160
    },
    {
      "epoch": 41.79,
      "learning_rate": 3.7499548641372725e-05,
      "loss": 0.2742,
      "step": 1170
    },
    {
      "epoch": 42.0,
      "eval_loss": 0.2939775884151459,
      "eval_runtime": 2.6378,
      "eval_samples_per_second": 117.522,
      "eval_steps_per_second": 1.896,
      "step": 1176
    },
    {
      "epoch": 42.14,
      "learning_rate": 3.7499336142412856e-05,
      "loss": 0.2757,
      "step": 1180
    },
    {
      "epoch": 42.5,
      "learning_rate": 3.749908277931758e-05,
      "loss": 0.2801,
      "step": 1190
    },
    {
      "epoch": 42.86,
      "learning_rate": 3.749878855263912e-05,
      "loss": 0.2791,
      "step": 1200
    },
    {
      "epoch": 43.0,
      "eval_loss": 0.29970163106918335,
      "eval_runtime": 2.6144,
      "eval_samples_per_second": 118.572,
      "eval_steps_per_second": 1.912,
      "step": 1204
    },
    {
      "epoch": 43.21,
      "learning_rate": 3.7498453463018734e-05,
      "loss": 0.2744,
      "step": 1210
    },
    {
      "epoch": 43.57,
      "learning_rate": 3.749807751118675e-05,
      "loss": 0.2793,
      "step": 1220
    },
    {
      "epoch": 43.93,
      "learning_rate": 3.749766069796256e-05,
      "loss": 0.2759,
      "step": 1230
    },
    {
      "epoch": 44.0,
      "eval_loss": 0.2951151728630066,
      "eval_runtime": 2.6705,
      "eval_samples_per_second": 116.083,
      "eval_steps_per_second": 1.872,
      "step": 1232
    },
    {
      "epoch": 44.29,
      "learning_rate": 3.749720302425459e-05,
      "loss": 0.2802,
      "step": 1240
    },
    {
      "epoch": 44.64,
      "learning_rate": 3.749670449106036e-05,
      "loss": 0.268,
      "step": 1250
    },
    {
      "epoch": 45.0,
      "learning_rate": 3.749616509946641e-05,
      "loss": 0.2819,
      "step": 1260
    },
    {
      "epoch": 45.0,
      "eval_loss": 0.2895747125148773,
      "eval_runtime": 2.6558,
      "eval_samples_per_second": 116.726,
      "eval_steps_per_second": 1.883,
      "step": 1260
    },
    {
      "epoch": 45.36,
      "learning_rate": 3.7495584850648345e-05,
      "loss": 0.2697,
      "step": 1270
    },
    {
      "epoch": 45.71,
      "learning_rate": 3.749496374587082e-05,
      "loss": 0.287,
      "step": 1280
    },
    {
      "epoch": 46.0,
      "eval_loss": 0.2938202917575836,
      "eval_runtime": 2.6182,
      "eval_samples_per_second": 118.403,
      "eval_steps_per_second": 1.91,
      "step": 1288
    },
    {
      "epoch": 46.07,
      "learning_rate": 3.749430178648752e-05,
      "loss": 0.2736,
      "step": 1290
    },
    {
      "epoch": 46.43,
      "learning_rate": 3.7493598973941194e-05,
      "loss": 0.2817,
      "step": 1300
    },
    {
      "epoch": 46.79,
      "learning_rate": 3.749285530976362e-05,
      "loss": 0.2711,
      "step": 1310
    },
    {
      "epoch": 47.0,
      "eval_loss": 0.2973059117794037,
      "eval_runtime": 2.6436,
      "eval_samples_per_second": 117.264,
      "eval_steps_per_second": 1.891,
      "step": 1316
    },
    {
      "epoch": 47.14,
      "learning_rate": 3.749207079557561e-05,
      "loss": 0.2761,
      "step": 1320
    },
    {
      "epoch": 47.5,
      "learning_rate": 3.749124543308701e-05,
      "loss": 0.2732,
      "step": 1330
    },
    {
      "epoch": 47.86,
      "learning_rate": 3.74903792240967e-05,
      "loss": 0.2782,
      "step": 1340
    },
    {
      "epoch": 48.0,
      "eval_loss": 0.2946121096611023,
      "eval_runtime": 2.6239,
      "eval_samples_per_second": 118.144,
      "eval_steps_per_second": 1.906,
      "step": 1344
    },
    {
      "epoch": 48.21,
      "learning_rate": 3.748947217049258e-05,
      "loss": 0.2682,
      "step": 1350
    },
    {
      "epoch": 48.57,
      "learning_rate": 3.7488524274251566e-05,
      "loss": 0.2765,
      "step": 1360
    },
    {
      "epoch": 48.93,
      "learning_rate": 3.748753553743959e-05,
      "loss": 0.2674,
      "step": 1370
    },
    {
      "epoch": 49.0,
      "eval_loss": 0.29130813479423523,
      "eval_runtime": 2.6539,
      "eval_samples_per_second": 116.81,
      "eval_steps_per_second": 1.884,
      "step": 1372
    },
    {
      "epoch": 49.29,
      "learning_rate": 3.748650596221162e-05,
      "loss": 0.2833,
      "step": 1380
    },
    {
      "epoch": 49.64,
      "learning_rate": 3.74854355508116e-05,
      "loss": 0.2701,
      "step": 1390
    },
    {
      "epoch": 50.0,
      "learning_rate": 3.748432430557249e-05,
      "loss": 0.268,
      "step": 1400
    },
    {
      "epoch": 50.0,
      "eval_loss": 0.29444822669029236,
      "eval_runtime": 2.6587,
      "eval_samples_per_second": 116.597,
      "eval_steps_per_second": 1.881,
      "step": 1400
    },
    {
      "epoch": 50.36,
      "learning_rate": 3.748317222891625e-05,
      "loss": 0.2743,
      "step": 1410
    },
    {
      "epoch": 50.71,
      "learning_rate": 3.748197932335383e-05,
      "loss": 0.2624,
      "step": 1420
    },
    {
      "epoch": 51.0,
      "eval_loss": 0.29401645064353943,
      "eval_runtime": 2.6105,
      "eval_samples_per_second": 118.751,
      "eval_steps_per_second": 1.915,
      "step": 1428
    },
    {
      "epoch": 51.07,
      "learning_rate": 3.748074559148516e-05,
      "loss": 0.2865,
      "step": 1430
    },
    {
      "epoch": 51.43,
      "learning_rate": 3.747947103599916e-05,
      "loss": 0.2717,
      "step": 1440
    },
    {
      "epoch": 51.79,
      "learning_rate": 3.747815565967371e-05,
      "loss": 0.2842,
      "step": 1450
    },
    {
      "epoch": 52.0,
      "eval_loss": 0.29776862263679504,
      "eval_runtime": 2.6164,
      "eval_samples_per_second": 118.483,
      "eval_steps_per_second": 1.911,
      "step": 1456
    },
    {
      "epoch": 52.14,
      "learning_rate": 3.747679946537569e-05,
      "loss": 0.2716,
      "step": 1460
    },
    {
      "epoch": 52.5,
      "learning_rate": 3.7475402456060905e-05,
      "loss": 0.2776,
      "step": 1470
    },
    {
      "epoch": 52.86,
      "learning_rate": 3.7473964634774137e-05,
      "loss": 0.2753,
      "step": 1480
    },
    {
      "epoch": 53.0,
      "eval_loss": 0.29514825344085693,
      "eval_runtime": 2.6208,
      "eval_samples_per_second": 118.283,
      "eval_steps_per_second": 1.908,
      "step": 1484
    },
    {
      "epoch": 53.21,
      "learning_rate": 3.747248600464912e-05,
      "loss": 0.268,
      "step": 1490
    },
    {
      "epoch": 53.57,
      "learning_rate": 3.747096656890852e-05,
      "loss": 0.2734,
      "step": 1500
    },
    {
      "epoch": 53.93,
      "learning_rate": 3.746940633086394e-05,
      "loss": 0.2733,
      "step": 1510
    },
    {
      "epoch": 54.0,
      "eval_loss": 0.28796395659446716,
      "eval_runtime": 2.6647,
      "eval_samples_per_second": 116.337,
      "eval_steps_per_second": 1.876,
      "step": 1512
    },
    {
      "epoch": 54.29,
      "learning_rate": 3.746780529391593e-05,
      "loss": 0.2638,
      "step": 1520
    },
    {
      "epoch": 54.64,
      "learning_rate": 3.7466163461553934e-05,
      "loss": 0.274,
      "step": 1530
    },
    {
      "epoch": 55.0,
      "learning_rate": 3.746448083735632e-05,
      "loss": 0.2782,
      "step": 1540
    },
    {
      "epoch": 55.0,
      "eval_loss": 0.29691949486732483,
      "eval_runtime": 2.6296,
      "eval_samples_per_second": 117.889,
      "eval_steps_per_second": 1.901,
      "step": 1540
    },
    {
      "epoch": 55.36,
      "learning_rate": 3.746275742499037e-05,
      "loss": 0.2749,
      "step": 1550
    },
    {
      "epoch": 55.71,
      "learning_rate": 3.746099322821226e-05,
      "loss": 0.2789,
      "step": 1560
    },
    {
      "epoch": 56.0,
      "eval_loss": 0.29188984632492065,
      "eval_runtime": 2.6147,
      "eval_samples_per_second": 118.559,
      "eval_steps_per_second": 1.912,
      "step": 1568
    },
    {
      "epoch": 56.07,
      "learning_rate": 3.7459188250867045e-05,
      "loss": 0.2537,
      "step": 1570
    },
    {
      "epoch": 56.43,
      "learning_rate": 3.7457342496888676e-05,
      "loss": 0.2718,
      "step": 1580
    },
    {
      "epoch": 56.79,
      "learning_rate": 3.745545597029996e-05,
      "loss": 0.2815,
      "step": 1590
    },
    {
      "epoch": 57.0,
      "eval_loss": 0.291569322347641,
      "eval_runtime": 2.6148,
      "eval_samples_per_second": 118.555,
      "eval_steps_per_second": 1.912,
      "step": 1596
    },
    {
      "epoch": 57.14,
      "learning_rate": 3.7453528675212585e-05,
      "loss": 0.2663,
      "step": 1600
    },
    {
      "epoch": 57.5,
      "learning_rate": 3.745156061582709e-05,
      "loss": 0.2785,
      "step": 1610
    },
    {
      "epoch": 57.86,
      "learning_rate": 3.744955179643285e-05,
      "loss": 0.2629,
      "step": 1620
    },
    {
      "epoch": 58.0,
      "eval_loss": 0.29474544525146484,
      "eval_runtime": 2.6396,
      "eval_samples_per_second": 117.441,
      "eval_steps_per_second": 1.894,
      "step": 1624
    },
    {
      "epoch": 58.21,
      "learning_rate": 3.744750222140808e-05,
      "loss": 0.2682,
      "step": 1630
    },
    {
      "epoch": 58.57,
      "learning_rate": 3.744541189521983e-05,
      "loss": 0.2666,
      "step": 1640
    },
    {
      "epoch": 58.93,
      "learning_rate": 3.7443280822423956e-05,
      "loss": 0.2716,
      "step": 1650
    },
    {
      "epoch": 59.0,
      "eval_loss": 0.28275418281555176,
      "eval_runtime": 2.6524,
      "eval_samples_per_second": 116.876,
      "eval_steps_per_second": 1.885,
      "step": 1652
    },
    {
      "epoch": 59.29,
      "learning_rate": 3.744110900766512e-05,
      "loss": 0.2775,
      "step": 1660
    },
    {
      "epoch": 59.64,
      "learning_rate": 3.7438896455676804e-05,
      "loss": 0.2767,
      "step": 1670
    },
    {
      "epoch": 60.0,
      "learning_rate": 3.7436643171281254e-05,
      "loss": 0.2623,
      "step": 1680
    },
    {
      "epoch": 60.0,
      "eval_loss": 0.29241943359375,
      "eval_runtime": 2.6403,
      "eval_samples_per_second": 117.412,
      "eval_steps_per_second": 1.894,
      "step": 1680
    },
    {
      "epoch": 60.36,
      "learning_rate": 1.3997036682248361e-05,
      "loss": 0.2725,
      "step": 1690
    },
    {
      "epoch": 60.71,
      "learning_rate": 1.3783159273580512e-05,
      "loss": 0.2773,
      "step": 1700
    },
    {
      "epoch": 61.0,
      "eval_loss": 0.27645179629325867,
      "eval_runtime": 2.6338,
      "eval_samples_per_second": 117.701,
      "eval_steps_per_second": 1.898,
      "step": 1708
    },
    {
      "epoch": 61.07,
      "learning_rate": 1.3569974670940111e-05,
      "loss": 0.2674,
      "step": 1710
    },
    {
      "epoch": 61.43,
      "learning_rate": 1.3357512610649675e-05,
      "loss": 0.2761,
      "step": 1720
    },
    {
      "epoch": 61.79,
      "learning_rate": 1.3145802728246982e-05,
      "loss": 0.268,
      "step": 1730
    },
    {
      "epoch": 62.0,
      "eval_loss": 0.2753865420818329,
      "eval_runtime": 3.0543,
      "eval_samples_per_second": 101.496,
      "eval_steps_per_second": 1.637,
      "step": 1736
    },
    {
      "epoch": 62.14,
      "learning_rate": 1.2934874554351332e-05,
      "loss": 0.2724,
      "step": 1740
    },
    {
      "epoch": 62.5,
      "learning_rate": 1.2724757510544413e-05,
      "loss": 0.2694,
      "step": 1750
    },
    {
      "epoch": 62.86,
      "learning_rate": 1.2515480905266435e-05,
      "loss": 0.2839,
      "step": 1760
    },
    {
      "epoch": 63.0,
      "eval_loss": 0.2744416296482086,
      "eval_runtime": 2.9673,
      "eval_samples_per_second": 104.471,
      "eval_steps_per_second": 1.685,
      "step": 1764
    },
    {
      "epoch": 63.21,
      "learning_rate": 1.2307073929727975e-05,
      "loss": 0.2731,
      "step": 1770
    },
    {
      "epoch": 63.57,
      "learning_rate": 1.2099565653838216e-05,
      "loss": 0.277,
      "step": 1780
    },
    {
      "epoch": 63.93,
      "learning_rate": 1.1892985022150073e-05,
      "loss": 0.2684,
      "step": 1790
    },
    {
      "epoch": 64.0,
      "eval_loss": 0.2744133472442627,
      "eval_runtime": 3.1325,
      "eval_samples_per_second": 98.963,
      "eval_steps_per_second": 1.596,
      "step": 1792
    },
    {
      "epoch": 64.29,
      "learning_rate": 1.1687360849822864e-05,
      "loss": 0.2694,
      "step": 1800
    },
    {
      "epoch": 64.64,
      "learning_rate": 1.1482721818602949e-05,
      "loss": 0.2639,
      "step": 1810
    },
    {
      "epoch": 65.0,
      "learning_rate": 1.1279096472823067e-05,
      "loss": 0.2865,
      "step": 1820
    },
    {
      "epoch": 65.0,
      "eval_loss": 0.2716241776943207,
      "eval_runtime": 2.9389,
      "eval_samples_per_second": 105.482,
      "eval_steps_per_second": 1.701,
      "step": 1820
    },
    {
      "epoch": 65.36,
      "learning_rate": 1.1076513215420738e-05,
      "loss": 0.2651,
      "step": 1830
    },
    {
      "epoch": 65.71,
      "learning_rate": 1.0875000303976493e-05,
      "loss": 0.2845,
      "step": 1840
    },
    {
      "epoch": 66.0,
      "eval_loss": 0.2768835127353668,
      "eval_runtime": 2.9046,
      "eval_samples_per_second": 106.726,
      "eval_steps_per_second": 1.721,
      "step": 1848
    },
    {
      "epoch": 66.07,
      "learning_rate": 1.0674585846772314e-05,
      "loss": 0.2768,
      "step": 1850
    },
    {
      "epoch": 66.43,
      "learning_rate": 1.04752977988709e-05,
      "loss": 0.2785,
      "step": 1860
    },
    {
      "epoch": 66.79,
      "learning_rate": 1.0277163958216333e-05,
      "loss": 0.2663,
      "step": 1870
    },
    {
      "epoch": 67.0,
      "eval_loss": 0.2754151523113251,
      "eval_runtime": 3.1034,
      "eval_samples_per_second": 99.891,
      "eval_steps_per_second": 1.611,
      "step": 1876
    },
    {
      "epoch": 67.14,
      "learning_rate": 1.0080211961756649e-05,
      "loss": 0.2632,
      "step": 1880
    },
    {
      "epoch": 67.5,
      "learning_rate": 9.884469281588848e-06,
      "loss": 0.2832,
      "step": 1890
    },
    {
      "epoch": 67.86,
      "learning_rate": 9.68996322112695e-06,
      "loss": 0.269,
      "step": 1900
    },
    {
      "epoch": 68.0,
      "eval_loss": 0.2737092077732086,
      "eval_runtime": 3.0377,
      "eval_samples_per_second": 102.051,
      "eval_steps_per_second": 1.646,
      "step": 1904
    },
    {
      "epoch": 68.21,
      "learning_rate": 9.496720911293496e-06,
      "loss": 0.2753,
      "step": 1910
    },
    {
      "epoch": 68.57,
      "learning_rate": 9.304769306735192e-06,
      "loss": 0.2793,
      "step": 1920
    },
    {
      "epoch": 68.93,
      "learning_rate": 9.114135182063088e-06,
      "loss": 0.2681,
      "step": 1930
    },
    {
      "epoch": 69.0,
      "eval_loss": 0.269718736410141,
      "eval_runtime": 3.08,
      "eval_samples_per_second": 100.648,
      "eval_steps_per_second": 1.623,
      "step": 1932
    },
    {
      "epoch": 69.29,
      "learning_rate": 8.924845128117907e-06,
      "loss": 0.2763,
      "step": 1940
    },
    {
      "epoch": 69.64,
      "learning_rate": 8.736925548260963e-06,
      "loss": 0.2679,
      "step": 1950
    },
    {
      "epoch": 70.0,
      "learning_rate": 8.550402654691271e-06,
      "loss": 0.2748,
      "step": 1960
    },
    {
      "epoch": 70.0,
      "eval_loss": 0.2778591513633728,
      "eval_runtime": 2.6486,
      "eval_samples_per_second": 117.044,
      "eval_steps_per_second": 1.888,
      "step": 1960
    },
    {
      "epoch": 70.36,
      "learning_rate": 8.365302464789306e-06,
      "loss": 0.2758,
      "step": 1970
    },
    {
      "epoch": 70.71,
      "learning_rate": 8.181650797487918e-06,
      "loss": 0.2769,
      "step": 1980
    },
    {
      "epoch": 71.0,
      "eval_loss": 0.272790789604187,
      "eval_runtime": 2.6447,
      "eval_samples_per_second": 117.214,
      "eval_steps_per_second": 1.891,
      "step": 1988
    },
    {
      "epoch": 71.07,
      "learning_rate": 7.999473269670997e-06,
      "loss": 0.2662,
      "step": 1990
    },
    {
      "epoch": 71.43,
      "learning_rate": 7.818795292600176e-06,
      "loss": 0.2549,
      "step": 2000
    },
    {
      "epoch": 71.79,
      "learning_rate": 7.639642068370416e-06,
      "loss": 0.2805,
      "step": 2010
    },
    {
      "epoch": 72.0,
      "eval_loss": 0.27293434739112854,
      "eval_runtime": 2.6471,
      "eval_samples_per_second": 117.111,
      "eval_steps_per_second": 1.889,
      "step": 2016
    },
    {
      "epoch": 72.14,
      "learning_rate": 7.462038586394576e-06,
      "loss": 0.2852,
      "step": 2020
    },
    {
      "epoch": 72.5,
      "learning_rate": 7.286009619917803e-06,
      "loss": 0.2597,
      "step": 2030
    },
    {
      "epoch": 72.86,
      "learning_rate": 7.1115797225619155e-06,
      "loss": 0.2771,
      "step": 2040
    },
    {
      "epoch": 73.0,
      "eval_loss": 0.272775262594223,
      "eval_runtime": 2.6711,
      "eval_samples_per_second": 116.058,
      "eval_steps_per_second": 1.872,
      "step": 2044
    },
    {
      "epoch": 73.21,
      "learning_rate": 6.938773224900619e-06,
      "loss": 0.2752,
      "step": 2050
    },
    {
      "epoch": 73.57,
      "learning_rate": 6.7676142310656305e-06,
      "loss": 0.2646,
      "step": 2060
    },
    {
      "epoch": 73.93,
      "learning_rate": 6.598126615384559e-06,
      "loss": 0.2717,
      "step": 2070
    },
    {
      "epoch": 74.0,
      "eval_loss": 0.27492260932922363,
      "eval_runtime": 2.6449,
      "eval_samples_per_second": 117.205,
      "eval_steps_per_second": 1.89,
      "step": 2072
    },
    {
      "epoch": 74.29,
      "learning_rate": 6.430334019050682e-06,
      "loss": 0.2793,
      "step": 2080
    },
    {
      "epoch": 74.64,
      "learning_rate": 6.264259846825417e-06,
      "loss": 0.2744,
      "step": 2090
    },
    {
      "epoch": 75.0,
      "learning_rate": 6.099927263773606e-06,
      "loss": 0.267,
      "step": 2100
    },
    {
      "epoch": 75.0,
      "eval_loss": 0.273212730884552,
      "eval_runtime": 2.6667,
      "eval_samples_per_second": 116.249,
      "eval_steps_per_second": 1.875,
      "step": 2100
    },
    {
      "epoch": 75.36,
      "learning_rate": 5.9373591920323426e-06,
      "loss": 0.2673,
      "step": 2110
    },
    {
      "epoch": 75.71,
      "learning_rate": 5.776578307613639e-06,
      "loss": 0.2812,
      "step": 2120
    },
    {
      "epoch": 76.0,
      "eval_loss": 0.27429237961769104,
      "eval_runtime": 2.66,
      "eval_samples_per_second": 116.541,
      "eval_steps_per_second": 1.88,
      "step": 2128
    },
    {
      "epoch": 76.07,
      "learning_rate": 5.6176070372414275e-06,
      "loss": 0.2674,
      "step": 2130
    },
    {
      "epoch": 76.43,
      "learning_rate": 5.460467555223369e-06,
      "loss": 0.2696,
      "step": 2140
    },
    {
      "epoch": 76.79,
      "learning_rate": 5.305181780357802e-06,
      "loss": 0.2749,
      "step": 2150
    },
    {
      "epoch": 77.0,
      "eval_loss": 0.2738768756389618,
      "eval_runtime": 2.6486,
      "eval_samples_per_second": 117.043,
      "eval_steps_per_second": 1.888,
      "step": 2156
    },
    {
      "epoch": 77.14,
      "learning_rate": 5.1517713728764e-06,
      "loss": 0.2837,
      "step": 2160
    },
    {
      "epoch": 77.5,
      "learning_rate": 5.000257731422843e-06,
      "loss": 0.2606,
      "step": 2170
    },
    {
      "epoch": 77.86,
      "learning_rate": 4.850661990068047e-06,
      "loss": 0.2746,
      "step": 2180
    },
    {
      "epoch": 78.0,
      "eval_loss": 0.27298134565353394,
      "eval_runtime": 2.6537,
      "eval_samples_per_second": 116.817,
      "eval_steps_per_second": 1.884,
      "step": 2184
    },
    {
      "epoch": 78.21,
      "learning_rate": 4.703005015362185e-06,
      "loss": 0.2828,
      "step": 2190
    },
    {
      "epoch": 78.57,
      "learning_rate": 4.557307403424147e-06,
      "loss": 0.2687,
      "step": 2200
    },
    {
      "epoch": 78.93,
      "learning_rate": 4.41358947706863e-06,
      "loss": 0.2707,
      "step": 2210
    },
    {
      "epoch": 79.0,
      "eval_loss": 0.274308443069458,
      "eval_runtime": 2.6512,
      "eval_samples_per_second": 116.929,
      "eval_steps_per_second": 1.886,
      "step": 2212
    },
    {
      "epoch": 79.29,
      "learning_rate": 4.271871282971398e-06,
      "loss": 0.2788,
      "step": 2220
    },
    {
      "epoch": 79.64,
      "learning_rate": 4.1321725888730675e-06,
      "loss": 0.2692,
      "step": 2230
    },
    {
      "epoch": 80.0,
      "learning_rate": 3.994512880821713e-06,
      "loss": 0.2644,
      "step": 2240
    },
    {
      "epoch": 80.0,
      "eval_loss": 0.2740114629268646,
      "eval_runtime": 2.6658,
      "eval_samples_per_second": 116.29,
      "eval_steps_per_second": 1.876,
      "step": 2240
    },
    {
      "epoch": 80.36,
      "learning_rate": 3.858911360454918e-06,
      "loss": 0.2685,
      "step": 2250
    },
    {
      "epoch": 80.71,
      "learning_rate": 3.72538694232134e-06,
      "loss": 0.2691,
      "step": 2260
    },
    {
      "epoch": 81.0,
      "eval_loss": 0.2726958394050598,
      "eval_runtime": 2.6341,
      "eval_samples_per_second": 117.686,
      "eval_steps_per_second": 1.898,
      "step": 2268
    },
    {
      "epoch": 81.07,
      "learning_rate": 3.593958251242444e-06,
      "loss": 0.2748,
      "step": 2270
    },
    {
      "epoch": 81.43,
      "learning_rate": 3.4646436197145335e-06,
      "loss": 0.2689,
      "step": 2280
    },
    {
      "epoch": 81.79,
      "learning_rate": 3.3374610853516906e-06,
      "loss": 0.2679,
      "step": 2290
    },
    {
      "epoch": 82.0,
      "eval_loss": 0.27709755301475525,
      "eval_runtime": 2.6545,
      "eval_samples_per_second": 116.784,
      "eval_steps_per_second": 1.884,
      "step": 2296
    },
    {
      "epoch": 82.14,
      "learning_rate": 3.212428388369718e-06,
      "loss": 0.2699,
      "step": 2300
    },
    {
      "epoch": 82.5,
      "learning_rate": 3.0895629691116723e-06,
      "loss": 0.2768,
      "step": 2310
    },
    {
      "epoch": 82.86,
      "learning_rate": 2.9688819656151164e-06,
      "loss": 0.2748,
      "step": 2320
    },
    {
      "epoch": 83.0,
      "eval_loss": 0.2743559777736664,
      "eval_runtime": 2.6236,
      "eval_samples_per_second": 118.158,
      "eval_steps_per_second": 1.906,
      "step": 2324
    },
    {
      "epoch": 83.21,
      "learning_rate": 2.8504022112216642e-06,
      "loss": 0.2635,
      "step": 2330
    },
    {
      "epoch": 83.57,
      "learning_rate": 2.734140232228908e-06,
      "loss": 0.2667,
      "step": 2340
    },
    {
      "epoch": 83.93,
      "learning_rate": 2.6201122455852582e-06,
      "loss": 0.2744,
      "step": 2350
    },
    {
      "epoch": 84.0,
      "eval_loss": 0.2703389823436737,
      "eval_runtime": 2.6131,
      "eval_samples_per_second": 118.631,
      "eval_steps_per_second": 1.913,
      "step": 2352
    },
    {
      "epoch": 84.29,
      "learning_rate": 2.5083341566278907e-06,
      "loss": 0.2761,
      "step": 2360
    },
    {
      "epoch": 84.64,
      "learning_rate": 2.398821556864163e-06,
      "loss": 0.2699,
      "step": 2370
    },
    {
      "epoch": 85.0,
      "learning_rate": 2.291589721796841e-06,
      "loss": 0.2715,
      "step": 2380
    },
    {
      "epoch": 85.0,
      "eval_loss": 0.27332332730293274,
      "eval_runtime": 2.6685,
      "eval_samples_per_second": 116.168,
      "eval_steps_per_second": 1.874,
      "step": 2380
    },
    {
      "epoch": 85.36,
      "learning_rate": 2.1866536087933395e-06,
      "loss": 0.263,
      "step": 2390
    },
    {
      "epoch": 85.71,
      "learning_rate": 2.084027854999404e-06,
      "loss": 0.2682,
      "step": 2400
    },
    {
      "epoch": 86.0,
      "eval_loss": 0.271467000246048,
      "eval_runtime": 2.6604,
      "eval_samples_per_second": 116.522,
      "eval_steps_per_second": 1.879,
      "step": 2408
    },
    {
      "epoch": 86.07,
      "learning_rate": 1.9837267752974077e-06,
      "loss": 0.2783,
      "step": 2410
    },
    {
      "epoch": 86.43,
      "learning_rate": 1.8857643603096489e-06,
      "loss": 0.2776,
      "step": 2420
    },
    {
      "epoch": 86.79,
      "learning_rate": 1.7901542744468242e-06,
      "loss": 0.2641,
      "step": 2430
    },
    {
      "epoch": 87.0,
      "eval_loss": 0.27218252420425415,
      "eval_runtime": 2.6297,
      "eval_samples_per_second": 117.884,
      "eval_steps_per_second": 1.901,
      "step": 2436
    },
    {
      "epoch": 87.14,
      "learning_rate": 1.6969098540020453e-06,
      "loss": 0.2704,
      "step": 2440
    },
    {
      "epoch": 87.5,
      "learning_rate": 1.6060441052906117e-06,
      "loss": 0.2769,
      "step": 2450
    },
    {
      "epoch": 87.86,
      "learning_rate": 1.517569702835792e-06,
      "loss": 0.274,
      "step": 2460
    },
    {
      "epoch": 88.0,
      "eval_loss": 0.27482396364212036,
      "eval_runtime": 2.6311,
      "eval_samples_per_second": 117.823,
      "eval_steps_per_second": 1.9,
      "step": 2464
    },
    {
      "epoch": 88.21,
      "learning_rate": 1.4314989876009292e-06,
      "loss": 0.2561,
      "step": 2470
    },
    {
      "epoch": 88.57,
      "learning_rate": 1.3478439652680215e-06,
      "loss": 0.281,
      "step": 2480
    },
    {
      "epoch": 88.93,
      "learning_rate": 1.2666163045631007e-06,
      "loss": 0.2669,
      "step": 2490
    },
    {
      "epoch": 89.0,
      "eval_loss": 0.27534210681915283,
      "eval_runtime": 2.6393,
      "eval_samples_per_second": 117.455,
      "eval_steps_per_second": 1.894,
      "step": 2492
    },
    {
      "epoch": 89.29,
      "learning_rate": 1.1878273356286133e-06,
      "loss": 0.2765,
      "step": 2500
    },
    {
      "epoch": 89.64,
      "learning_rate": 1.1114880484430204e-06,
      "loss": 0.2635,
      "step": 2510
    },
    {
      "epoch": 90.0,
      "learning_rate": 1.0376090912878263e-06,
      "loss": 0.2707,
      "step": 2520
    },
    {
      "epoch": 90.0,
      "eval_loss": 0.2724357843399048,
      "eval_runtime": 2.6596,
      "eval_samples_per_second": 116.558,
      "eval_steps_per_second": 1.88,
      "step": 2520
    },
    {
      "epoch": 90.36,
      "learning_rate": 9.662007692623237e-07,
      "loss": 0.2667,
      "step": 2530
    },
    {
      "epoch": 90.71,
      "learning_rate": 8.972730428461452e-07,
      "loss": 0.2755,
      "step": 2540
    },
    {
      "epoch": 91.0,
      "eval_loss": 0.27034902572631836,
      "eval_runtime": 2.6286,
      "eval_samples_per_second": 117.934,
      "eval_steps_per_second": 1.902,
      "step": 2548
    },
    {
      "epoch": 91.07,
      "learning_rate": 8.308355265099334e-07,
      "loss": 0.2692,
      "step": 2550
    },
    {
      "epoch": 91.43,
      "learning_rate": 7.668974873742282e-07,
      "loss": 0.263,
      "step": 2560
    },
    {
      "epoch": 91.79,
      "learning_rate": 7.054678439168689e-07,
      "loss": 0.2769,
      "step": 2570
    },
    {
      "epoch": 92.0,
      "eval_loss": 0.27368903160095215,
      "eval_runtime": 2.6502,
      "eval_samples_per_second": 116.973,
      "eval_steps_per_second": 1.887,
      "step": 2576
    },
    {
      "epoch": 92.14,
      "learning_rate": 6.465551647289506e-07,
      "loss": 0.2722,
      "step": 2580
    },
    {
      "epoch": 92.5,
      "learning_rate": 5.901676673196507e-07,
      "loss": 0.2828,
      "step": 2590
    },
    {
      "epoch": 92.86,
      "learning_rate": 5.36313216969983e-07,
      "loss": 0.2659,
      "step": 2600
    },
    {
      "epoch": 93.0,
      "eval_loss": 0.2720969319343567,
      "eval_runtime": 2.6064,
      "eval_samples_per_second": 118.938,
      "eval_steps_per_second": 1.918,
      "step": 2604
    },
    {
      "epoch": 93.21,
      "learning_rate": 4.849993256357071e-07,
      "loss": 0.2608,
      "step": 2610
    },
    {
      "epoch": 93.57,
      "learning_rate": 4.362331508995204e-07,
      "loss": 0.2731,
      "step": 2620
    },
    {
      "epoch": 93.93,
      "learning_rate": 3.9002149497265255e-07,
      "loss": 0.2674,
      "step": 2630
    },
    {
      "epoch": 94.0,
      "eval_loss": 0.2763327658176422,
      "eval_runtime": 2.6289,
      "eval_samples_per_second": 117.922,
      "eval_steps_per_second": 1.902,
      "step": 2632
    },
    {
      "epoch": 94.29,
      "learning_rate": 3.463708037460683e-07,
      "loss": 0.277,
      "step": 2640
    },
    {
      "epoch": 94.64,
      "learning_rate": 3.052871658913533e-07,
      "loss": 0.2746,
      "step": 2650
    },
    {
      "epoch": 95.0,
      "learning_rate": 2.6677631201141856e-07,
      "loss": 0.2723,
      "step": 2660
    },
    {
      "epoch": 95.0,
      "eval_loss": 0.27226099371910095,
      "eval_runtime": 2.6569,
      "eval_samples_per_second": 116.678,
      "eval_steps_per_second": 1.882,
      "step": 2660
    },
    {
      "epoch": 95.36,
      "learning_rate": 2.3084361384116886e-07,
      "loss": 0.261,
      "step": 2670
    },
    {
      "epoch": 95.71,
      "learning_rate": 1.9749408349821265e-07,
      "loss": 0.2723,
      "step": 2680
    },
    {
      "epoch": 96.0,
      "eval_loss": 0.2743915319442749,
      "eval_runtime": 2.6421,
      "eval_samples_per_second": 117.329,
      "eval_steps_per_second": 1.892,
      "step": 2688
    },
    {
      "epoch": 96.07,
      "learning_rate": 1.6673237278375036e-07,
      "loss": 0.2775,
      "step": 2690
    },
    {
      "epoch": 96.43,
      "learning_rate": 1.3856277253369782e-07,
      "loss": 0.2813,
      "step": 2700
    },
    {
      "epoch": 96.79,
      "learning_rate": 1.1298921202018765e-07,
      "loss": 0.272,
      "step": 2710
    },
    {
      "epoch": 97.0,
      "eval_loss": 0.26857295632362366,
      "eval_runtime": 2.6133,
      "eval_samples_per_second": 118.624,
      "eval_steps_per_second": 1.913,
      "step": 2716
    },
    {
      "epoch": 97.14,
      "learning_rate": 9.001525840348495e-08,
      "loss": 0.2601,
      "step": 2720
    },
    {
      "epoch": 97.5,
      "learning_rate": 6.964411623440816e-08,
      "loss": 0.2756,
      "step": 2730
    },
    {
      "epoch": 97.86,
      "learning_rate": 5.187862700735332e-08,
      "loss": 0.27,
      "step": 2740
    },
    {
      "epoch": 98.0,
      "eval_loss": 0.27275460958480835,
      "eval_runtime": 2.6408,
      "eval_samples_per_second": 117.389,
      "eval_steps_per_second": 1.893,
      "step": 2744
    },
    {
      "epoch": 98.21,
      "learning_rate": 3.672126876393816e-08,
      "loss": 0.2757,
      "step": 2750
    },
    {
      "epoch": 98.57,
      "learning_rate": 2.4174155747345297e-08,
      "loss": 0.2662,
      "step": 2760
    },
    {
      "epoch": 98.93,
      "learning_rate": 1.4239038107416445e-08,
      "loss": 0.2721,
      "step": 2770
    },
    {
      "epoch": 99.0,
      "eval_loss": 0.27428731322288513,
      "eval_runtime": 2.6271,
      "eval_samples_per_second": 118.001,
      "eval_steps_per_second": 1.903,
      "step": 2772
    },
    {
      "epoch": 99.29,
      "learning_rate": 6.917301656537283e-09,
      "loss": 0.2792,
      "step": 2780
    },
    {
      "epoch": 99.64,
      "learning_rate": 2.2099676763295825e-09,
      "loss": 0.2751,
      "step": 2790
    },
    {
      "epoch": 100.0,
      "learning_rate": 1.176927751922463e-10,
      "loss": 0.2692,
      "step": 2800
    },
    {
      "epoch": 100.0,
      "eval_loss": 0.2747902274131775,
      "eval_runtime": 2.6706,
      "eval_samples_per_second": 116.08,
      "eval_steps_per_second": 1.872,
      "step": 2800
    },
    {
      "epoch": 100.0,
      "step": 2800,
      "total_flos": 1.7697633664499712e+19,
      "train_loss": 0.1087875749383654,
      "train_runtime": 1716.2917,
      "train_samples_per_second": 102.022,
      "train_steps_per_second": 1.631
    }
  ],
  "max_steps": 2800,
  "num_train_epochs": 100,
  "total_flos": 1.7697633664499712e+19,
  "trial_name": null,
  "trial_params": null
}