{
  "best_metric": 0.75,
  "best_model_checkpoint": "MAE-CT-CPC-Dicotomized-n0-m1-v8-1/checkpoint-980",
  "epoch": 49.02,
  "eval_steps": 500,
  "global_step": 3500,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.002857142857142857,
      "grad_norm": 3.5134518146514893,
      "learning_rate": 2.8571428571428575e-07,
      "loss": 0.2572,
      "step": 10
    },
    {
      "epoch": 0.005714285714285714,
      "grad_norm": 19.781003952026367,
      "learning_rate": 5.714285714285715e-07,
      "loss": 0.2784,
      "step": 20
    },
    {
      "epoch": 0.008571428571428572,
      "grad_norm": 27.620603561401367,
      "learning_rate": 8.571428571428572e-07,
      "loss": 0.4606,
      "step": 30
    },
    {
      "epoch": 0.011428571428571429,
      "grad_norm": 5.682175159454346,
      "learning_rate": 1.142857142857143e-06,
      "loss": 0.1624,
      "step": 40
    },
    {
      "epoch": 0.014285714285714285,
      "grad_norm": 57.06011199951172,
      "learning_rate": 1.4285714285714286e-06,
      "loss": 0.5168,
      "step": 50
    },
    {
      "epoch": 0.017142857142857144,
      "grad_norm": 12.413666725158691,
      "learning_rate": 1.7142857142857145e-06,
      "loss": 0.3112,
      "step": 60
    },
    {
      "epoch": 0.02,
      "grad_norm": 40.50763702392578,
      "learning_rate": 2.0000000000000003e-06,
      "loss": 0.3313,
      "step": 70
    },
    {
      "epoch": 0.02,
      "eval_accuracy": 0.6875,
      "eval_loss": 0.8278725147247314,
      "eval_runtime": 7.2014,
      "eval_samples_per_second": 4.444,
      "eval_steps_per_second": 1.111,
      "step": 70
    },
    {
      "epoch": 1.002857142857143,
      "grad_norm": 2.85071063041687,
      "learning_rate": 2.285714285714286e-06,
      "loss": 0.2007,
      "step": 80
    },
    {
      "epoch": 1.0057142857142858,
      "grad_norm": 3.0488057136535645,
      "learning_rate": 2.571428571428571e-06,
      "loss": 0.4132,
      "step": 90
    },
    {
      "epoch": 1.0085714285714287,
      "grad_norm": 1.0658628940582275,
      "learning_rate": 2.8571428571428573e-06,
      "loss": 0.2365,
      "step": 100
    },
    {
      "epoch": 1.0114285714285713,
      "grad_norm": 32.61592483520508,
      "learning_rate": 3.142857142857143e-06,
      "loss": 0.2773,
      "step": 110
    },
    {
      "epoch": 1.0142857142857142,
      "grad_norm": 0.24897827208042145,
      "learning_rate": 3.428571428571429e-06,
      "loss": 0.2216,
      "step": 120
    },
    {
      "epoch": 1.0171428571428571,
      "grad_norm": 40.215003967285156,
      "learning_rate": 3.7142857142857146e-06,
      "loss": 0.6413,
      "step": 130
    },
    {
      "epoch": 1.02,
      "grad_norm": 21.867782592773438,
      "learning_rate": 4.000000000000001e-06,
      "loss": 0.4602,
      "step": 140
    },
    {
      "epoch": 1.02,
      "eval_accuracy": 0.65625,
      "eval_loss": 0.7750873565673828,
      "eval_runtime": 6.0768,
      "eval_samples_per_second": 5.266,
      "eval_steps_per_second": 1.316,
      "step": 140
    },
    {
      "epoch": 2.0028571428571427,
      "grad_norm": 39.138465881347656,
      "learning_rate": 4.2857142857142855e-06,
      "loss": 0.1423,
      "step": 150
    },
    {
      "epoch": 2.005714285714286,
      "grad_norm": 0.8473556041717529,
      "learning_rate": 4.571428571428572e-06,
      "loss": 0.1991,
      "step": 160
    },
    {
      "epoch": 2.0085714285714285,
      "grad_norm": 31.097688674926758,
      "learning_rate": 4.857142857142858e-06,
      "loss": 0.327,
      "step": 170
    },
    {
      "epoch": 2.0114285714285716,
      "grad_norm": 0.581165611743927,
      "learning_rate": 5.142857142857142e-06,
      "loss": 0.332,
      "step": 180
    },
    {
      "epoch": 2.0142857142857142,
      "grad_norm": 54.0560188293457,
      "learning_rate": 5.428571428571429e-06,
      "loss": 0.3274,
      "step": 190
    },
    {
      "epoch": 2.0171428571428573,
      "grad_norm": 8.153368949890137,
      "learning_rate": 5.7142857142857145e-06,
      "loss": 0.1702,
      "step": 200
    },
    {
      "epoch": 2.02,
      "grad_norm": 82.87539672851562,
      "learning_rate": 6e-06,
      "loss": 0.413,
      "step": 210
    },
    {
      "epoch": 2.02,
      "eval_accuracy": 0.65625,
      "eval_loss": 0.9297134280204773,
      "eval_runtime": 6.3778,
      "eval_samples_per_second": 5.017,
      "eval_steps_per_second": 1.254,
      "step": 210
    },
    {
      "epoch": 3.0028571428571427,
      "grad_norm": 6.549503803253174,
      "learning_rate": 6.285714285714286e-06,
      "loss": 0.126,
      "step": 220
    },
    {
      "epoch": 3.005714285714286,
      "grad_norm": 0.07536959648132324,
      "learning_rate": 6.571428571428572e-06,
      "loss": 0.2436,
      "step": 230
    },
    {
      "epoch": 3.0085714285714285,
      "grad_norm": 75.6919174194336,
      "learning_rate": 6.857142857142858e-06,
      "loss": 0.3704,
      "step": 240
    },
    {
      "epoch": 3.0114285714285716,
      "grad_norm": 6.202108860015869,
      "learning_rate": 7.1428571428571436e-06,
      "loss": 0.2685,
      "step": 250
    },
    {
      "epoch": 3.0142857142857142,
      "grad_norm": 5.261022567749023,
      "learning_rate": 7.428571428571429e-06,
      "loss": 0.0579,
      "step": 260
    },
    {
      "epoch": 3.0171428571428573,
      "grad_norm": 62.4367790222168,
      "learning_rate": 7.714285714285716e-06,
      "loss": 0.4266,
      "step": 270
    },
    {
      "epoch": 3.02,
      "grad_norm": 0.3603902757167816,
      "learning_rate": 8.000000000000001e-06,
      "loss": 0.2196,
      "step": 280
    },
    {
      "epoch": 3.02,
      "eval_accuracy": 0.59375,
      "eval_loss": 1.307774543762207,
      "eval_runtime": 6.077,
      "eval_samples_per_second": 5.266,
      "eval_steps_per_second": 1.316,
      "step": 280
    },
    {
      "epoch": 4.002857142857143,
      "grad_norm": 82.43415832519531,
      "learning_rate": 8.285714285714287e-06,
      "loss": 0.1117,
      "step": 290
    },
    {
      "epoch": 4.005714285714285,
      "grad_norm": 1.1366825103759766,
      "learning_rate": 8.571428571428571e-06,
      "loss": 0.1042,
      "step": 300
    },
    {
      "epoch": 4.008571428571429,
      "grad_norm": 0.5070340037345886,
      "learning_rate": 8.857142857142858e-06,
      "loss": 0.3576,
      "step": 310
    },
    {
      "epoch": 4.011428571428572,
      "grad_norm": 0.18760715425014496,
      "learning_rate": 9.142857142857144e-06,
      "loss": 0.2198,
      "step": 320
    },
    {
      "epoch": 4.014285714285714,
      "grad_norm": 0.5741114020347595,
      "learning_rate": 9.42857142857143e-06,
      "loss": 0.111,
      "step": 330
    },
    {
      "epoch": 4.017142857142857,
      "grad_norm": 0.07025701552629471,
      "learning_rate": 9.714285714285715e-06,
      "loss": 0.149,
      "step": 340
    },
    {
      "epoch": 4.02,
      "grad_norm": 62.01303482055664,
      "learning_rate": 1e-05,
      "loss": 0.4898,
      "step": 350
    },
    {
      "epoch": 4.02,
      "eval_accuracy": 0.625,
      "eval_loss": 1.286909580230713,
      "eval_runtime": 6.1054,
      "eval_samples_per_second": 5.241,
      "eval_steps_per_second": 1.31,
      "step": 350
    },
    {
      "epoch": 5.002857142857143,
      "grad_norm": 59.823509216308594,
      "learning_rate": 9.968253968253969e-06,
      "loss": 0.2052,
      "step": 360
    },
    {
      "epoch": 5.005714285714285,
      "grad_norm": 0.020512551069259644,
      "learning_rate": 9.936507936507937e-06,
      "loss": 0.2319,
      "step": 370
    },
    {
      "epoch": 5.008571428571429,
      "grad_norm": 0.013519791886210442,
      "learning_rate": 9.904761904761906e-06,
      "loss": 0.2941,
      "step": 380
    },
    {
      "epoch": 5.011428571428572,
      "grad_norm": 103.2642822265625,
      "learning_rate": 9.873015873015874e-06,
      "loss": 0.123,
      "step": 390
    },
    {
      "epoch": 5.014285714285714,
      "grad_norm": 0.5531640648841858,
      "learning_rate": 9.841269841269842e-06,
      "loss": 0.3053,
      "step": 400
    },
    {
      "epoch": 5.017142857142857,
      "grad_norm": 8.770979881286621,
      "learning_rate": 9.80952380952381e-06,
      "loss": 0.3701,
      "step": 410
    },
    {
      "epoch": 5.02,
      "grad_norm": 5.519145965576172,
      "learning_rate": 9.777777777777779e-06,
      "loss": 0.1867,
      "step": 420
    },
    {
      "epoch": 5.02,
      "eval_accuracy": 0.625,
      "eval_loss": 1.5883848667144775,
      "eval_runtime": 6.5433,
      "eval_samples_per_second": 4.891,
      "eval_steps_per_second": 1.223,
      "step": 420
    },
    {
      "epoch": 6.002857142857143,
      "grad_norm": 114.1990966796875,
      "learning_rate": 9.746031746031747e-06,
      "loss": 0.1602,
      "step": 430
    },
    {
      "epoch": 6.005714285714285,
      "grad_norm": 0.034505173563957214,
      "learning_rate": 9.714285714285715e-06,
      "loss": 0.003,
      "step": 440
    },
    {
      "epoch": 6.008571428571429,
      "grad_norm": 3.3536415100097656,
      "learning_rate": 9.682539682539683e-06,
      "loss": 0.158,
      "step": 450
    },
    {
      "epoch": 6.011428571428572,
      "grad_norm": 0.010583457536995411,
      "learning_rate": 9.650793650793652e-06,
      "loss": 0.3164,
      "step": 460
    },
    {
      "epoch": 6.014285714285714,
      "grad_norm": 109.80305480957031,
      "learning_rate": 9.61904761904762e-06,
      "loss": 0.1235,
      "step": 470
    },
    {
      "epoch": 6.017142857142857,
      "grad_norm": 0.5373286008834839,
      "learning_rate": 9.587301587301588e-06,
      "loss": 0.3013,
      "step": 480
    },
    {
      "epoch": 6.02,
      "grad_norm": 9.764594078063965,
      "learning_rate": 9.555555555555556e-06,
      "loss": 0.0394,
      "step": 490
    },
    {
      "epoch": 6.02,
      "eval_accuracy": 0.625,
      "eval_loss": 1.7786798477172852,
      "eval_runtime": 6.3022,
      "eval_samples_per_second": 5.078,
      "eval_steps_per_second": 1.269,
      "step": 490
    },
    {
      "epoch": 7.002857142857143,
      "grad_norm": 0.10722041130065918,
      "learning_rate": 9.523809523809525e-06,
      "loss": 0.0015,
      "step": 500
    },
    {
      "epoch": 7.005714285714285,
      "grad_norm": 0.04226737469434738,
      "learning_rate": 9.492063492063493e-06,
      "loss": 0.201,
      "step": 510
    },
    {
      "epoch": 7.008571428571429,
      "grad_norm": 1.7212629318237305,
      "learning_rate": 9.460317460317461e-06,
      "loss": 0.0604,
      "step": 520
    },
    {
      "epoch": 7.011428571428572,
      "grad_norm": 0.01773512363433838,
      "learning_rate": 9.42857142857143e-06,
      "loss": 0.0661,
      "step": 530
    },
    {
      "epoch": 7.014285714285714,
      "grad_norm": 0.024557197466492653,
      "learning_rate": 9.396825396825398e-06,
      "loss": 0.0466,
      "step": 540
    },
    {
      "epoch": 7.017142857142857,
      "grad_norm": 0.01698165014386177,
      "learning_rate": 9.365079365079366e-06,
      "loss": 0.0009,
      "step": 550
    },
    {
      "epoch": 7.02,
      "grad_norm": 0.40697580575942993,
      "learning_rate": 9.333333333333334e-06,
      "loss": 0.1218,
      "step": 560
    },
    {
      "epoch": 7.02,
      "eval_accuracy": 0.59375,
      "eval_loss": 1.79988694190979,
      "eval_runtime": 5.9173,
      "eval_samples_per_second": 5.408,
      "eval_steps_per_second": 1.352,
      "step": 560
    },
    {
      "epoch": 8.002857142857144,
      "grad_norm": 0.09231330454349518,
      "learning_rate": 9.301587301587303e-06,
      "loss": 0.5704,
      "step": 570
    },
    {
      "epoch": 8.005714285714285,
      "grad_norm": 0.06326180696487427,
      "learning_rate": 9.26984126984127e-06,
      "loss": 0.1072,
      "step": 580
    },
    {
      "epoch": 8.008571428571429,
      "grad_norm": 0.077267587184906,
      "learning_rate": 9.238095238095239e-06,
      "loss": 0.0308,
      "step": 590
    },
    {
      "epoch": 8.01142857142857,
      "grad_norm": 0.05788983032107353,
      "learning_rate": 9.206349206349207e-06,
      "loss": 0.1882,
      "step": 600
    },
    {
      "epoch": 8.014285714285714,
      "grad_norm": 154.5283203125,
      "learning_rate": 9.174603174603176e-06,
      "loss": 0.4305,
      "step": 610
    },
    {
      "epoch": 8.017142857142858,
      "grad_norm": 0.03793035447597504,
      "learning_rate": 9.142857142857144e-06,
      "loss": 0.3163,
      "step": 620
    },
    {
      "epoch": 8.02,
      "grad_norm": 1.2432268857955933,
      "learning_rate": 9.111111111111112e-06,
      "loss": 0.192,
      "step": 630
    },
    {
      "epoch": 8.02,
      "eval_accuracy": 0.625,
      "eval_loss": 1.7294597625732422,
      "eval_runtime": 6.4051,
      "eval_samples_per_second": 4.996,
      "eval_steps_per_second": 1.249,
      "step": 630
    },
    {
      "epoch": 9.002857142857144,
      "grad_norm": 33.93547439575195,
      "learning_rate": 9.07936507936508e-06,
      "loss": 0.3914,
      "step": 640
    },
    {
      "epoch": 9.005714285714285,
      "grad_norm": 266.6748962402344,
      "learning_rate": 9.047619047619049e-06,
      "loss": 0.0957,
      "step": 650
    },
    {
      "epoch": 9.008571428571429,
      "grad_norm": 1.426464557647705,
      "learning_rate": 9.015873015873017e-06,
      "loss": 0.0762,
      "step": 660
    },
    {
      "epoch": 9.01142857142857,
      "grad_norm": 28.881240844726562,
      "learning_rate": 8.984126984126985e-06,
      "loss": 0.5325,
      "step": 670
    },
    {
      "epoch": 9.014285714285714,
      "grad_norm": 0.008092758245766163,
      "learning_rate": 8.952380952380953e-06,
      "loss": 0.3857,
      "step": 680
    },
    {
      "epoch": 9.017142857142858,
      "grad_norm": 53.07158279418945,
      "learning_rate": 8.920634920634922e-06,
      "loss": 0.6686,
      "step": 690
    },
    {
      "epoch": 9.02,
      "grad_norm": 147.34286499023438,
      "learning_rate": 8.888888888888888e-06,
      "loss": 0.3974,
      "step": 700
    },
    {
      "epoch": 9.02,
      "eval_accuracy": 0.65625,
      "eval_loss": 1.4708929061889648,
      "eval_runtime": 6.0018,
      "eval_samples_per_second": 5.332,
      "eval_steps_per_second": 1.333,
      "step": 700
    },
    {
      "epoch": 10.002857142857144,
      "grad_norm": 0.08053329586982727,
      "learning_rate": 8.857142857142858e-06,
      "loss": 0.2338,
      "step": 710
    },
    {
      "epoch": 10.005714285714285,
      "grad_norm": 0.04598785564303398,
      "learning_rate": 8.825396825396827e-06,
      "loss": 0.2478,
      "step": 720
    },
    {
      "epoch": 10.008571428571429,
      "grad_norm": 0.032458722591400146,
      "learning_rate": 8.793650793650795e-06,
      "loss": 0.3256,
      "step": 730
    },
    {
      "epoch": 10.01142857142857,
      "grad_norm": 3.2461020946502686,
      "learning_rate": 8.761904761904763e-06,
      "loss": 0.542,
      "step": 740
    },
    {
      "epoch": 10.014285714285714,
      "grad_norm": 38.98529815673828,
      "learning_rate": 8.730158730158731e-06,
      "loss": 0.3005,
      "step": 750
    },
    {
      "epoch": 10.017142857142858,
      "grad_norm": 5.08363151550293,
      "learning_rate": 8.6984126984127e-06,
      "loss": 0.0276,
      "step": 760
    },
    {
      "epoch": 10.02,
      "grad_norm": 1.1978472471237183,
      "learning_rate": 8.666666666666668e-06,
      "loss": 0.4529,
      "step": 770
    },
    {
      "epoch": 10.02,
      "eval_accuracy": 0.625,
      "eval_loss": 1.6416354179382324,
      "eval_runtime": 6.0866,
      "eval_samples_per_second": 5.257,
      "eval_steps_per_second": 1.314,
      "step": 770
    },
    {
      "epoch": 11.002857142857144,
      "grad_norm": 54.74907684326172,
      "learning_rate": 8.634920634920636e-06,
      "loss": 0.1156,
      "step": 780
    },
    {
      "epoch": 11.005714285714285,
      "grad_norm": 5.403316974639893,
      "learning_rate": 8.603174603174604e-06,
      "loss": 0.0258,
      "step": 790
    },
    {
      "epoch": 11.008571428571429,
      "grad_norm": 0.43036800622940063,
      "learning_rate": 8.571428571428571e-06,
      "loss": 0.1555,
      "step": 800
    },
    {
      "epoch": 11.01142857142857,
      "grad_norm": 176.08836364746094,
      "learning_rate": 8.53968253968254e-06,
      "loss": 0.5625,
      "step": 810
    },
    {
      "epoch": 11.014285714285714,
      "grad_norm": 52.73113250732422,
      "learning_rate": 8.507936507936509e-06,
      "loss": 0.1161,
      "step": 820
    },
    {
      "epoch": 11.017142857142858,
      "grad_norm": 0.23291529715061188,
      "learning_rate": 8.476190476190477e-06,
      "loss": 0.0022,
      "step": 830
    },
    {
      "epoch": 11.02,
      "grad_norm": 0.01016359031200409,
      "learning_rate": 8.444444444444446e-06,
      "loss": 0.0831,
      "step": 840
    },
    {
      "epoch": 11.02,
      "eval_accuracy": 0.6875,
      "eval_loss": 2.043224334716797,
      "eval_runtime": 5.9603,
      "eval_samples_per_second": 5.369,
      "eval_steps_per_second": 1.342,
      "step": 840
    },
    {
      "epoch": 12.002857142857144,
      "grad_norm": 0.025461383163928986,
      "learning_rate": 8.412698412698414e-06,
      "loss": 0.2046,
      "step": 850
    },
    {
      "epoch": 12.005714285714285,
      "grad_norm": 192.4024658203125,
      "learning_rate": 8.380952380952382e-06,
      "loss": 0.3202,
      "step": 860
    },
    {
      "epoch": 12.008571428571429,
      "grad_norm": 0.008105052635073662,
      "learning_rate": 8.34920634920635e-06,
      "loss": 0.2158,
      "step": 870
    },
    {
      "epoch": 12.01142857142857,
      "grad_norm": 0.009196307510137558,
      "learning_rate": 8.317460317460319e-06,
      "loss": 0.1244,
      "step": 880
    },
    {
      "epoch": 12.014285714285714,
      "grad_norm": 192.51708984375,
      "learning_rate": 8.285714285714287e-06,
      "loss": 0.1151,
      "step": 890
    },
    {
      "epoch": 12.017142857142858,
      "grad_norm": 1.3545656204223633,
      "learning_rate": 8.253968253968254e-06,
      "loss": 0.1114,
      "step": 900
    },
    {
      "epoch": 12.02,
      "grad_norm": 0.9690192341804504,
      "learning_rate": 8.222222222222222e-06,
      "loss": 0.0104,
      "step": 910
    },
    {
      "epoch": 12.02,
      "eval_accuracy": 0.59375,
      "eval_loss": 2.1283388137817383,
      "eval_runtime": 6.3439,
      "eval_samples_per_second": 5.044,
      "eval_steps_per_second": 1.261,
      "step": 910
    },
    {
      "epoch": 13.002857142857144,
      "grad_norm": 0.026668542996048927,
      "learning_rate": 8.190476190476192e-06,
      "loss": 0.0327,
      "step": 920
    },
    {
      "epoch": 13.005714285714285,
      "grad_norm": 0.01906030997633934,
      "learning_rate": 8.15873015873016e-06,
      "loss": 0.1063,
      "step": 930
    },
    {
      "epoch": 13.008571428571429,
      "grad_norm": 0.13048304617404938,
      "learning_rate": 8.126984126984128e-06,
      "loss": 0.1734,
      "step": 940
    },
    {
      "epoch": 13.01142857142857,
      "grad_norm": 0.012808074243366718,
      "learning_rate": 8.095238095238097e-06,
      "loss": 0.087,
      "step": 950
    },
    {
      "epoch": 13.014285714285714,
      "grad_norm": 219.556640625,
      "learning_rate": 8.063492063492065e-06,
      "loss": 0.2768,
      "step": 960
    },
    {
      "epoch": 13.017142857142858,
      "grad_norm": 0.009385952726006508,
      "learning_rate": 8.031746031746033e-06,
      "loss": 0.005,
      "step": 970
    },
    {
      "epoch": 13.02,
      "grad_norm": 21.89446449279785,
      "learning_rate": 8.000000000000001e-06,
      "loss": 0.0287,
      "step": 980
    },
    {
      "epoch": 13.02,
      "eval_accuracy": 0.75,
      "eval_loss": 1.853172779083252,
      "eval_runtime": 5.936,
      "eval_samples_per_second": 5.391,
      "eval_steps_per_second": 1.348,
      "step": 980
    },
    {
      "epoch": 14.002857142857144,
      "grad_norm": 0.0067256903275847435,
      "learning_rate": 7.968253968253968e-06,
      "loss": 0.0056,
      "step": 990
    },
    {
      "epoch": 14.005714285714285,
      "grad_norm": 309.82196044921875,
      "learning_rate": 7.936507936507936e-06,
      "loss": 0.1169,
      "step": 1000
    },
    {
      "epoch": 14.008571428571429,
      "grad_norm": 214.8688201904297,
      "learning_rate": 7.904761904761904e-06,
      "loss": 0.1285,
      "step": 1010
    },
    {
      "epoch": 14.01142857142857,
      "grad_norm": 0.005147908348590136,
      "learning_rate": 7.873015873015873e-06,
      "loss": 0.0166,
      "step": 1020
    },
    {
      "epoch": 14.014285714285714,
      "grad_norm": 0.013515499420464039,
      "learning_rate": 7.841269841269843e-06,
      "loss": 0.0549,
      "step": 1030
    },
    {
      "epoch": 14.017142857142858,
      "grad_norm": 0.005485304165631533,
      "learning_rate": 7.809523809523811e-06,
      "loss": 0.0002,
      "step": 1040
    },
    {
      "epoch": 14.02,
      "grad_norm": 0.026508044451475143,
      "learning_rate": 7.77777777777778e-06,
      "loss": 0.0426,
      "step": 1050
    },
    {
      "epoch": 14.02,
      "eval_accuracy": 0.6875,
      "eval_loss": 2.092482089996338,
      "eval_runtime": 6.4209,
      "eval_samples_per_second": 4.984,
      "eval_steps_per_second": 1.246,
      "step": 1050
    },
    {
      "epoch": 15.002857142857144,
      "grad_norm": 0.49312490224838257,
      "learning_rate": 7.746031746031747e-06,
      "loss": 0.0009,
      "step": 1060
    },
    {
      "epoch": 15.005714285714285,
      "grad_norm": 0.02295481227338314,
      "learning_rate": 7.714285714285716e-06,
      "loss": 0.0036,
      "step": 1070
    },
    {
      "epoch": 15.008571428571429,
      "grad_norm": 0.009320501238107681,
      "learning_rate": 7.682539682539684e-06,
      "loss": 0.1635,
      "step": 1080
    },
    {
      "epoch": 15.01142857142857,
      "grad_norm": 0.01872088387608528,
      "learning_rate": 7.65079365079365e-06,
      "loss": 0.4035,
      "step": 1090
    },
    {
      "epoch": 15.014285714285714,
      "grad_norm": 0.007027905434370041,
      "learning_rate": 7.61904761904762e-06,
      "loss": 0.0004,
      "step": 1100
    },
    {
      "epoch": 15.017142857142858,
      "grad_norm": 0.006057267542928457,
      "learning_rate": 7.587301587301588e-06,
      "loss": 0.3872,
      "step": 1110
    },
    {
      "epoch": 15.02,
      "grad_norm": 3.802086591720581,
      "learning_rate": 7.555555555555556e-06,
      "loss": 0.0145,
      "step": 1120
    },
    {
      "epoch": 15.02,
      "eval_accuracy": 0.625,
      "eval_loss": 2.351853370666504,
      "eval_runtime": 5.7934,
      "eval_samples_per_second": 5.524,
      "eval_steps_per_second": 1.381,
      "step": 1120
    },
    {
      "epoch": 16.002857142857142,
      "grad_norm": 92.55747985839844,
      "learning_rate": 7.523809523809524e-06,
      "loss": 0.1877,
      "step": 1130
    },
    {
      "epoch": 16.005714285714287,
      "grad_norm": 0.005619277711957693,
      "learning_rate": 7.492063492063493e-06,
      "loss": 0.001,
      "step": 1140
    },
    {
      "epoch": 16.00857142857143,
      "grad_norm": 196.1888885498047,
      "learning_rate": 7.460317460317461e-06,
      "loss": 0.1907,
      "step": 1150
    },
    {
      "epoch": 16.01142857142857,
      "grad_norm": 0.005707439035177231,
      "learning_rate": 7.428571428571429e-06,
      "loss": 0.1524,
      "step": 1160
    },
    {
      "epoch": 16.014285714285716,
      "grad_norm": 0.018987352028489113,
      "learning_rate": 7.3968253968253975e-06,
      "loss": 0.2278,
      "step": 1170
    },
    {
      "epoch": 16.017142857142858,
      "grad_norm": 0.029227713122963905,
      "learning_rate": 7.3650793650793666e-06,
      "loss": 0.0281,
      "step": 1180
    },
    {
      "epoch": 16.02,
      "grad_norm": 0.006308805197477341,
      "learning_rate": 7.333333333333333e-06,
      "loss": 0.0024,
      "step": 1190
    },
    {
      "epoch": 16.02,
      "eval_accuracy": 0.625,
      "eval_loss": 2.070525646209717,
      "eval_runtime": 5.8174,
      "eval_samples_per_second": 5.501,
      "eval_steps_per_second": 1.375,
      "step": 1190
    },
    {
      "epoch": 17.002857142857142,
      "grad_norm": 0.0030912691727280617,
      "learning_rate": 7.301587301587301e-06,
      "loss": 0.1571,
      "step": 1200
    },
    {
      "epoch": 17.005714285714287,
      "grad_norm": 0.024552060291171074,
      "learning_rate": 7.2698412698412705e-06,
      "loss": 0.0002,
      "step": 1210
    },
    {
      "epoch": 17.00857142857143,
      "grad_norm": 0.01592804305255413,
      "learning_rate": 7.238095238095239e-06,
      "loss": 0.0002,
      "step": 1220
    },
    {
      "epoch": 17.01142857142857,
      "grad_norm": 0.039011064916849136,
      "learning_rate": 7.206349206349207e-06,
      "loss": 0.0003,
      "step": 1230
    },
    {
      "epoch": 17.014285714285716,
      "grad_norm": 0.0051954942755401134,
      "learning_rate": 7.174603174603175e-06,
      "loss": 0.0028,
      "step": 1240
    },
    {
      "epoch": 17.017142857142858,
      "grad_norm": 0.014356645755469799,
      "learning_rate": 7.1428571428571436e-06,
      "loss": 0.0001,
      "step": 1250
    },
    {
      "epoch": 17.02,
      "grad_norm": 0.04629682004451752,
      "learning_rate": 7.111111111111112e-06,
      "loss": 0.0176,
      "step": 1260
    },
    {
      "epoch": 17.02,
      "eval_accuracy": 0.625,
      "eval_loss": 2.360275983810425,
      "eval_runtime": 6.3101,
      "eval_samples_per_second": 5.071,
      "eval_steps_per_second": 1.268,
      "step": 1260
    },
    {
      "epoch": 18.002857142857142,
      "grad_norm": 0.002282822737470269,
      "learning_rate": 7.07936507936508e-06,
      "loss": 0.0003,
      "step": 1270
    },
    {
      "epoch": 18.005714285714287,
      "grad_norm": 0.007674033287912607,
      "learning_rate": 7.047619047619048e-06,
      "loss": 0.005,
      "step": 1280
    },
    {
      "epoch": 18.00857142857143,
      "grad_norm": 107.018310546875,
      "learning_rate": 7.015873015873016e-06,
      "loss": 0.223,
      "step": 1290
    },
    {
      "epoch": 18.01142857142857,
      "grad_norm": 0.027542075142264366,
      "learning_rate": 6.984126984126984e-06,
      "loss": 0.2571,
      "step": 1300
    },
    {
      "epoch": 18.014285714285716,
      "grad_norm": 0.013514521531760693,
      "learning_rate": 6.952380952380952e-06,
      "loss": 0.1242,
      "step": 1310
    },
    {
      "epoch": 18.017142857142858,
      "grad_norm": 0.004058144986629486,
      "learning_rate": 6.920634920634921e-06,
      "loss": 0.0026,
      "step": 1320
    },
    {
      "epoch": 18.02,
      "grad_norm": 0.052907612174749374,
      "learning_rate": 6.88888888888889e-06,
      "loss": 0.0023,
      "step": 1330
    },
    {
      "epoch": 18.02,
      "eval_accuracy": 0.65625,
      "eval_loss": 2.3409135341644287,
      "eval_runtime": 5.6456,
      "eval_samples_per_second": 5.668,
      "eval_steps_per_second": 1.417,
      "step": 1330
    },
    {
      "epoch": 19.002857142857142,
      "grad_norm": 0.003937090281397104,
      "learning_rate": 6.857142857142858e-06,
      "loss": 0.0083,
      "step": 1340
    },
    {
      "epoch": 19.005714285714287,
      "grad_norm": 0.17072078585624695,
      "learning_rate": 6.825396825396826e-06,
      "loss": 0.0016,
      "step": 1350
    },
    {
      "epoch": 19.00857142857143,
      "grad_norm": 0.0025130421854555607,
      "learning_rate": 6.7936507936507944e-06,
      "loss": 0.0018,
      "step": 1360
    },
    {
      "epoch": 19.01142857142857,
      "grad_norm": 0.0014831767184659839,
      "learning_rate": 6.761904761904763e-06,
      "loss": 0.007,
      "step": 1370
    },
    {
      "epoch": 19.014285714285716,
      "grad_norm": 48.3748893737793,
      "learning_rate": 6.730158730158731e-06,
      "loss": 0.2262,
      "step": 1380
    },
    {
      "epoch": 19.017142857142858,
      "grad_norm": 0.0189586840569973,
      "learning_rate": 6.698412698412698e-06,
      "loss": 0.1803,
      "step": 1390
    },
    {
      "epoch": 19.02,
      "grad_norm": 0.003652828047052026,
      "learning_rate": 6.666666666666667e-06,
      "loss": 0.1111,
      "step": 1400
    },
    {
      "epoch": 19.02,
      "eval_accuracy": 0.625,
      "eval_loss": 2.3460936546325684,
      "eval_runtime": 5.6345,
      "eval_samples_per_second": 5.679,
      "eval_steps_per_second": 1.42,
      "step": 1400
    },
    {
      "epoch": 20.002857142857142,
      "grad_norm": 0.033134013414382935,
      "learning_rate": 6.634920634920635e-06,
      "loss": 0.2262,
      "step": 1410
    },
    {
      "epoch": 20.005714285714287,
      "grad_norm": 0.015590427443385124,
      "learning_rate": 6.603174603174603e-06,
      "loss": 0.0606,
      "step": 1420
    },
    {
      "epoch": 20.00857142857143,
      "grad_norm": 0.008228067308664322,
      "learning_rate": 6.571428571428572e-06,
      "loss": 0.0001,
      "step": 1430
    },
    {
      "epoch": 20.01142857142857,
      "grad_norm": 0.007313844282180071,
      "learning_rate": 6.5396825396825405e-06,
      "loss": 0.0812,
      "step": 1440
    },
    {
      "epoch": 20.014285714285716,
      "grad_norm": 0.03062896430492401,
      "learning_rate": 6.507936507936509e-06,
      "loss": 0.3162,
      "step": 1450
    },
    {
      "epoch": 20.017142857142858,
      "grad_norm": 0.00798776838928461,
      "learning_rate": 6.476190476190477e-06,
      "loss": 0.0034,
      "step": 1460
    },
    {
      "epoch": 20.02,
      "grad_norm": 0.0768931582570076,
      "learning_rate": 6.444444444444445e-06,
      "loss": 0.0002,
      "step": 1470
    },
    {
      "epoch": 20.02,
      "eval_accuracy": 0.6875,
      "eval_loss": 2.0998449325561523,
      "eval_runtime": 6.206,
      "eval_samples_per_second": 5.156,
      "eval_steps_per_second": 1.289,
      "step": 1470
    },
    {
      "epoch": 21.002857142857142,
      "grad_norm": 0.08019956201314926,
      "learning_rate": 6.412698412698414e-06,
      "loss": 0.0007,
      "step": 1480
    },
    {
      "epoch": 21.005714285714287,
      "grad_norm": 0.008561481721699238,
      "learning_rate": 6.380952380952381e-06,
      "loss": 0.0003,
      "step": 1490
    },
    {
      "epoch": 21.00857142857143,
      "grad_norm": 0.17695729434490204,
      "learning_rate": 6.349206349206349e-06,
      "loss": 0.0085,
      "step": 1500
    },
    {
      "epoch": 21.01142857142857,
      "grad_norm": 0.0022817952558398247,
      "learning_rate": 6.3174603174603175e-06,
      "loss": 0.0001,
      "step": 1510
    },
    {
      "epoch": 21.014285714285716,
      "grad_norm": 0.004223952069878578,
      "learning_rate": 6.285714285714286e-06,
      "loss": 0.0008,
      "step": 1520
    },
    {
      "epoch": 21.017142857142858,
      "grad_norm": 0.004527420271188021,
      "learning_rate": 6.253968253968254e-06,
      "loss": 0.0002,
      "step": 1530
    },
    {
      "epoch": 21.02,
      "grad_norm": 0.020056281238794327,
      "learning_rate": 6.222222222222223e-06,
      "loss": 0.0432,
      "step": 1540
    },
    {
      "epoch": 21.02,
      "eval_accuracy": 0.59375,
      "eval_loss": 2.2702555656433105,
      "eval_runtime": 5.806,
      "eval_samples_per_second": 5.512,
      "eval_steps_per_second": 1.378,
      "step": 1540
    },
    {
      "epoch": 22.002857142857142,
      "grad_norm": 0.002663273364305496,
      "learning_rate": 6.1904761904761914e-06,
      "loss": 0.1748,
      "step": 1550
    },
    {
      "epoch": 22.005714285714287,
      "grad_norm": 0.002945307642221451,
      "learning_rate": 6.15873015873016e-06,
      "loss": 0.0364,
      "step": 1560
    },
    {
      "epoch": 22.00857142857143,
      "grad_norm": 6.704668045043945,
      "learning_rate": 6.126984126984128e-06,
      "loss": 0.0104,
      "step": 1570
    },
    {
      "epoch": 22.01142857142857,
      "grad_norm": 0.0020908783189952374,
      "learning_rate": 6.095238095238096e-06,
      "loss": 0.0893,
      "step": 1580
    },
    {
      "epoch": 22.014285714285716,
      "grad_norm": 0.004042261280119419,
      "learning_rate": 6.063492063492064e-06,
      "loss": 0.1432,
      "step": 1590
    },
    {
      "epoch": 22.017142857142858,
      "grad_norm": 0.026853501796722412,
      "learning_rate": 6.031746031746032e-06,
      "loss": 0.0002,
      "step": 1600
    },
    {
      "epoch": 22.02,
      "grad_norm": 0.002750640967860818,
      "learning_rate": 6e-06,
      "loss": 0.006,
      "step": 1610
    },
    {
      "epoch": 22.02,
      "eval_accuracy": 0.65625,
      "eval_loss": 2.3306009769439697,
      "eval_runtime": 5.6486,
      "eval_samples_per_second": 5.665,
      "eval_steps_per_second": 1.416,
      "step": 1610
    },
    {
      "epoch": 23.002857142857142,
      "grad_norm": 0.0037402939051389694,
      "learning_rate": 5.968253968253968e-06,
      "loss": 0.0001,
      "step": 1620
    },
    {
      "epoch": 23.005714285714287,
      "grad_norm": 0.002422990510240197,
      "learning_rate": 5.936507936507937e-06,
      "loss": 0.0014,
      "step": 1630
    },
    {
      "epoch": 23.00857142857143,
      "grad_norm": 0.004334081895649433,
      "learning_rate": 5.904761904761905e-06,
      "loss": 0.0052,
      "step": 1640
    },
    {
      "epoch": 23.01142857142857,
      "grad_norm": 0.0010784378973767161,
      "learning_rate": 5.873015873015874e-06,
      "loss": 0.1291,
      "step": 1650
    },
    {
      "epoch": 23.014285714285716,
      "grad_norm": 0.0017067866865545511,
      "learning_rate": 5.841269841269842e-06,
      "loss": 0.0001,
      "step": 1660
    },
    {
      "epoch": 23.017142857142858,
      "grad_norm": 0.08872966468334198,
      "learning_rate": 5.8095238095238106e-06,
      "loss": 0.0001,
      "step": 1670
    },
    {
      "epoch": 23.02,
      "grad_norm": 0.001540614292025566,
      "learning_rate": 5.777777777777778e-06,
      "loss": 0.0014,
      "step": 1680
    },
    {
      "epoch": 23.02,
      "eval_accuracy": 0.625,
      "eval_loss": 2.446324348449707,
      "eval_runtime": 6.0676,
      "eval_samples_per_second": 5.274,
      "eval_steps_per_second": 1.318,
      "step": 1680
    },
    {
      "epoch": 24.002857142857142,
      "grad_norm": 0.005013082176446915,
      "learning_rate": 5.746031746031746e-06,
      "loss": 0.0001,
      "step": 1690
    },
    {
      "epoch": 24.005714285714287,
      "grad_norm": 0.09714154154062271,
      "learning_rate": 5.7142857142857145e-06,
      "loss": 0.0016,
      "step": 1700
    },
    {
      "epoch": 24.00857142857143,
      "grad_norm": 0.0058003757148981094,
      "learning_rate": 5.682539682539683e-06,
      "loss": 0.0001,
      "step": 1710
    },
    {
      "epoch": 24.01142857142857,
      "grad_norm": 0.0029599149711430073,
      "learning_rate": 5.650793650793651e-06,
      "loss": 0.0024,
      "step": 1720
    },
    {
      "epoch": 24.014285714285716,
      "grad_norm": 0.001937561435624957,
      "learning_rate": 5.619047619047619e-06,
      "loss": 0.1056,
      "step": 1730
    },
    {
      "epoch": 24.017142857142858,
      "grad_norm": 0.02229410596191883,
      "learning_rate": 5.5873015873015876e-06,
      "loss": 0.0001,
      "step": 1740
    },
    {
      "epoch": 24.02,
      "grad_norm": 0.004641120787709951,
      "learning_rate": 5.555555555555557e-06,
      "loss": 0.0173,
      "step": 1750
    },
    {
      "epoch": 24.02,
      "eval_accuracy": 0.65625,
      "eval_loss": 2.556457042694092,
      "eval_runtime": 5.7748,
      "eval_samples_per_second": 5.541,
      "eval_steps_per_second": 1.385,
      "step": 1750
    },
    {
      "epoch": 25.002857142857142,
      "grad_norm": 0.0013144315453246236,
      "learning_rate": 5.523809523809525e-06,
      "loss": 0.0001,
      "step": 1760
    },
    {
      "epoch": 25.005714285714287,
      "grad_norm": 0.00479417759925127,
      "learning_rate": 5.492063492063493e-06,
      "loss": 0.071,
      "step": 1770
    },
    {
      "epoch": 25.00857142857143,
      "grad_norm": 0.0011662091128528118,
      "learning_rate": 5.460317460317461e-06,
      "loss": 0.0001,
      "step": 1780
    },
    {
      "epoch": 25.01142857142857,
      "grad_norm": 0.003162315348163247,
      "learning_rate": 5.428571428571429e-06,
      "loss": 0.0001,
      "step": 1790
    },
    {
      "epoch": 25.014285714285716,
      "grad_norm": 0.004806616809219122,
      "learning_rate": 5.396825396825397e-06,
      "loss": 0.0,
      "step": 1800
    },
    {
      "epoch": 25.017142857142858,
      "grad_norm": 0.010089408606290817,
      "learning_rate": 5.365079365079365e-06,
      "loss": 0.0001,
      "step": 1810
    },
    {
      "epoch": 25.02,
      "grad_norm": 0.0010941866785287857,
      "learning_rate": 5.333333333333334e-06,
      "loss": 0.0001,
      "step": 1820
    },
    {
      "epoch": 25.02,
      "eval_accuracy": 0.625,
      "eval_loss": 2.7107367515563965,
      "eval_runtime": 6.0469,
      "eval_samples_per_second": 5.292,
      "eval_steps_per_second": 1.323,
      "step": 1820
    },
    {
      "epoch": 26.002857142857142,
      "grad_norm": 0.003195547964423895,
      "learning_rate": 5.301587301587302e-06,
      "loss": 0.1521,
      "step": 1830
    },
    {
      "epoch": 26.005714285714287,
      "grad_norm": 0.009643026627600193,
      "learning_rate": 5.26984126984127e-06,
      "loss": 0.0005,
      "step": 1840
    },
    {
      "epoch": 26.00857142857143,
      "grad_norm": 0.0033331862650811672,
      "learning_rate": 5.2380952380952384e-06,
      "loss": 0.0001,
      "step": 1850
    },
    {
      "epoch": 26.01142857142857,
      "grad_norm": 0.0029240387957543135,
      "learning_rate": 5.2063492063492076e-06,
      "loss": 0.0001,
      "step": 1860
    },
    {
      "epoch": 26.014285714285716,
      "grad_norm": 0.007532194256782532,
      "learning_rate": 5.174603174603176e-06,
      "loss": 0.0001,
      "step": 1870
    },
    {
      "epoch": 26.017142857142858,
      "grad_norm": 0.00938828382641077,
      "learning_rate": 5.142857142857142e-06,
      "loss": 0.0001,
      "step": 1880
    },
    {
      "epoch": 26.02,
      "grad_norm": 0.0021244913805276155,
      "learning_rate": 5.1111111111111115e-06,
      "loss": 0.0001,
      "step": 1890
    },
    {
      "epoch": 26.02,
      "eval_accuracy": 0.59375,
      "eval_loss": 3.092498302459717,
      "eval_runtime": 6.514,
      "eval_samples_per_second": 4.912,
      "eval_steps_per_second": 1.228,
      "step": 1890
    },
    {
      "epoch": 27.002857142857142,
      "grad_norm": 0.0038645656313747168,
      "learning_rate": 5.07936507936508e-06,
      "loss": 0.0001,
      "step": 1900
    },
    {
      "epoch": 27.005714285714287,
      "grad_norm": 0.0035917272325605154,
      "learning_rate": 5.047619047619048e-06,
      "loss": 0.0001,
      "step": 1910
    },
    {
      "epoch": 27.00857142857143,
      "grad_norm": 0.002653130330145359,
      "learning_rate": 5.015873015873016e-06,
      "loss": 0.0001,
      "step": 1920
    },
    {
      "epoch": 27.01142857142857,
      "grad_norm": 0.001524316263385117,
      "learning_rate": 4.9841269841269845e-06,
      "loss": 0.0001,
      "step": 1930
    },
    {
      "epoch": 27.014285714285716,
      "grad_norm": 0.00874999351799488,
      "learning_rate": 4.952380952380953e-06,
      "loss": 0.0001,
      "step": 1940
    },
    {
      "epoch": 27.017142857142858,
      "grad_norm": 0.0011307639069855213,
      "learning_rate": 4.920634920634921e-06,
      "loss": 0.0001,
      "step": 1950
    },
    {
      "epoch": 27.02,
      "grad_norm": 0.0028375948313623667,
      "learning_rate": 4.888888888888889e-06,
      "loss": 0.1792,
      "step": 1960
    },
    {
      "epoch": 27.02,
      "eval_accuracy": 0.625,
      "eval_loss": 2.536710500717163,
      "eval_runtime": 5.634,
      "eval_samples_per_second": 5.68,
      "eval_steps_per_second": 1.42,
      "step": 1960
    },
    {
      "epoch": 28.002857142857142,
      "grad_norm": 0.00281306984834373,
      "learning_rate": 4.857142857142858e-06,
      "loss": 0.0049,
      "step": 1970
    },
    {
      "epoch": 28.005714285714287,
      "grad_norm": 0.001993451965972781,
      "learning_rate": 4.825396825396826e-06,
      "loss": 0.0001,
      "step": 1980
    },
    {
      "epoch": 28.00857142857143,
      "grad_norm": 0.0021242112852633,
      "learning_rate": 4.793650793650794e-06,
      "loss": 0.0,
      "step": 1990
    },
    {
      "epoch": 28.01142857142857,
      "grad_norm": 0.0019546120893210173,
      "learning_rate": 4.761904761904762e-06,
      "loss": 0.0001,
      "step": 2000
    },
    {
      "epoch": 28.014285714285716,
      "grad_norm": 0.0021625503432005644,
      "learning_rate": 4.730158730158731e-06,
      "loss": 0.0144,
      "step": 2010
    },
    {
      "epoch": 28.017142857142858,
      "grad_norm": 0.001239254605025053,
      "learning_rate": 4.698412698412699e-06,
      "loss": 0.0001,
      "step": 2020
    },
    {
      "epoch": 28.02,
      "grad_norm": 0.0012608050601556897,
      "learning_rate": 4.666666666666667e-06,
      "loss": 0.0004,
      "step": 2030
    },
    {
      "epoch": 28.02,
      "eval_accuracy": 0.5625,
      "eval_loss": 2.9952054023742676,
      "eval_runtime": 6.2717,
      "eval_samples_per_second": 5.102,
      "eval_steps_per_second": 1.276,
      "step": 2030
    },
    {
      "epoch": 29.002857142857142,
      "grad_norm": 0.0023157112300395966,
      "learning_rate": 4.634920634920635e-06,
      "loss": 0.0001,
      "step": 2040
    },
    {
      "epoch": 29.005714285714287,
      "grad_norm": 0.0012295986525714397,
      "learning_rate": 4.603174603174604e-06,
      "loss": 0.0001,
      "step": 2050
    },
    {
      "epoch": 29.00857142857143,
      "grad_norm": 0.005791760981082916,
      "learning_rate": 4.571428571428572e-06,
      "loss": 0.0001,
      "step": 2060
    },
    {
      "epoch": 29.01142857142857,
      "grad_norm": 0.003394305007532239,
      "learning_rate": 4.53968253968254e-06,
      "loss": 0.1151,
      "step": 2070
    },
    {
      "epoch": 29.014285714285716,
      "grad_norm": 0.0017611499642953277,
      "learning_rate": 4.5079365079365085e-06,
      "loss": 0.0001,
      "step": 2080
    },
    {
      "epoch": 29.017142857142858,
      "grad_norm": 0.0028206913266330957,
      "learning_rate": 4.476190476190477e-06,
      "loss": 0.0001,
      "step": 2090
    },
    {
      "epoch": 29.02,
      "grad_norm": 0.0028997170738875866,
      "learning_rate": 4.444444444444444e-06,
      "loss": 0.0001,
      "step": 2100
    },
    {
      "epoch": 29.02,
      "eval_accuracy": 0.625,
      "eval_loss": 2.7962350845336914,
      "eval_runtime": 6.6512,
      "eval_samples_per_second": 4.811,
      "eval_steps_per_second": 1.203,
      "step": 2100
    },
    {
      "epoch": 30.002857142857142,
      "grad_norm": 0.004685261752456427,
      "learning_rate": 4.412698412698413e-06,
      "loss": 0.0001,
      "step": 2110
    },
    {
      "epoch": 30.005714285714287,
      "grad_norm": 0.0024346013087779284,
      "learning_rate": 4.3809523809523815e-06,
      "loss": 0.0,
      "step": 2120
    },
    {
      "epoch": 30.00857142857143,
      "grad_norm": 0.0026598083786666393,
      "learning_rate": 4.34920634920635e-06,
      "loss": 0.0,
      "step": 2130
    },
    {
      "epoch": 30.01142857142857,
      "grad_norm": 0.002897858852520585,
      "learning_rate": 4.317460317460318e-06,
      "loss": 0.0001,
      "step": 2140
    },
    {
      "epoch": 30.014285714285716,
      "grad_norm": 0.01607625186443329,
      "learning_rate": 4.2857142857142855e-06,
      "loss": 0.0001,
      "step": 2150
    },
    {
      "epoch": 30.017142857142858,
      "grad_norm": 0.002586986403912306,
      "learning_rate": 4.2539682539682546e-06,
      "loss": 0.0,
      "step": 2160
    },
    {
      "epoch": 30.02,
      "grad_norm": 0.0013718613190576434,
      "learning_rate": 4.222222222222223e-06,
      "loss": 0.0,
      "step": 2170
    },
    {
      "epoch": 30.02,
      "eval_accuracy": 0.59375,
      "eval_loss": 3.108156204223633,
      "eval_runtime": 5.9931,
      "eval_samples_per_second": 5.34,
      "eval_steps_per_second": 1.335,
      "step": 2170
    },
    {
      "epoch": 31.002857142857142,
      "grad_norm": 0.004130430519580841,
      "learning_rate": 4.190476190476191e-06,
      "loss": 0.0001,
      "step": 2180
    },
    {
      "epoch": 31.005714285714287,
      "grad_norm": 0.0036993578542023897,
      "learning_rate": 4.158730158730159e-06,
      "loss": 0.0001,
      "step": 2190
    },
    {
      "epoch": 31.00857142857143,
      "grad_norm": 0.0024625733494758606,
      "learning_rate": 4.126984126984127e-06,
      "loss": 0.0008,
      "step": 2200
    },
    {
      "epoch": 31.01142857142857,
      "grad_norm": 0.007389547768980265,
      "learning_rate": 4.095238095238096e-06,
      "loss": 0.1028,
      "step": 2210
    },
    {
      "epoch": 31.014285714285716,
      "grad_norm": 0.001320167793892324,
      "learning_rate": 4.063492063492064e-06,
      "loss": 0.0004,
      "step": 2220
    },
    {
      "epoch": 31.017142857142858,
      "grad_norm": 466.27154541015625,
      "learning_rate": 4.031746031746032e-06,
      "loss": 0.069,
      "step": 2230
    },
    {
      "epoch": 31.02,
      "grad_norm": 0.004771101288497448,
      "learning_rate": 4.000000000000001e-06,
      "loss": 0.0001,
      "step": 2240
    },
    {
      "epoch": 31.02,
      "eval_accuracy": 0.625,
      "eval_loss": 2.686725616455078,
      "eval_runtime": 7.1944,
      "eval_samples_per_second": 4.448,
      "eval_steps_per_second": 1.112,
      "step": 2240
    },
    {
      "epoch": 32.002857142857145,
      "grad_norm": 0.004328535404056311,
      "learning_rate": 3.968253968253968e-06,
      "loss": 0.0001,
      "step": 2250
    },
    {
      "epoch": 32.005714285714284,
      "grad_norm": 0.007427292410284281,
      "learning_rate": 3.936507936507936e-06,
      "loss": 0.0001,
      "step": 2260
    },
    {
      "epoch": 32.00857142857143,
      "grad_norm": 0.001498078927397728,
      "learning_rate": 3.9047619047619055e-06,
      "loss": 0.0024,
      "step": 2270
    },
    {
      "epoch": 32.011428571428574,
      "grad_norm": 0.0010804787743836641,
      "learning_rate": 3.873015873015874e-06,
      "loss": 0.0,
      "step": 2280
    },
    {
      "epoch": 32.01428571428571,
      "grad_norm": 0.07372881472110748,
      "learning_rate": 3.841269841269842e-06,
      "loss": 0.0284,
      "step": 2290
    },
    {
      "epoch": 32.01714285714286,
      "grad_norm": 0.002815135521814227,
      "learning_rate": 3.80952380952381e-06,
      "loss": 0.0,
      "step": 2300
    },
    {
      "epoch": 32.02,
      "grad_norm": 0.0011799400672316551,
      "learning_rate": 3.777777777777778e-06,
      "loss": 0.0,
      "step": 2310
    },
    {
      "epoch": 32.02,
      "eval_accuracy": 0.625,
      "eval_loss": 2.6702141761779785,
      "eval_runtime": 7.3014,
      "eval_samples_per_second": 4.383,
      "eval_steps_per_second": 1.096,
      "step": 2310
    },
    {
      "epoch": 33.002857142857145,
      "grad_norm": 0.004017932340502739,
      "learning_rate": 3.7460317460317463e-06,
      "loss": 0.0046,
      "step": 2320
    },
    {
      "epoch": 33.005714285714284,
      "grad_norm": 0.002392595401033759,
      "learning_rate": 3.7142857142857146e-06,
      "loss": 0.0,
      "step": 2330
    },
    {
      "epoch": 33.00857142857143,
      "grad_norm": 0.0008441254030913115,
      "learning_rate": 3.6825396825396833e-06,
      "loss": 0.0,
      "step": 2340
    },
    {
      "epoch": 33.011428571428574,
      "grad_norm": 0.01818065159022808,
      "learning_rate": 3.6507936507936507e-06,
      "loss": 0.0,
      "step": 2350
    },
    {
      "epoch": 33.01428571428571,
      "grad_norm": 0.0024400062393397093,
      "learning_rate": 3.6190476190476194e-06,
      "loss": 0.0002,
      "step": 2360
    },
    {
      "epoch": 33.01714285714286,
      "grad_norm": 0.001259080832824111,
      "learning_rate": 3.5873015873015877e-06,
      "loss": 0.0244,
      "step": 2370
    },
    {
      "epoch": 33.02,
      "grad_norm": 0.0022309625055640936,
      "learning_rate": 3.555555555555556e-06,
      "loss": 0.0,
      "step": 2380
    },
    {
      "epoch": 33.02,
      "eval_accuracy": 0.59375,
      "eval_loss": 3.3402888774871826,
      "eval_runtime": 7.3496,
      "eval_samples_per_second": 4.354,
      "eval_steps_per_second": 1.088,
      "step": 2380
    },
    {
      "epoch": 34.002857142857145,
      "grad_norm": 0.001507407519966364,
      "learning_rate": 3.523809523809524e-06,
      "loss": 0.1484,
      "step": 2390
    },
    {
      "epoch": 34.005714285714284,
      "grad_norm": 0.001335164182819426,
      "learning_rate": 3.492063492063492e-06,
      "loss": 0.0,
      "step": 2400
    },
    {
      "epoch": 34.00857142857143,
      "grad_norm": 0.0010363466572016478,
      "learning_rate": 3.4603174603174607e-06,
      "loss": 0.0,
      "step": 2410
    },
    {
      "epoch": 34.011428571428574,
      "grad_norm": 0.0027341984678059816,
      "learning_rate": 3.428571428571429e-06,
      "loss": 0.0,
      "step": 2420
    },
    {
      "epoch": 34.01428571428571,
      "grad_norm": 0.007578617427498102,
      "learning_rate": 3.3968253968253972e-06,
      "loss": 0.0001,
      "step": 2430
    },
    {
      "epoch": 34.01714285714286,
      "grad_norm": 0.001968264114111662,
      "learning_rate": 3.3650793650793655e-06,
      "loss": 0.0,
      "step": 2440
    },
    {
      "epoch": 34.02,
      "grad_norm": 0.0030165906064212322,
      "learning_rate": 3.3333333333333333e-06,
      "loss": 0.0,
      "step": 2450
    },
    {
      "epoch": 34.02,
      "eval_accuracy": 0.625,
      "eval_loss": 3.053858757019043,
      "eval_runtime": 8.5887,
      "eval_samples_per_second": 3.726,
      "eval_steps_per_second": 0.931,
      "step": 2450
    },
    {
      "epoch": 35.002857142857145,
      "grad_norm": 0.0015692274319007993,
      "learning_rate": 3.3015873015873016e-06,
      "loss": 0.0,
      "step": 2460
    },
    {
      "epoch": 35.005714285714284,
      "grad_norm": 0.0013480527559295297,
      "learning_rate": 3.2698412698412703e-06,
      "loss": 0.0,
      "step": 2470
    },
    {
      "epoch": 35.00857142857143,
      "grad_norm": 0.003836671821773052,
      "learning_rate": 3.2380952380952385e-06,
      "loss": 0.1318,
      "step": 2480
    },
    {
      "epoch": 35.011428571428574,
      "grad_norm": 0.0024338432122021914,
      "learning_rate": 3.206349206349207e-06,
      "loss": 0.0,
      "step": 2490
    },
    {
      "epoch": 35.01428571428571,
      "grad_norm": 0.0123060941696167,
      "learning_rate": 3.1746031746031746e-06,
      "loss": 0.0001,
      "step": 2500
    },
    {
      "epoch": 35.01714285714286,
      "grad_norm": 0.001270064152777195,
      "learning_rate": 3.142857142857143e-06,
      "loss": 0.0,
      "step": 2510
    },
    {
      "epoch": 35.02,
      "grad_norm": 0.0015140828909352422,
      "learning_rate": 3.1111111111111116e-06,
      "loss": 0.2031,
      "step": 2520
    },
    {
      "epoch": 35.02,
      "eval_accuracy": 0.6875,
      "eval_loss": 2.73325252532959,
      "eval_runtime": 8.3821,
      "eval_samples_per_second": 3.818,
      "eval_steps_per_second": 0.954,
      "step": 2520
    },
    {
      "epoch": 36.002857142857145,
      "grad_norm": 0.0005715539446100593,
      "learning_rate": 3.07936507936508e-06,
      "loss": 0.0,
      "step": 2530
    },
    {
      "epoch": 36.005714285714284,
      "grad_norm": 0.003236456075683236,
      "learning_rate": 3.047619047619048e-06,
      "loss": 0.0,
      "step": 2540
    },
    {
      "epoch": 36.00857142857143,
      "grad_norm": 0.0009249376598745584,
      "learning_rate": 3.015873015873016e-06,
      "loss": 0.2215,
      "step": 2550
    },
    {
      "epoch": 36.011428571428574,
      "grad_norm": 0.0008299104520119727,
      "learning_rate": 2.984126984126984e-06,
      "loss": 0.031,
      "step": 2560
    },
    {
      "epoch": 36.01428571428571,
      "grad_norm": 0.0019169561564922333,
      "learning_rate": 2.9523809523809525e-06,
      "loss": 0.2959,
      "step": 2570
    },
    {
      "epoch": 36.01714285714286,
      "grad_norm": 0.016660558059811592,
      "learning_rate": 2.920634920634921e-06,
      "loss": 0.0,
      "step": 2580
    },
    {
      "epoch": 36.02,
      "grad_norm": 0.0024716516491025686,
      "learning_rate": 2.888888888888889e-06,
      "loss": 0.0,
      "step": 2590
    },
    {
      "epoch": 36.02,
      "eval_accuracy": 0.59375,
      "eval_loss": 2.9419589042663574,
      "eval_runtime": 7.6689,
      "eval_samples_per_second": 4.173,
      "eval_steps_per_second": 1.043,
      "step": 2590
    },
    {
      "epoch": 37.002857142857145,
      "grad_norm": 0.0009388060425408185,
      "learning_rate": 2.8571428571428573e-06,
      "loss": 0.0,
      "step": 2600
    },
    {
      "epoch": 37.005714285714284,
      "grad_norm": 0.0014066151343286037,
      "learning_rate": 2.8253968253968255e-06,
      "loss": 0.0049,
      "step": 2610
    },
    {
      "epoch": 37.00857142857143,
      "grad_norm": 0.0009500043233856559,
      "learning_rate": 2.7936507936507938e-06,
      "loss": 0.0,
      "step": 2620
    },
    {
      "epoch": 37.011428571428574,
      "grad_norm": 0.0007660359842702746,
      "learning_rate": 2.7619047619047625e-06,
      "loss": 0.0002,
      "step": 2630
    },
    {
      "epoch": 37.01428571428571,
      "grad_norm": 0.005225882399827242,
      "learning_rate": 2.7301587301587303e-06,
      "loss": 0.0,
      "step": 2640
    },
    {
      "epoch": 37.01714285714286,
      "grad_norm": 0.0036217437591403723,
      "learning_rate": 2.6984126984126986e-06,
      "loss": 0.0001,
      "step": 2650
    },
    {
      "epoch": 37.02,
      "grad_norm": 0.0021994500420987606,
      "learning_rate": 2.666666666666667e-06,
      "loss": 0.0,
      "step": 2660
    },
    {
      "epoch": 37.02,
      "eval_accuracy": 0.65625,
      "eval_loss": 2.919912338256836,
      "eval_runtime": 7.4701,
      "eval_samples_per_second": 4.284,
      "eval_steps_per_second": 1.071,
      "step": 2660
    },
    {
      "epoch": 38.002857142857145,
      "grad_norm": 0.0009795749792829156,
      "learning_rate": 2.634920634920635e-06,
      "loss": 0.0,
      "step": 2670
    },
    {
      "epoch": 38.005714285714284,
      "grad_norm": 0.0015403656288981438,
      "learning_rate": 2.6031746031746038e-06,
      "loss": 0.0007,
      "step": 2680
    },
    {
      "epoch": 38.00857142857143,
      "grad_norm": 0.007823006249964237,
      "learning_rate": 2.571428571428571e-06,
      "loss": 0.0701,
      "step": 2690
    },
    {
      "epoch": 38.011428571428574,
      "grad_norm": 0.12522433698177338,
      "learning_rate": 2.53968253968254e-06,
      "loss": 0.0643,
      "step": 2700
    },
    {
      "epoch": 38.01428571428571,
      "grad_norm": 0.0010730504291132092,
      "learning_rate": 2.507936507936508e-06,
      "loss": 0.0001,
      "step": 2710
    },
    {
      "epoch": 38.01714285714286,
      "grad_norm": 0.001265152939595282,
      "learning_rate": 2.4761904761904764e-06,
      "loss": 0.2465,
      "step": 2720
    },
    {
      "epoch": 38.02,
      "grad_norm": 0.45712485909461975,
      "learning_rate": 2.4444444444444447e-06,
      "loss": 0.0018,
      "step": 2730
    },
    {
      "epoch": 38.02,
      "eval_accuracy": 0.59375,
      "eval_loss": 2.8933959007263184,
      "eval_runtime": 7.7323,
      "eval_samples_per_second": 4.138,
      "eval_steps_per_second": 1.035,
      "step": 2730
    },
    {
      "epoch": 39.002857142857145,
      "grad_norm": 0.0031093251891434193,
      "learning_rate": 2.412698412698413e-06,
      "loss": 0.0042,
      "step": 2740
    },
    {
      "epoch": 39.005714285714284,
      "grad_norm": 0.004323534667491913,
      "learning_rate": 2.380952380952381e-06,
      "loss": 0.0,
      "step": 2750
    },
    {
      "epoch": 39.00857142857143,
      "grad_norm": 0.0008992227958515286,
      "learning_rate": 2.3492063492063494e-06,
      "loss": 0.0,
      "step": 2760
    },
    {
      "epoch": 39.011428571428574,
      "grad_norm": 0.0029326954390853643,
      "learning_rate": 2.3174603174603177e-06,
      "loss": 0.0,
      "step": 2770
    },
    {
      "epoch": 39.01428571428571,
      "grad_norm": 0.0016185936983674765,
      "learning_rate": 2.285714285714286e-06,
      "loss": 0.0,
      "step": 2780
    },
    {
      "epoch": 39.01714285714286,
      "grad_norm": 0.005735237151384354,
      "learning_rate": 2.2539682539682542e-06,
      "loss": 0.0,
      "step": 2790
    },
    {
      "epoch": 39.02,
      "grad_norm": 0.0014396852347999811,
      "learning_rate": 2.222222222222222e-06,
      "loss": 0.0,
      "step": 2800
    },
    {
      "epoch": 39.02,
      "eval_accuracy": 0.625,
      "eval_loss": 2.832045078277588,
      "eval_runtime": 7.5616,
      "eval_samples_per_second": 4.232,
      "eval_steps_per_second": 1.058,
      "step": 2800
    },
    {
      "epoch": 40.002857142857145,
      "grad_norm": 0.0041121807880699635,
      "learning_rate": 2.1904761904761908e-06,
      "loss": 0.0,
      "step": 2810
    },
    {
      "epoch": 40.005714285714284,
      "grad_norm": 0.0013577585341408849,
      "learning_rate": 2.158730158730159e-06,
      "loss": 0.0,
      "step": 2820
    },
    {
      "epoch": 40.00857142857143,
      "grad_norm": 0.0017459145747125149,
      "learning_rate": 2.1269841269841273e-06,
      "loss": 0.0,
      "step": 2830
    },
    {
      "epoch": 40.011428571428574,
      "grad_norm": 0.0019927776884287596,
      "learning_rate": 2.0952380952380955e-06,
      "loss": 0.0,
      "step": 2840
    },
    {
      "epoch": 40.01428571428571,
      "grad_norm": 0.0009335738723166287,
      "learning_rate": 2.0634920634920634e-06,
      "loss": 0.0,
      "step": 2850
    },
    {
      "epoch": 40.01714285714286,
      "grad_norm": 0.0010714407544583082,
      "learning_rate": 2.031746031746032e-06,
      "loss": 0.0,
      "step": 2860
    },
    {
      "epoch": 40.02,
      "grad_norm": 0.0007551051676273346,
      "learning_rate": 2.0000000000000003e-06,
      "loss": 0.0,
      "step": 2870
    },
    {
      "epoch": 40.02,
      "eval_accuracy": 0.625,
      "eval_loss": 2.8381476402282715,
      "eval_runtime": 7.9752,
      "eval_samples_per_second": 4.012,
      "eval_steps_per_second": 1.003,
      "step": 2870
    },
    {
      "epoch": 41.002857142857145,
      "grad_norm": 0.001309627783484757,
      "learning_rate": 1.968253968253968e-06,
      "loss": 0.0001,
      "step": 2880
    },
    {
      "epoch": 41.005714285714284,
      "grad_norm": 0.0011593608651310205,
      "learning_rate": 1.936507936507937e-06,
      "loss": 0.0,
      "step": 2890
    },
    {
      "epoch": 41.00857142857143,
      "grad_norm": 0.0005718428292311728,
      "learning_rate": 1.904761904761905e-06,
      "loss": 0.0,
      "step": 2900
    },
    {
      "epoch": 41.011428571428574,
      "grad_norm": 0.0008164609898813069,
      "learning_rate": 1.8730158730158732e-06,
      "loss": 0.0,
      "step": 2910
    },
    {
      "epoch": 41.01428571428571,
      "grad_norm": 0.0068624867126345634,
      "learning_rate": 1.8412698412698416e-06,
      "loss": 0.0,
      "step": 2920
    },
    {
      "epoch": 41.01714285714286,
      "grad_norm": 0.002447180449962616,
      "learning_rate": 1.8095238095238097e-06,
      "loss": 0.0,
      "step": 2930
    },
    {
      "epoch": 41.02,
      "grad_norm": 0.001345404889434576,
      "learning_rate": 1.777777777777778e-06,
      "loss": 0.0,
      "step": 2940
    },
    {
      "epoch": 41.02,
      "eval_accuracy": 0.625,
      "eval_loss": 2.8438665866851807,
      "eval_runtime": 6.6318,
      "eval_samples_per_second": 4.825,
      "eval_steps_per_second": 1.206,
      "step": 2940
    },
    {
      "epoch": 42.002857142857145,
      "grad_norm": 0.004871055483818054,
      "learning_rate": 1.746031746031746e-06,
      "loss": 0.0,
      "step": 2950
    },
    {
      "epoch": 42.005714285714284,
      "grad_norm": 0.0018843573052436113,
      "learning_rate": 1.7142857142857145e-06,
      "loss": 0.0,
      "step": 2960
    },
    {
      "epoch": 42.00857142857143,
      "grad_norm": 0.0010428063105791807,
      "learning_rate": 1.6825396825396827e-06,
      "loss": 0.0001,
      "step": 2970
    },
    {
      "epoch": 42.011428571428574,
      "grad_norm": 0.0007727840566076338,
      "learning_rate": 1.6507936507936508e-06,
      "loss": 0.0,
      "step": 2980
    },
    {
      "epoch": 42.01428571428571,
      "grad_norm": 0.0022347676567733288,
      "learning_rate": 1.6190476190476193e-06,
      "loss": 0.0,
      "step": 2990
    },
    {
      "epoch": 42.01714285714286,
      "grad_norm": 0.0014822926605120301,
      "learning_rate": 1.5873015873015873e-06,
      "loss": 0.0,
      "step": 3000
    },
    {
      "epoch": 42.02,
      "grad_norm": 0.0008702924242243171,
      "learning_rate": 1.5555555555555558e-06,
      "loss": 0.0,
      "step": 3010
    },
    {
      "epoch": 42.02,
      "eval_accuracy": 0.625,
      "eval_loss": 2.8524856567382812,
      "eval_runtime": 6.573,
      "eval_samples_per_second": 4.868,
      "eval_steps_per_second": 1.217,
      "step": 3010
    },
    {
      "epoch": 43.002857142857145,
      "grad_norm": 0.001486993976868689,
      "learning_rate": 1.523809523809524e-06,
      "loss": 0.0,
      "step": 3020
    },
    {
      "epoch": 43.005714285714284,
      "grad_norm": 0.0009380248957313597,
      "learning_rate": 1.492063492063492e-06,
      "loss": 0.0,
      "step": 3030
    },
    {
      "epoch": 43.00857142857143,
      "grad_norm": 0.004946484230458736,
      "learning_rate": 1.4603174603174606e-06,
      "loss": 0.0,
      "step": 3040
    },
    {
      "epoch": 43.011428571428574,
      "grad_norm": 0.0016762095037847757,
      "learning_rate": 1.4285714285714286e-06,
      "loss": 0.0002,
      "step": 3050
    },
    {
      "epoch": 43.01428571428571,
      "grad_norm": 0.001995038939639926,
      "learning_rate": 1.3968253968253969e-06,
      "loss": 0.0,
      "step": 3060
    },
    {
      "epoch": 43.01714285714286,
      "grad_norm": 0.0014240635791793466,
      "learning_rate": 1.3650793650793652e-06,
      "loss": 0.0001,
      "step": 3070
    },
    {
      "epoch": 43.02,
      "grad_norm": 0.0009568611858412623,
      "learning_rate": 1.3333333333333334e-06,
      "loss": 0.0,
      "step": 3080
    },
    {
      "epoch": 43.02,
      "eval_accuracy": 0.59375,
      "eval_loss": 2.6921167373657227,
      "eval_runtime": 7.5296,
      "eval_samples_per_second": 4.25,
      "eval_steps_per_second": 1.062,
      "step": 3080
    },
    {
      "epoch": 44.002857142857145,
      "grad_norm": 0.01802194118499756,
      "learning_rate": 1.3015873015873019e-06,
      "loss": 0.0,
      "step": 3090
    },
    {
      "epoch": 44.005714285714284,
      "grad_norm": 0.000703324971254915,
      "learning_rate": 1.26984126984127e-06,
      "loss": 0.0,
      "step": 3100
    },
    {
      "epoch": 44.00857142857143,
      "grad_norm": 0.0015024804743006825,
      "learning_rate": 1.2380952380952382e-06,
      "loss": 0.0,
      "step": 3110
    },
    {
      "epoch": 44.011428571428574,
      "grad_norm": 0.0023550696205347776,
      "learning_rate": 1.2063492063492065e-06,
      "loss": 0.0,
      "step": 3120
    },
    {
      "epoch": 44.01428571428571,
      "grad_norm": 0.0009432346560060978,
      "learning_rate": 1.1746031746031747e-06,
      "loss": 0.0001,
      "step": 3130
    },
    {
      "epoch": 44.01714285714286,
      "grad_norm": 0.0014145496534183621,
      "learning_rate": 1.142857142857143e-06,
      "loss": 0.0,
      "step": 3140
    },
    {
      "epoch": 44.02,
      "grad_norm": 0.0011549839982762933,
      "learning_rate": 1.111111111111111e-06,
      "loss": 0.0,
      "step": 3150
    },
    {
      "epoch": 44.02,
      "eval_accuracy": 0.625,
      "eval_loss": 2.9579904079437256,
      "eval_runtime": 7.1267,
      "eval_samples_per_second": 4.49,
      "eval_steps_per_second": 1.123,
      "step": 3150
    },
    {
      "epoch": 45.002857142857145,
      "grad_norm": 0.004226373508572578,
      "learning_rate": 1.0793650793650795e-06,
      "loss": 0.0003,
      "step": 3160
    },
    {
      "epoch": 45.005714285714284,
      "grad_norm": 0.0005224282504059374,
      "learning_rate": 1.0476190476190478e-06,
      "loss": 0.0,
      "step": 3170
    },
    {
      "epoch": 45.00857142857143,
      "grad_norm": 0.0016394226113334298,
      "learning_rate": 1.015873015873016e-06,
      "loss": 0.0,
      "step": 3180
    },
    {
      "epoch": 45.011428571428574,
      "grad_norm": 0.004058617167174816,
      "learning_rate": 9.84126984126984e-07,
      "loss": 0.0,
      "step": 3190
    },
    {
      "epoch": 45.01428571428571,
      "grad_norm": 0.0013102118391543627,
      "learning_rate": 9.523809523809525e-07,
      "loss": 0.0,
      "step": 3200
    },
    {
      "epoch": 45.01714285714286,
      "grad_norm": 0.0019316988764330745,
      "learning_rate": 9.206349206349208e-07,
      "loss": 0.0,
      "step": 3210
    },
    {
      "epoch": 45.02,
      "grad_norm": 0.000915765471290797,
      "learning_rate": 8.88888888888889e-07,
      "loss": 0.0,
      "step": 3220
    },
    {
      "epoch": 45.02,
      "eval_accuracy": 0.59375,
      "eval_loss": 3.029210090637207,
      "eval_runtime": 6.7888,
      "eval_samples_per_second": 4.714,
      "eval_steps_per_second": 1.178,
      "step": 3220
    },
    {
      "epoch": 46.002857142857145,
      "grad_norm": 0.001245733117684722,
      "learning_rate": 8.571428571428572e-07,
      "loss": 0.068,
      "step": 3230
    },
    {
      "epoch": 46.005714285714284,
      "grad_norm": 0.0010560519294813275,
      "learning_rate": 8.253968253968254e-07,
      "loss": 0.0,
      "step": 3240
    },
    {
      "epoch": 46.00857142857143,
      "grad_norm": 0.0011368553387001157,
      "learning_rate": 7.936507936507937e-07,
      "loss": 0.0,
      "step": 3250
    },
    {
      "epoch": 46.011428571428574,
      "grad_norm": 0.00201380904763937,
      "learning_rate": 7.61904761904762e-07,
      "loss": 0.0,
      "step": 3260
    },
    {
      "epoch": 46.01428571428571,
      "grad_norm": 0.0008550912025384605,
      "learning_rate": 7.301587301587303e-07,
      "loss": 0.0,
      "step": 3270
    },
    {
      "epoch": 46.01714285714286,
      "grad_norm": 0.0008424161933362484,
      "learning_rate": 6.984126984126984e-07,
      "loss": 0.0,
      "step": 3280
    },
    {
      "epoch": 46.02,
      "grad_norm": 0.0006465838523581624,
      "learning_rate": 6.666666666666667e-07,
      "loss": 0.0,
      "step": 3290
    },
    {
      "epoch": 46.02,
      "eval_accuracy": 0.625,
      "eval_loss": 3.1119141578674316,
      "eval_runtime": 7.4394,
      "eval_samples_per_second": 4.301,
      "eval_steps_per_second": 1.075,
      "step": 3290
    },
    {
      "epoch": 47.002857142857145,
      "grad_norm": 0.0010397338774055243,
      "learning_rate": 6.34920634920635e-07,
      "loss": 0.0,
      "step": 3300
    },
    {
      "epoch": 47.005714285714284,
      "grad_norm": 0.0018721016822382808,
      "learning_rate": 6.031746031746032e-07,
      "loss": 0.0,
      "step": 3310
    },
    {
      "epoch": 47.00857142857143,
      "grad_norm": 0.0009640470379963517,
      "learning_rate": 5.714285714285715e-07,
      "loss": 0.0,
      "step": 3320
    },
    {
      "epoch": 47.011428571428574,
      "grad_norm": 0.0024156668223440647,
      "learning_rate": 5.396825396825398e-07,
      "loss": 0.0,
      "step": 3330
    },
    {
      "epoch": 47.01428571428571,
      "grad_norm": 0.0012575440341606736,
      "learning_rate": 5.07936507936508e-07,
      "loss": 0.0,
      "step": 3340
    },
    {
      "epoch": 47.01714285714286,
      "grad_norm": 0.0007373109692707658,
      "learning_rate": 4.7619047619047623e-07,
      "loss": 0.0,
      "step": 3350
    },
    {
      "epoch": 47.02,
      "grad_norm": 0.0050420379266142845,
      "learning_rate": 4.444444444444445e-07,
      "loss": 0.0,
      "step": 3360
    },
    {
      "epoch": 47.02,
      "eval_accuracy": 0.625,
      "eval_loss": 3.1150264739990234,
      "eval_runtime": 7.5053,
      "eval_samples_per_second": 4.264,
      "eval_steps_per_second": 1.066,
      "step": 3360
    },
    {
      "epoch": 48.002857142857145,
      "grad_norm": 0.001978852553293109,
      "learning_rate": 4.126984126984127e-07,
      "loss": 0.0,
      "step": 3370
    },
    {
      "epoch": 48.005714285714284,
      "grad_norm": 0.0011000612284988165,
      "learning_rate": 3.80952380952381e-07,
      "loss": 0.0007,
      "step": 3380
    },
    {
      "epoch": 48.00857142857143,
      "grad_norm": 0.0011735723819583654,
      "learning_rate": 3.492063492063492e-07,
      "loss": 0.0,
      "step": 3390
    },
    {
      "epoch": 48.011428571428574,
      "grad_norm": 0.0010873243445530534,
      "learning_rate": 3.174603174603175e-07,
      "loss": 0.0,
      "step": 3400
    },
    {
      "epoch": 48.01428571428571,
      "grad_norm": 0.0011251752730458975,
      "learning_rate": 2.8571428571428575e-07,
      "loss": 0.0,
      "step": 3410
    },
    {
      "epoch": 48.01714285714286,
      "grad_norm": 0.0013193455524742603,
      "learning_rate": 2.53968253968254e-07,
      "loss": 0.0,
      "step": 3420
    },
    {
      "epoch": 48.02,
      "grad_norm": 0.0006919482257217169,
      "learning_rate": 2.2222222222222224e-07,
      "loss": 0.0,
      "step": 3430
    },
    {
      "epoch": 48.02,
      "eval_accuracy": 0.625,
      "eval_loss": 3.112720489501953,
      "eval_runtime": 7.7277,
      "eval_samples_per_second": 4.141,
      "eval_steps_per_second": 1.035,
      "step": 3430
    },
    {
      "epoch": 49.002857142857145,
      "grad_norm": 0.0010557180503383279,
      "learning_rate": 1.904761904761905e-07,
      "loss": 0.0,
      "step": 3440
    },
    {
      "epoch": 49.005714285714284,
      "grad_norm": 0.006352465599775314,
      "learning_rate": 1.5873015873015874e-07,
      "loss": 0.0,
      "step": 3450
    },
    {
      "epoch": 49.00857142857143,
      "grad_norm": 0.0013192709302529693,
      "learning_rate": 1.26984126984127e-07,
      "loss": 0.0,
      "step": 3460
    },
    {
      "epoch": 49.011428571428574,
      "grad_norm": 0.0010165185667574406,
      "learning_rate": 9.523809523809525e-08,
      "loss": 0.0,
      "step": 3470
    },
    {
      "epoch": 49.01428571428571,
      "grad_norm": 0.000668003223836422,
      "learning_rate": 6.34920634920635e-08,
      "loss": 0.0,
      "step": 3480
    },
    {
      "epoch": 49.01714285714286,
      "grad_norm": 0.0007290692301467061,
      "learning_rate": 3.174603174603175e-08,
      "loss": 0.0,
      "step": 3490
    },
    {
      "epoch": 49.02,
      "grad_norm": 0.0005605400656349957,
      "learning_rate": 0.0,
      "loss": 0.0,
      "step": 3500
    },
    {
      "epoch": 49.02,
      "eval_accuracy": 0.625,
      "eval_loss": 3.1125903129577637,
      "eval_runtime": 7.757,
      "eval_samples_per_second": 4.125,
      "eval_steps_per_second": 1.031,
      "step": 3500
    },
    {
      "epoch": 49.02,
      "step": 3500,
      "total_flos": 6.147436841415475e+19,
      "train_loss": 0.08523796891186586,
      "train_runtime": 6238.6099,
      "train_samples_per_second": 2.244,
      "train_steps_per_second": 0.561
    },
    {
      "epoch": 49.02,
      "eval_accuracy": 0.8372093023255814,
      "eval_loss": 0.8805378675460815,
      "eval_runtime": 9.9977,
      "eval_samples_per_second": 4.301,
      "eval_steps_per_second": 1.1,
      "step": 3500
    },
    {
      "epoch": 49.02,
      "eval_accuracy": 0.8372093023255814,
      "eval_loss": 0.8805379867553711,
      "eval_runtime": 9.0595,
      "eval_samples_per_second": 4.746,
      "eval_steps_per_second": 1.214,
      "step": 3500
    }
  ],
  "logging_steps": 10,
  "max_steps": 3500,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 9223372036854775807,
  "save_steps": 500,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": true
      },
      "attributes": {}
    }
  },
  "total_flos": 6.147436841415475e+19,
  "train_batch_size": 4,
  "trial_name": null,
  "trial_params": null
}