{
  "best_metric": 0.01655612699687481,
  "best_model_checkpoint": "runs/deepseek_20240423-162824/checkpoint-3000",
  "epoch": 1.5665387326701654,
  "eval_steps": 1000,
  "global_step": 5000,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.0,
      "grad_norm": 11.5,
      "learning_rate": 4.0000000000000003e-07,
      "loss": 1.5189,
      "step": 10
    },
    {
      "epoch": 0.01,
      "grad_norm": 16.375,
      "learning_rate": 8.000000000000001e-07,
      "loss": 1.5082,
      "step": 20
    },
    {
      "epoch": 0.01,
      "grad_norm": 21.75,
      "learning_rate": 1.2000000000000002e-06,
      "loss": 1.4715,
      "step": 30
    },
    {
      "epoch": 0.01,
      "grad_norm": 13.8125,
      "learning_rate": 1.6000000000000001e-06,
      "loss": 1.4515,
      "step": 40
    },
    {
      "epoch": 0.02,
      "grad_norm": 9.8125,
      "learning_rate": 2.0000000000000003e-06,
      "loss": 1.5133,
      "step": 50
    },
    {
      "epoch": 0.02,
      "grad_norm": 10.25,
      "learning_rate": 2.4000000000000003e-06,
      "loss": 1.313,
      "step": 60
    },
    {
      "epoch": 0.02,
      "grad_norm": 18.125,
      "learning_rate": 2.8000000000000003e-06,
      "loss": 1.4489,
      "step": 70
    },
    {
      "epoch": 0.03,
      "grad_norm": 28.0,
      "learning_rate": 3.2000000000000003e-06,
      "loss": 1.2688,
      "step": 80
    },
    {
      "epoch": 0.03,
      "grad_norm": 9.25,
      "learning_rate": 3.6000000000000003e-06,
      "loss": 1.2326,
      "step": 90
    },
    {
      "epoch": 0.03,
      "grad_norm": 25.625,
      "learning_rate": 4.000000000000001e-06,
      "loss": 1.1791,
      "step": 100
    },
    {
      "epoch": 0.03,
      "grad_norm": 6.96875,
      "learning_rate": 4.4e-06,
      "loss": 0.9111,
      "step": 110
    },
    {
      "epoch": 0.04,
      "grad_norm": 6.84375,
      "learning_rate": 4.800000000000001e-06,
      "loss": 0.8633,
      "step": 120
    },
    {
      "epoch": 0.04,
      "grad_norm": 7.9375,
      "learning_rate": 5.2e-06,
      "loss": 0.5553,
      "step": 130
    },
    {
      "epoch": 0.04,
      "grad_norm": 14.5625,
      "learning_rate": 5.600000000000001e-06,
      "loss": 0.5916,
      "step": 140
    },
    {
      "epoch": 0.05,
      "grad_norm": 8.8125,
      "learning_rate": 6e-06,
      "loss": 0.4327,
      "step": 150
    },
    {
      "epoch": 0.05,
      "grad_norm": 3.9375,
      "learning_rate": 6.4000000000000006e-06,
      "loss": 0.2799,
      "step": 160
    },
    {
      "epoch": 0.05,
      "grad_norm": 2.34375,
      "learning_rate": 6.800000000000001e-06,
      "loss": 0.2369,
      "step": 170
    },
    {
      "epoch": 0.06,
      "grad_norm": 0.171875,
      "learning_rate": 7.2000000000000005e-06,
      "loss": 0.2997,
      "step": 180
    },
    {
      "epoch": 0.06,
      "grad_norm": 0.287109375,
      "learning_rate": 7.600000000000001e-06,
      "loss": 0.0946,
      "step": 190
    },
    {
      "epoch": 0.06,
      "grad_norm": 18.25,
      "learning_rate": 8.000000000000001e-06,
      "loss": 0.1124,
      "step": 200
    },
    {
      "epoch": 0.07,
      "grad_norm": 1.546875,
      "learning_rate": 8.400000000000001e-06,
      "loss": 0.1131,
      "step": 210
    },
    {
      "epoch": 0.07,
      "grad_norm": 0.71484375,
      "learning_rate": 8.8e-06,
      "loss": 0.0652,
      "step": 220
    },
    {
      "epoch": 0.07,
      "grad_norm": 1.859375,
      "learning_rate": 9.200000000000002e-06,
      "loss": 0.1486,
      "step": 230
    },
    {
      "epoch": 0.08,
      "grad_norm": 1.65625,
      "learning_rate": 9.600000000000001e-06,
      "loss": 0.2169,
      "step": 240
    },
    {
      "epoch": 0.08,
      "grad_norm": 0.1357421875,
      "learning_rate": 1e-05,
      "loss": 0.2408,
      "step": 250
    },
    {
      "epoch": 0.08,
      "grad_norm": 0.055908203125,
      "learning_rate": 1.04e-05,
      "loss": 0.1166,
      "step": 260
    },
    {
      "epoch": 0.08,
      "grad_norm": 5.375,
      "learning_rate": 1.0800000000000002e-05,
      "loss": 0.0114,
      "step": 270
    },
    {
      "epoch": 0.09,
      "grad_norm": 0.0927734375,
      "learning_rate": 1.1200000000000001e-05,
      "loss": 0.1631,
      "step": 280
    },
    {
      "epoch": 0.09,
      "grad_norm": 0.07177734375,
      "learning_rate": 1.16e-05,
      "loss": 0.0603,
      "step": 290
    },
    {
      "epoch": 0.09,
      "grad_norm": 0.482421875,
      "learning_rate": 1.2e-05,
      "loss": 0.1077,
      "step": 300
    },
    {
      "epoch": 0.1,
      "grad_norm": 12.5625,
      "learning_rate": 1.2400000000000002e-05,
      "loss": 0.0832,
      "step": 310
    },
    {
      "epoch": 0.1,
      "grad_norm": 0.048095703125,
      "learning_rate": 1.2800000000000001e-05,
      "loss": 0.0496,
      "step": 320
    },
    {
      "epoch": 0.1,
      "grad_norm": 0.400390625,
      "learning_rate": 1.3200000000000002e-05,
      "loss": 0.1024,
      "step": 330
    },
    {
      "epoch": 0.11,
      "grad_norm": 20.625,
      "learning_rate": 1.3600000000000002e-05,
      "loss": 0.1108,
      "step": 340
    },
    {
      "epoch": 0.11,
      "grad_norm": 0.01611328125,
      "learning_rate": 1.4e-05,
      "loss": 0.0336,
      "step": 350
    },
    {
      "epoch": 0.11,
      "grad_norm": 0.01544189453125,
      "learning_rate": 1.4400000000000001e-05,
      "loss": 0.0165,
      "step": 360
    },
    {
      "epoch": 0.12,
      "grad_norm": 28.0,
      "learning_rate": 1.48e-05,
      "loss": 0.113,
      "step": 370
    },
    {
      "epoch": 0.12,
      "grad_norm": 0.0155029296875,
      "learning_rate": 1.5200000000000002e-05,
      "loss": 0.0221,
      "step": 380
    },
    {
      "epoch": 0.12,
      "grad_norm": 0.0164794921875,
      "learning_rate": 1.5600000000000003e-05,
      "loss": 0.1383,
      "step": 390
    },
    {
      "epoch": 0.13,
      "grad_norm": 1.3203125,
      "learning_rate": 1.6000000000000003e-05,
      "loss": 0.0589,
      "step": 400
    },
    {
      "epoch": 0.13,
      "grad_norm": 0.00909423828125,
      "learning_rate": 1.64e-05,
      "loss": 0.055,
      "step": 410
    },
    {
      "epoch": 0.13,
      "grad_norm": 0.236328125,
      "learning_rate": 1.6800000000000002e-05,
      "loss": 0.0195,
      "step": 420
    },
    {
      "epoch": 0.13,
      "grad_norm": 0.01025390625,
      "learning_rate": 1.72e-05,
      "loss": 0.1477,
      "step": 430
    },
    {
      "epoch": 0.14,
      "grad_norm": 0.007415771484375,
      "learning_rate": 1.76e-05,
      "loss": 0.0708,
      "step": 440
    },
    {
      "epoch": 0.14,
      "grad_norm": 0.01214599609375,
      "learning_rate": 1.8e-05,
      "loss": 0.0808,
      "step": 450
    },
    {
      "epoch": 0.14,
      "grad_norm": 11.3125,
      "learning_rate": 1.8400000000000003e-05,
      "loss": 0.0689,
      "step": 460
    },
    {
      "epoch": 0.15,
      "grad_norm": 0.013427734375,
      "learning_rate": 1.88e-05,
      "loss": 0.1786,
      "step": 470
    },
    {
      "epoch": 0.15,
      "grad_norm": 0.0113525390625,
      "learning_rate": 1.9200000000000003e-05,
      "loss": 0.0897,
      "step": 480
    },
    {
      "epoch": 0.15,
      "grad_norm": 0.2490234375,
      "learning_rate": 1.9600000000000002e-05,
      "loss": 0.0934,
      "step": 490
    },
    {
      "epoch": 0.16,
      "grad_norm": 13.0,
      "learning_rate": 2e-05,
      "loss": 0.1318,
      "step": 500
    },
    {
      "epoch": 0.16,
      "grad_norm": 0.0162353515625,
      "learning_rate": 1.9955555555555557e-05,
      "loss": 0.0156,
      "step": 510
    },
    {
      "epoch": 0.16,
      "grad_norm": 0.0693359375,
      "learning_rate": 1.9911111111111112e-05,
      "loss": 0.067,
      "step": 520
    },
    {
      "epoch": 0.17,
      "grad_norm": 0.007232666015625,
      "learning_rate": 1.9866666666666667e-05,
      "loss": 0.1199,
      "step": 530
    },
    {
      "epoch": 0.17,
      "grad_norm": 1.796875,
      "learning_rate": 1.9822222222222226e-05,
      "loss": 0.0882,
      "step": 540
    },
    {
      "epoch": 0.17,
      "grad_norm": 10.5,
      "learning_rate": 1.977777777777778e-05,
      "loss": 0.1482,
      "step": 550
    },
    {
      "epoch": 0.18,
      "grad_norm": 0.0086669921875,
      "learning_rate": 1.9733333333333336e-05,
      "loss": 0.0937,
      "step": 560
    },
    {
      "epoch": 0.18,
      "grad_norm": 0.328125,
      "learning_rate": 1.968888888888889e-05,
      "loss": 0.0845,
      "step": 570
    },
    {
      "epoch": 0.18,
      "grad_norm": 17.75,
      "learning_rate": 1.9644444444444447e-05,
      "loss": 0.1225,
      "step": 580
    },
    {
      "epoch": 0.18,
      "grad_norm": 0.0068359375,
      "learning_rate": 1.9600000000000002e-05,
      "loss": 0.1117,
      "step": 590
    },
    {
      "epoch": 0.19,
      "grad_norm": 0.00830078125,
      "learning_rate": 1.9555555555555557e-05,
      "loss": 0.1494,
      "step": 600
    },
    {
      "epoch": 0.19,
      "grad_norm": 0.01019287109375,
      "learning_rate": 1.9511111111111113e-05,
      "loss": 0.0575,
      "step": 610
    },
    {
      "epoch": 0.19,
      "grad_norm": 1.5,
      "learning_rate": 1.9466666666666668e-05,
      "loss": 0.0366,
      "step": 620
    },
    {
      "epoch": 0.2,
      "grad_norm": 0.035400390625,
      "learning_rate": 1.9422222222222223e-05,
      "loss": 0.0528,
      "step": 630
    },
    {
      "epoch": 0.2,
      "grad_norm": 0.00860595703125,
      "learning_rate": 1.9377777777777778e-05,
      "loss": 0.0518,
      "step": 640
    },
    {
      "epoch": 0.2,
      "grad_norm": 0.00732421875,
      "learning_rate": 1.9333333333333333e-05,
      "loss": 0.2077,
      "step": 650
    },
    {
      "epoch": 0.21,
      "grad_norm": 3.921875,
      "learning_rate": 1.928888888888889e-05,
      "loss": 0.0559,
      "step": 660
    },
    {
      "epoch": 0.21,
      "grad_norm": 0.01165771484375,
      "learning_rate": 1.9244444444444444e-05,
      "loss": 0.139,
      "step": 670
    },
    {
      "epoch": 0.21,
      "grad_norm": 0.03271484375,
      "learning_rate": 1.9200000000000003e-05,
      "loss": 0.0641,
      "step": 680
    },
    {
      "epoch": 0.22,
      "grad_norm": 0.007232666015625,
      "learning_rate": 1.9155555555555558e-05,
      "loss": 0.0861,
      "step": 690
    },
    {
      "epoch": 0.22,
      "grad_norm": 0.0341796875,
      "learning_rate": 1.9111111111111113e-05,
      "loss": 0.0353,
      "step": 700
    },
    {
      "epoch": 0.22,
      "grad_norm": 6.34375,
      "learning_rate": 1.9066666666666668e-05,
      "loss": 0.0446,
      "step": 710
    },
    {
      "epoch": 0.23,
      "grad_norm": 0.01019287109375,
      "learning_rate": 1.9022222222222223e-05,
      "loss": 0.1009,
      "step": 720
    },
    {
      "epoch": 0.23,
      "grad_norm": 0.007293701171875,
      "learning_rate": 1.897777777777778e-05,
      "loss": 0.0141,
      "step": 730
    },
    {
      "epoch": 0.23,
      "grad_norm": 0.009521484375,
      "learning_rate": 1.8933333333333334e-05,
      "loss": 0.106,
      "step": 740
    },
    {
      "epoch": 0.23,
      "grad_norm": 12.3125,
      "learning_rate": 1.888888888888889e-05,
      "loss": 0.1166,
      "step": 750
    },
    {
      "epoch": 0.24,
      "grad_norm": 12.875,
      "learning_rate": 1.8844444444444444e-05,
      "loss": 0.1105,
      "step": 760
    },
    {
      "epoch": 0.24,
      "grad_norm": 21.25,
      "learning_rate": 1.88e-05,
      "loss": 0.0407,
      "step": 770
    },
    {
      "epoch": 0.24,
      "grad_norm": 18.25,
      "learning_rate": 1.8755555555555558e-05,
      "loss": 0.1127,
      "step": 780
    },
    {
      "epoch": 0.25,
      "grad_norm": 0.032958984375,
      "learning_rate": 1.8711111111111113e-05,
      "loss": 0.022,
      "step": 790
    },
    {
      "epoch": 0.25,
      "grad_norm": 8.6875,
      "learning_rate": 1.866666666666667e-05,
      "loss": 0.1099,
      "step": 800
    },
    {
      "epoch": 0.25,
      "grad_norm": 0.005950927734375,
      "learning_rate": 1.8622222222222224e-05,
      "loss": 0.0456,
      "step": 810
    },
    {
      "epoch": 0.26,
      "grad_norm": 0.005279541015625,
      "learning_rate": 1.857777777777778e-05,
      "loss": 0.086,
      "step": 820
    },
    {
      "epoch": 0.26,
      "grad_norm": 0.03662109375,
      "learning_rate": 1.8533333333333334e-05,
      "loss": 0.0814,
      "step": 830
    },
    {
      "epoch": 0.26,
      "grad_norm": 0.0089111328125,
      "learning_rate": 1.848888888888889e-05,
      "loss": 0.0382,
      "step": 840
    },
    {
      "epoch": 0.27,
      "grad_norm": 292.0,
      "learning_rate": 1.8444444444444448e-05,
      "loss": 0.0244,
      "step": 850
    },
    {
      "epoch": 0.27,
      "grad_norm": 0.01263427734375,
      "learning_rate": 1.8400000000000003e-05,
      "loss": 0.0213,
      "step": 860
    },
    {
      "epoch": 0.27,
      "grad_norm": 0.004608154296875,
      "learning_rate": 1.835555555555556e-05,
      "loss": 0.0438,
      "step": 870
    },
    {
      "epoch": 0.28,
      "grad_norm": 0.005340576171875,
      "learning_rate": 1.8311111111111114e-05,
      "loss": 0.2058,
      "step": 880
    },
    {
      "epoch": 0.28,
      "grad_norm": 0.04931640625,
      "learning_rate": 1.826666666666667e-05,
      "loss": 0.0169,
      "step": 890
    },
    {
      "epoch": 0.28,
      "grad_norm": 9.5,
      "learning_rate": 1.8222222222222224e-05,
      "loss": 0.0693,
      "step": 900
    },
    {
      "epoch": 0.29,
      "grad_norm": 0.00396728515625,
      "learning_rate": 1.817777777777778e-05,
      "loss": 0.0502,
      "step": 910
    },
    {
      "epoch": 0.29,
      "grad_norm": 0.0206298828125,
      "learning_rate": 1.8133333333333335e-05,
      "loss": 0.0479,
      "step": 920
    },
    {
      "epoch": 0.29,
      "grad_norm": 0.046875,
      "learning_rate": 1.808888888888889e-05,
      "loss": 0.0184,
      "step": 930
    },
    {
      "epoch": 0.29,
      "grad_norm": 13.6875,
      "learning_rate": 1.8044444444444445e-05,
      "loss": 0.0315,
      "step": 940
    },
    {
      "epoch": 0.3,
      "grad_norm": 10.0,
      "learning_rate": 1.8e-05,
      "loss": 0.0702,
      "step": 950
    },
    {
      "epoch": 0.3,
      "grad_norm": 7.84375,
      "learning_rate": 1.7955555555555556e-05,
      "loss": 0.162,
      "step": 960
    },
    {
      "epoch": 0.3,
      "grad_norm": 11.25,
      "learning_rate": 1.791111111111111e-05,
      "loss": 0.0772,
      "step": 970
    },
    {
      "epoch": 0.31,
      "grad_norm": 0.0206298828125,
      "learning_rate": 1.7866666666666666e-05,
      "loss": 0.078,
      "step": 980
    },
    {
      "epoch": 0.31,
      "grad_norm": 0.0546875,
      "learning_rate": 1.782222222222222e-05,
      "loss": 0.0804,
      "step": 990
    },
    {
      "epoch": 0.31,
      "grad_norm": 0.053466796875,
      "learning_rate": 1.7777777777777777e-05,
      "loss": 0.0512,
      "step": 1000
    },
    {
      "epoch": 0.31,
      "eval_loss": 0.026573657989501953,
      "eval_runtime": 53.6666,
      "eval_samples_per_second": 18.634,
      "eval_steps_per_second": 18.634,
      "step": 1000
    },
    {
      "epoch": 0.32,
      "grad_norm": 0.0069580078125,
      "learning_rate": 1.7733333333333335e-05,
      "loss": 0.0366,
      "step": 1010
    },
    {
      "epoch": 0.32,
      "grad_norm": 0.00848388671875,
      "learning_rate": 1.768888888888889e-05,
      "loss": 0.0946,
      "step": 1020
    },
    {
      "epoch": 0.32,
      "grad_norm": 0.0133056640625,
      "learning_rate": 1.7644444444444446e-05,
      "loss": 0.138,
      "step": 1030
    },
    {
      "epoch": 0.33,
      "grad_norm": 3.0625,
      "learning_rate": 1.76e-05,
      "loss": 0.0519,
      "step": 1040
    },
    {
      "epoch": 0.33,
      "grad_norm": 9.8125,
      "learning_rate": 1.7555555555555556e-05,
      "loss": 0.0585,
      "step": 1050
    },
    {
      "epoch": 0.33,
      "grad_norm": 0.00537109375,
      "learning_rate": 1.751111111111111e-05,
      "loss": 0.0561,
      "step": 1060
    },
    {
      "epoch": 0.34,
      "grad_norm": 0.0040283203125,
      "learning_rate": 1.7466666666666667e-05,
      "loss": 0.0136,
      "step": 1070
    },
    {
      "epoch": 0.34,
      "grad_norm": 6.03125,
      "learning_rate": 1.7422222222222222e-05,
      "loss": 0.0829,
      "step": 1080
    },
    {
      "epoch": 0.34,
      "grad_norm": 0.92578125,
      "learning_rate": 1.737777777777778e-05,
      "loss": 0.0386,
      "step": 1090
    },
    {
      "epoch": 0.34,
      "grad_norm": 0.0128173828125,
      "learning_rate": 1.7333333333333336e-05,
      "loss": 0.0538,
      "step": 1100
    },
    {
      "epoch": 0.35,
      "grad_norm": 4.0625,
      "learning_rate": 1.728888888888889e-05,
      "loss": 0.0989,
      "step": 1110
    },
    {
      "epoch": 0.35,
      "grad_norm": 0.0106201171875,
      "learning_rate": 1.7244444444444446e-05,
      "loss": 0.0283,
      "step": 1120
    },
    {
      "epoch": 0.35,
      "grad_norm": 21.125,
      "learning_rate": 1.72e-05,
      "loss": 0.0995,
      "step": 1130
    },
    {
      "epoch": 0.36,
      "grad_norm": 13.1875,
      "learning_rate": 1.7155555555555557e-05,
      "loss": 0.07,
      "step": 1140
    },
    {
      "epoch": 0.36,
      "grad_norm": 0.00689697265625,
      "learning_rate": 1.7111111111111112e-05,
      "loss": 0.0948,
      "step": 1150
    },
    {
      "epoch": 0.36,
      "grad_norm": 0.00555419921875,
      "learning_rate": 1.706666666666667e-05,
      "loss": 0.0404,
      "step": 1160
    },
    {
      "epoch": 0.37,
      "grad_norm": 0.0062255859375,
      "learning_rate": 1.7022222222222226e-05,
      "loss": 0.034,
      "step": 1170
    },
    {
      "epoch": 0.37,
      "grad_norm": 0.0037841796875,
      "learning_rate": 1.697777777777778e-05,
      "loss": 0.0375,
      "step": 1180
    },
    {
      "epoch": 0.37,
      "grad_norm": 2.5,
      "learning_rate": 1.6933333333333336e-05,
      "loss": 0.0445,
      "step": 1190
    },
    {
      "epoch": 0.38,
      "grad_norm": 0.00439453125,
      "learning_rate": 1.688888888888889e-05,
      "loss": 0.116,
      "step": 1200
    },
    {
      "epoch": 0.38,
      "grad_norm": 3.09375,
      "learning_rate": 1.6844444444444447e-05,
      "loss": 0.0127,
      "step": 1210
    },
    {
      "epoch": 0.38,
      "grad_norm": 2.1875,
      "learning_rate": 1.6800000000000002e-05,
      "loss": 0.003,
      "step": 1220
    },
    {
      "epoch": 0.39,
      "grad_norm": 0.01165771484375,
      "learning_rate": 1.6755555555555557e-05,
      "loss": 0.0315,
      "step": 1230
    },
    {
      "epoch": 0.39,
      "grad_norm": 0.004302978515625,
      "learning_rate": 1.6711111111111112e-05,
      "loss": 0.057,
      "step": 1240
    },
    {
      "epoch": 0.39,
      "grad_norm": 0.0023651123046875,
      "learning_rate": 1.6666666666666667e-05,
      "loss": 0.0189,
      "step": 1250
    },
    {
      "epoch": 0.39,
      "grad_norm": 0.01080322265625,
      "learning_rate": 1.6622222222222223e-05,
      "loss": 0.0538,
      "step": 1260
    },
    {
      "epoch": 0.4,
      "grad_norm": 4.6875,
      "learning_rate": 1.6577777777777778e-05,
      "loss": 0.0093,
      "step": 1270
    },
    {
      "epoch": 0.4,
      "grad_norm": 0.0032958984375,
      "learning_rate": 1.6533333333333333e-05,
      "loss": 0.0212,
      "step": 1280
    },
    {
      "epoch": 0.4,
      "grad_norm": 0.02685546875,
      "learning_rate": 1.648888888888889e-05,
      "loss": 0.0651,
      "step": 1290
    },
    {
      "epoch": 0.41,
      "grad_norm": 0.0625,
      "learning_rate": 1.6444444444444444e-05,
      "loss": 0.0743,
      "step": 1300
    },
    {
      "epoch": 0.41,
      "grad_norm": 5.8125,
      "learning_rate": 1.64e-05,
      "loss": 0.049,
      "step": 1310
    },
    {
      "epoch": 0.41,
      "grad_norm": 0.005828857421875,
      "learning_rate": 1.6355555555555557e-05,
      "loss": 0.0571,
      "step": 1320
    },
    {
      "epoch": 0.42,
      "grad_norm": 12.375,
      "learning_rate": 1.6311111111111113e-05,
      "loss": 0.0484,
      "step": 1330
    },
    {
      "epoch": 0.42,
      "grad_norm": 14.25,
      "learning_rate": 1.6266666666666668e-05,
      "loss": 0.0679,
      "step": 1340
    },
    {
      "epoch": 0.42,
      "grad_norm": 0.003143310546875,
      "learning_rate": 1.6222222222222223e-05,
      "loss": 0.0008,
      "step": 1350
    },
    {
      "epoch": 0.43,
      "grad_norm": 11.375,
      "learning_rate": 1.617777777777778e-05,
      "loss": 0.0967,
      "step": 1360
    },
    {
      "epoch": 0.43,
      "grad_norm": 0.0031890869140625,
      "learning_rate": 1.6133333333333334e-05,
      "loss": 0.0503,
      "step": 1370
    },
    {
      "epoch": 0.43,
      "grad_norm": 12.0625,
      "learning_rate": 1.608888888888889e-05,
      "loss": 0.0661,
      "step": 1380
    },
    {
      "epoch": 0.44,
      "grad_norm": 0.01019287109375,
      "learning_rate": 1.6044444444444444e-05,
      "loss": 0.0462,
      "step": 1390
    },
    {
      "epoch": 0.44,
      "grad_norm": 0.00653076171875,
      "learning_rate": 1.6000000000000003e-05,
      "loss": 0.0442,
      "step": 1400
    },
    {
      "epoch": 0.44,
      "grad_norm": 0.00262451171875,
      "learning_rate": 1.5955555555555558e-05,
      "loss": 0.0092,
      "step": 1410
    },
    {
      "epoch": 0.44,
      "grad_norm": 9.1875,
      "learning_rate": 1.5911111111111113e-05,
      "loss": 0.0714,
      "step": 1420
    },
    {
      "epoch": 0.45,
      "grad_norm": 8.9375,
      "learning_rate": 1.586666666666667e-05,
      "loss": 0.0202,
      "step": 1430
    },
    {
      "epoch": 0.45,
      "grad_norm": 11.0625,
      "learning_rate": 1.5822222222222224e-05,
      "loss": 0.1116,
      "step": 1440
    },
    {
      "epoch": 0.45,
      "grad_norm": 8.125,
      "learning_rate": 1.577777777777778e-05,
      "loss": 0.1274,
      "step": 1450
    },
    {
      "epoch": 0.46,
      "grad_norm": 1.0078125,
      "learning_rate": 1.5733333333333334e-05,
      "loss": 0.021,
      "step": 1460
    },
    {
      "epoch": 0.46,
      "grad_norm": 0.0030364990234375,
      "learning_rate": 1.5688888888888893e-05,
      "loss": 0.0515,
      "step": 1470
    },
    {
      "epoch": 0.46,
      "grad_norm": 13.375,
      "learning_rate": 1.5644444444444448e-05,
      "loss": 0.1106,
      "step": 1480
    },
    {
      "epoch": 0.47,
      "grad_norm": 0.00885009765625,
      "learning_rate": 1.5600000000000003e-05,
      "loss": 0.0628,
      "step": 1490
    },
    {
      "epoch": 0.47,
      "grad_norm": 2.765625,
      "learning_rate": 1.555555555555556e-05,
      "loss": 0.041,
      "step": 1500
    },
    {
      "epoch": 0.47,
      "grad_norm": 0.0220947265625,
      "learning_rate": 1.5511111111111114e-05,
      "loss": 0.052,
      "step": 1510
    },
    {
      "epoch": 0.48,
      "grad_norm": 3.40625,
      "learning_rate": 1.546666666666667e-05,
      "loss": 0.0283,
      "step": 1520
    },
    {
      "epoch": 0.48,
      "grad_norm": 0.00982666015625,
      "learning_rate": 1.5422222222222224e-05,
      "loss": 0.0568,
      "step": 1530
    },
    {
      "epoch": 0.48,
      "grad_norm": 0.00213623046875,
      "learning_rate": 1.537777777777778e-05,
      "loss": 0.0382,
      "step": 1540
    },
    {
      "epoch": 0.49,
      "grad_norm": 0.0166015625,
      "learning_rate": 1.5333333333333334e-05,
      "loss": 0.0779,
      "step": 1550
    },
    {
      "epoch": 0.49,
      "grad_norm": 5.375,
      "learning_rate": 1.528888888888889e-05,
      "loss": 0.009,
      "step": 1560
    },
    {
      "epoch": 0.49,
      "grad_norm": 0.00193023681640625,
      "learning_rate": 1.5244444444444447e-05,
      "loss": 0.028,
      "step": 1570
    },
    {
      "epoch": 0.5,
      "grad_norm": 0.002105712890625,
      "learning_rate": 1.5200000000000002e-05,
      "loss": 0.0671,
      "step": 1580
    },
    {
      "epoch": 0.5,
      "grad_norm": 0.024169921875,
      "learning_rate": 1.5155555555555557e-05,
      "loss": 0.0574,
      "step": 1590
    },
    {
      "epoch": 0.5,
      "grad_norm": 8.9375,
      "learning_rate": 1.5111111111111112e-05,
      "loss": 0.0782,
      "step": 1600
    },
    {
      "epoch": 0.5,
      "grad_norm": 0.044921875,
      "learning_rate": 1.5066666666666668e-05,
      "loss": 0.0549,
      "step": 1610
    },
    {
      "epoch": 0.51,
      "grad_norm": 0.62890625,
      "learning_rate": 1.5022222222222223e-05,
      "loss": 0.082,
      "step": 1620
    },
    {
      "epoch": 0.51,
      "grad_norm": 0.0032958984375,
      "learning_rate": 1.497777777777778e-05,
      "loss": 0.0751,
      "step": 1630
    },
    {
      "epoch": 0.51,
      "grad_norm": 0.001983642578125,
      "learning_rate": 1.4933333333333335e-05,
      "loss": 0.1232,
      "step": 1640
    },
    {
      "epoch": 0.52,
      "grad_norm": 11.4375,
      "learning_rate": 1.488888888888889e-05,
      "loss": 0.0654,
      "step": 1650
    },
    {
      "epoch": 0.52,
      "grad_norm": 11.4375,
      "learning_rate": 1.4844444444444445e-05,
      "loss": 0.0872,
      "step": 1660
    },
    {
      "epoch": 0.52,
      "grad_norm": 0.0027313232421875,
      "learning_rate": 1.48e-05,
      "loss": 0.0189,
      "step": 1670
    },
    {
      "epoch": 0.53,
      "grad_norm": 12.75,
      "learning_rate": 1.4755555555555556e-05,
      "loss": 0.1015,
      "step": 1680
    },
    {
      "epoch": 0.53,
      "grad_norm": 4.875,
      "learning_rate": 1.4711111111111111e-05,
      "loss": 0.0415,
      "step": 1690
    },
    {
      "epoch": 0.53,
      "grad_norm": 0.00811767578125,
      "learning_rate": 1.4666666666666666e-05,
      "loss": 0.0442,
      "step": 1700
    },
    {
      "epoch": 0.54,
      "grad_norm": 0.9140625,
      "learning_rate": 1.4622222222222225e-05,
      "loss": 0.0741,
      "step": 1710
    },
    {
      "epoch": 0.54,
      "grad_norm": 0.0306396484375,
      "learning_rate": 1.457777777777778e-05,
      "loss": 0.0945,
      "step": 1720
    },
    {
      "epoch": 0.54,
      "grad_norm": 0.443359375,
      "learning_rate": 1.4533333333333335e-05,
      "loss": 0.0565,
      "step": 1730
    },
    {
      "epoch": 0.55,
      "grad_norm": 9.0,
      "learning_rate": 1.448888888888889e-05,
      "loss": 0.0312,
      "step": 1740
    },
    {
      "epoch": 0.55,
      "grad_norm": 0.028564453125,
      "learning_rate": 1.4444444444444446e-05,
      "loss": 0.0108,
      "step": 1750
    },
    {
      "epoch": 0.55,
      "grad_norm": 0.00244140625,
      "learning_rate": 1.4400000000000001e-05,
      "loss": 0.0006,
      "step": 1760
    },
    {
      "epoch": 0.55,
      "grad_norm": 0.00408935546875,
      "learning_rate": 1.4355555555555556e-05,
      "loss": 0.0609,
      "step": 1770
    },
    {
      "epoch": 0.56,
      "grad_norm": 10.75,
      "learning_rate": 1.4311111111111111e-05,
      "loss": 0.0827,
      "step": 1780
    },
    {
      "epoch": 0.56,
      "grad_norm": 3.640625,
      "learning_rate": 1.4266666666666668e-05,
      "loss": 0.0014,
      "step": 1790
    },
    {
      "epoch": 0.56,
      "grad_norm": 23.125,
      "learning_rate": 1.4222222222222224e-05,
      "loss": 0.0715,
      "step": 1800
    },
    {
      "epoch": 0.57,
      "grad_norm": 19.625,
      "learning_rate": 1.4177777777777779e-05,
      "loss": 0.0377,
      "step": 1810
    },
    {
      "epoch": 0.57,
      "grad_norm": 10.8125,
      "learning_rate": 1.4133333333333334e-05,
      "loss": 0.072,
      "step": 1820
    },
    {
      "epoch": 0.57,
      "grad_norm": 0.0022125244140625,
      "learning_rate": 1.408888888888889e-05,
      "loss": 0.0065,
      "step": 1830
    },
    {
      "epoch": 0.58,
      "grad_norm": 0.00177001953125,
      "learning_rate": 1.4044444444444445e-05,
      "loss": 0.0679,
      "step": 1840
    },
    {
      "epoch": 0.58,
      "grad_norm": 0.0037384033203125,
      "learning_rate": 1.4e-05,
      "loss": 0.0588,
      "step": 1850
    },
    {
      "epoch": 0.58,
      "grad_norm": 0.1513671875,
      "learning_rate": 1.3955555555555558e-05,
      "loss": 0.049,
      "step": 1860
    },
    {
      "epoch": 0.59,
      "grad_norm": 0.0042724609375,
      "learning_rate": 1.3911111111111114e-05,
      "loss": 0.1017,
      "step": 1870
    },
    {
      "epoch": 0.59,
      "grad_norm": 0.004150390625,
      "learning_rate": 1.3866666666666669e-05,
      "loss": 0.0498,
      "step": 1880
    },
    {
      "epoch": 0.59,
      "grad_norm": 0.002655029296875,
      "learning_rate": 1.3822222222222224e-05,
      "loss": 0.0259,
      "step": 1890
    },
    {
      "epoch": 0.6,
      "grad_norm": 9.875,
      "learning_rate": 1.377777777777778e-05,
      "loss": 0.0414,
      "step": 1900
    },
    {
      "epoch": 0.6,
      "grad_norm": 7.78125,
      "learning_rate": 1.3733333333333335e-05,
      "loss": 0.0436,
      "step": 1910
    },
    {
      "epoch": 0.6,
      "grad_norm": 0.00250244140625,
      "learning_rate": 1.368888888888889e-05,
      "loss": 0.0468,
      "step": 1920
    },
    {
      "epoch": 0.6,
      "grad_norm": 6.5625,
      "learning_rate": 1.3644444444444445e-05,
      "loss": 0.0382,
      "step": 1930
    },
    {
      "epoch": 0.61,
      "grad_norm": 5.84375,
      "learning_rate": 1.3600000000000002e-05,
      "loss": 0.0076,
      "step": 1940
    },
    {
      "epoch": 0.61,
      "grad_norm": 0.00191497802734375,
      "learning_rate": 1.3555555555555557e-05,
      "loss": 0.027,
      "step": 1950
    },
    {
      "epoch": 0.61,
      "grad_norm": 0.002288818359375,
      "learning_rate": 1.3511111111111112e-05,
      "loss": 0.0724,
      "step": 1960
    },
    {
      "epoch": 0.62,
      "grad_norm": 0.00250244140625,
      "learning_rate": 1.3466666666666668e-05,
      "loss": 0.0323,
      "step": 1970
    },
    {
      "epoch": 0.62,
      "grad_norm": 0.0025177001953125,
      "learning_rate": 1.3422222222222223e-05,
      "loss": 0.0943,
      "step": 1980
    },
    {
      "epoch": 0.62,
      "grad_norm": 0.0093994140625,
      "learning_rate": 1.3377777777777778e-05,
      "loss": 0.0142,
      "step": 1990
    },
    {
      "epoch": 0.63,
      "grad_norm": 0.00162506103515625,
      "learning_rate": 1.3333333333333333e-05,
      "loss": 0.0292,
      "step": 2000
    },
    {
      "epoch": 0.63,
      "eval_loss": 0.02100822888314724,
      "eval_runtime": 53.6726,
      "eval_samples_per_second": 18.631,
      "eval_steps_per_second": 18.631,
      "step": 2000
    },
    {
      "epoch": 0.63,
      "grad_norm": 0.0021820068359375,
      "learning_rate": 1.3288888888888889e-05,
      "loss": 0.0266,
      "step": 2010
    },
    {
      "epoch": 0.63,
      "grad_norm": 0.022216796875,
      "learning_rate": 1.3244444444444447e-05,
      "loss": 0.0234,
      "step": 2020
    },
    {
      "epoch": 0.64,
      "grad_norm": 6.8125,
      "learning_rate": 1.3200000000000002e-05,
      "loss": 0.0527,
      "step": 2030
    },
    {
      "epoch": 0.64,
      "grad_norm": 0.0016632080078125,
      "learning_rate": 1.3155555555555558e-05,
      "loss": 0.0001,
      "step": 2040
    },
    {
      "epoch": 0.64,
      "grad_norm": 0.234375,
      "learning_rate": 1.3111111111111113e-05,
      "loss": 0.0284,
      "step": 2050
    },
    {
      "epoch": 0.65,
      "grad_norm": 0.001708984375,
      "learning_rate": 1.3066666666666668e-05,
      "loss": 0.0677,
      "step": 2060
    },
    {
      "epoch": 0.65,
      "grad_norm": 4.90625,
      "learning_rate": 1.3022222222222223e-05,
      "loss": 0.0746,
      "step": 2070
    },
    {
      "epoch": 0.65,
      "grad_norm": 0.0027008056640625,
      "learning_rate": 1.2977777777777779e-05,
      "loss": 0.0461,
      "step": 2080
    },
    {
      "epoch": 0.65,
      "grad_norm": 4.75,
      "learning_rate": 1.2933333333333334e-05,
      "loss": 0.0427,
      "step": 2090
    },
    {
      "epoch": 0.66,
      "grad_norm": 5.625,
      "learning_rate": 1.288888888888889e-05,
      "loss": 0.0952,
      "step": 2100
    },
    {
      "epoch": 0.66,
      "grad_norm": 0.0031890869140625,
      "learning_rate": 1.2844444444444446e-05,
      "loss": 0.0388,
      "step": 2110
    },
    {
      "epoch": 0.66,
      "grad_norm": 0.012939453125,
      "learning_rate": 1.2800000000000001e-05,
      "loss": 0.0518,
      "step": 2120
    },
    {
      "epoch": 0.67,
      "grad_norm": 0.002288818359375,
      "learning_rate": 1.2755555555555556e-05,
      "loss": 0.0452,
      "step": 2130
    },
    {
      "epoch": 0.67,
      "grad_norm": 0.0019989013671875,
      "learning_rate": 1.2711111111111112e-05,
      "loss": 0.0421,
      "step": 2140
    },
    {
      "epoch": 0.67,
      "grad_norm": 0.00439453125,
      "learning_rate": 1.2666666666666667e-05,
      "loss": 0.0824,
      "step": 2150
    },
    {
      "epoch": 0.68,
      "grad_norm": 0.007171630859375,
      "learning_rate": 1.2622222222222222e-05,
      "loss": 0.0774,
      "step": 2160
    },
    {
      "epoch": 0.68,
      "grad_norm": 0.00160980224609375,
      "learning_rate": 1.257777777777778e-05,
      "loss": 0.0417,
      "step": 2170
    },
    {
      "epoch": 0.68,
      "grad_norm": 0.609375,
      "learning_rate": 1.2533333333333336e-05,
      "loss": 0.0208,
      "step": 2180
    },
    {
      "epoch": 0.69,
      "grad_norm": 0.00156402587890625,
      "learning_rate": 1.2488888888888891e-05,
      "loss": 0.1365,
      "step": 2190
    },
    {
      "epoch": 0.69,
      "grad_norm": 0.0016021728515625,
      "learning_rate": 1.2444444444444446e-05,
      "loss": 0.0498,
      "step": 2200
    },
    {
      "epoch": 0.69,
      "grad_norm": 0.0013885498046875,
      "learning_rate": 1.2400000000000002e-05,
      "loss": 0.0679,
      "step": 2210
    },
    {
      "epoch": 0.7,
      "grad_norm": 4.40625,
      "learning_rate": 1.2355555555555557e-05,
      "loss": 0.0176,
      "step": 2220
    },
    {
      "epoch": 0.7,
      "grad_norm": 0.00153350830078125,
      "learning_rate": 1.2311111111111112e-05,
      "loss": 0.0328,
      "step": 2230
    },
    {
      "epoch": 0.7,
      "grad_norm": 11.6875,
      "learning_rate": 1.2266666666666667e-05,
      "loss": 0.1294,
      "step": 2240
    },
    {
      "epoch": 0.7,
      "grad_norm": 8.4375,
      "learning_rate": 1.2222222222222224e-05,
      "loss": 0.0423,
      "step": 2250
    },
    {
      "epoch": 0.71,
      "grad_norm": 0.007720947265625,
      "learning_rate": 1.217777777777778e-05,
      "loss": 0.0088,
      "step": 2260
    },
    {
      "epoch": 0.71,
      "grad_norm": 11.625,
      "learning_rate": 1.2133333333333335e-05,
      "loss": 0.0754,
      "step": 2270
    },
    {
      "epoch": 0.71,
      "grad_norm": 0.00177001953125,
      "learning_rate": 1.208888888888889e-05,
      "loss": 0.0148,
      "step": 2280
    },
    {
      "epoch": 0.72,
      "grad_norm": 0.00142669677734375,
      "learning_rate": 1.2044444444444445e-05,
      "loss": 0.0322,
      "step": 2290
    },
    {
      "epoch": 0.72,
      "grad_norm": 0.0019683837890625,
      "learning_rate": 1.2e-05,
      "loss": 0.067,
      "step": 2300
    },
    {
      "epoch": 0.72,
      "grad_norm": 0.001708984375,
      "learning_rate": 1.1955555555555556e-05,
      "loss": 0.0345,
      "step": 2310
    },
    {
      "epoch": 0.73,
      "grad_norm": 0.004852294921875,
      "learning_rate": 1.191111111111111e-05,
      "loss": 0.0604,
      "step": 2320
    },
    {
      "epoch": 0.73,
      "grad_norm": 0.0029296875,
      "learning_rate": 1.186666666666667e-05,
      "loss": 0.0611,
      "step": 2330
    },
    {
      "epoch": 0.73,
      "grad_norm": 1.8984375,
      "learning_rate": 1.1822222222222225e-05,
      "loss": 0.0205,
      "step": 2340
    },
    {
      "epoch": 0.74,
      "grad_norm": 1.9296875,
      "learning_rate": 1.177777777777778e-05,
      "loss": 0.0836,
      "step": 2350
    },
    {
      "epoch": 0.74,
      "grad_norm": 0.002685546875,
      "learning_rate": 1.1733333333333335e-05,
      "loss": 0.0416,
      "step": 2360
    },
    {
      "epoch": 0.74,
      "grad_norm": 24.375,
      "learning_rate": 1.168888888888889e-05,
      "loss": 0.0652,
      "step": 2370
    },
    {
      "epoch": 0.75,
      "grad_norm": 0.002777099609375,
      "learning_rate": 1.1644444444444446e-05,
      "loss": 0.0372,
      "step": 2380
    },
    {
      "epoch": 0.75,
      "grad_norm": 0.0022125244140625,
      "learning_rate": 1.16e-05,
      "loss": 0.05,
      "step": 2390
    },
    {
      "epoch": 0.75,
      "grad_norm": 0.00193023681640625,
      "learning_rate": 1.1555555555555556e-05,
      "loss": 0.0199,
      "step": 2400
    },
    {
      "epoch": 0.76,
      "grad_norm": 18.25,
      "learning_rate": 1.1511111111111113e-05,
      "loss": 0.0264,
      "step": 2410
    },
    {
      "epoch": 0.76,
      "grad_norm": 0.00185394287109375,
      "learning_rate": 1.1466666666666668e-05,
      "loss": 0.0507,
      "step": 2420
    },
    {
      "epoch": 0.76,
      "grad_norm": 0.0027008056640625,
      "learning_rate": 1.1422222222222223e-05,
      "loss": 0.032,
      "step": 2430
    },
    {
      "epoch": 0.76,
      "grad_norm": 16.5,
      "learning_rate": 1.1377777777777779e-05,
      "loss": 0.1437,
      "step": 2440
    },
    {
      "epoch": 0.77,
      "grad_norm": 9.0625,
      "learning_rate": 1.1333333333333334e-05,
      "loss": 0.0642,
      "step": 2450
    },
    {
      "epoch": 0.77,
      "grad_norm": 16.125,
      "learning_rate": 1.1288888888888889e-05,
      "loss": 0.0265,
      "step": 2460
    },
    {
      "epoch": 0.77,
      "grad_norm": 0.6484375,
      "learning_rate": 1.1244444444444444e-05,
      "loss": 0.0216,
      "step": 2470
    },
    {
      "epoch": 0.78,
      "grad_norm": 0.00168609619140625,
      "learning_rate": 1.1200000000000001e-05,
      "loss": 0.0574,
      "step": 2480
    },
    {
      "epoch": 0.78,
      "grad_norm": 4.15625,
      "learning_rate": 1.1155555555555556e-05,
      "loss": 0.0344,
      "step": 2490
    },
    {
      "epoch": 0.78,
      "grad_norm": 0.00142669677734375,
      "learning_rate": 1.1111111111111113e-05,
      "loss": 0.0645,
      "step": 2500
    },
    {
      "epoch": 0.79,
      "grad_norm": 0.003387451171875,
      "learning_rate": 1.1066666666666669e-05,
      "loss": 0.0654,
      "step": 2510
    },
    {
      "epoch": 0.79,
      "grad_norm": 0.005523681640625,
      "learning_rate": 1.1022222222222224e-05,
      "loss": 0.0329,
      "step": 2520
    },
    {
      "epoch": 0.79,
      "grad_norm": 10.6875,
      "learning_rate": 1.0977777777777779e-05,
      "loss": 0.1523,
      "step": 2530
    },
    {
      "epoch": 0.8,
      "grad_norm": 0.0018310546875,
      "learning_rate": 1.0933333333333334e-05,
      "loss": 0.0013,
      "step": 2540
    },
    {
      "epoch": 0.8,
      "grad_norm": 0.00225830078125,
      "learning_rate": 1.088888888888889e-05,
      "loss": 0.0073,
      "step": 2550
    },
    {
      "epoch": 0.8,
      "grad_norm": 0.00180816650390625,
      "learning_rate": 1.0844444444444446e-05,
      "loss": 0.03,
      "step": 2560
    },
    {
      "epoch": 0.81,
      "grad_norm": 8.125,
      "learning_rate": 1.0800000000000002e-05,
      "loss": 0.0208,
      "step": 2570
    },
    {
      "epoch": 0.81,
      "grad_norm": 14.6875,
      "learning_rate": 1.0755555555555557e-05,
      "loss": 0.0357,
      "step": 2580
    },
    {
      "epoch": 0.81,
      "grad_norm": 15.5,
      "learning_rate": 1.0711111111111112e-05,
      "loss": 0.0667,
      "step": 2590
    },
    {
      "epoch": 0.81,
      "grad_norm": 0.00140380859375,
      "learning_rate": 1.0666666666666667e-05,
      "loss": 0.0441,
      "step": 2600
    },
    {
      "epoch": 0.82,
      "grad_norm": 0.00180816650390625,
      "learning_rate": 1.0622222222222223e-05,
      "loss": 0.0164,
      "step": 2610
    },
    {
      "epoch": 0.82,
      "grad_norm": 0.003631591796875,
      "learning_rate": 1.0577777777777778e-05,
      "loss": 0.0373,
      "step": 2620
    },
    {
      "epoch": 0.82,
      "grad_norm": 0.0233154296875,
      "learning_rate": 1.0533333333333333e-05,
      "loss": 0.064,
      "step": 2630
    },
    {
      "epoch": 0.83,
      "grad_norm": 0.00152587890625,
      "learning_rate": 1.048888888888889e-05,
      "loss": 0.0459,
      "step": 2640
    },
    {
      "epoch": 0.83,
      "grad_norm": 14.125,
      "learning_rate": 1.0444444444444445e-05,
      "loss": 0.0832,
      "step": 2650
    },
    {
      "epoch": 0.83,
      "grad_norm": 11.375,
      "learning_rate": 1.04e-05,
      "loss": 0.0607,
      "step": 2660
    },
    {
      "epoch": 0.84,
      "grad_norm": 0.095703125,
      "learning_rate": 1.0355555555555557e-05,
      "loss": 0.0198,
      "step": 2670
    },
    {
      "epoch": 0.84,
      "grad_norm": 0.0022430419921875,
      "learning_rate": 1.0311111111111113e-05,
      "loss": 0.0521,
      "step": 2680
    },
    {
      "epoch": 0.84,
      "grad_norm": 0.0016326904296875,
      "learning_rate": 1.0266666666666668e-05,
      "loss": 0.0647,
      "step": 2690
    },
    {
      "epoch": 0.85,
      "grad_norm": 0.51171875,
      "learning_rate": 1.0222222222222223e-05,
      "loss": 0.0028,
      "step": 2700
    },
    {
      "epoch": 0.85,
      "grad_norm": 5.53125,
      "learning_rate": 1.0177777777777778e-05,
      "loss": 0.045,
      "step": 2710
    },
    {
      "epoch": 0.85,
      "grad_norm": 4.84375,
      "learning_rate": 1.0133333333333335e-05,
      "loss": 0.0326,
      "step": 2720
    },
    {
      "epoch": 0.86,
      "grad_norm": 0.026611328125,
      "learning_rate": 1.008888888888889e-05,
      "loss": 0.0315,
      "step": 2730
    },
    {
      "epoch": 0.86,
      "grad_norm": 0.001556396484375,
      "learning_rate": 1.0044444444444446e-05,
      "loss": 0.0117,
      "step": 2740
    },
    {
      "epoch": 0.86,
      "grad_norm": 0.005767822265625,
      "learning_rate": 1e-05,
      "loss": 0.0485,
      "step": 2750
    },
    {
      "epoch": 0.86,
      "grad_norm": 0.005889892578125,
      "learning_rate": 9.955555555555556e-06,
      "loss": 0.0631,
      "step": 2760
    },
    {
      "epoch": 0.87,
      "grad_norm": 0.0016021728515625,
      "learning_rate": 9.911111111111113e-06,
      "loss": 0.0489,
      "step": 2770
    },
    {
      "epoch": 0.87,
      "grad_norm": 0.00189208984375,
      "learning_rate": 9.866666666666668e-06,
      "loss": 0.0797,
      "step": 2780
    },
    {
      "epoch": 0.87,
      "grad_norm": 0.00347900390625,
      "learning_rate": 9.822222222222223e-06,
      "loss": 0.0224,
      "step": 2790
    },
    {
      "epoch": 0.88,
      "grad_norm": 0.0020599365234375,
      "learning_rate": 9.777777777777779e-06,
      "loss": 0.0776,
      "step": 2800
    },
    {
      "epoch": 0.88,
      "grad_norm": 0.00156402587890625,
      "learning_rate": 9.733333333333334e-06,
      "loss": 0.0961,
      "step": 2810
    },
    {
      "epoch": 0.88,
      "grad_norm": 4.125,
      "learning_rate": 9.688888888888889e-06,
      "loss": 0.0648,
      "step": 2820
    },
    {
      "epoch": 0.89,
      "grad_norm": 2.84375,
      "learning_rate": 9.644444444444444e-06,
      "loss": 0.0402,
      "step": 2830
    },
    {
      "epoch": 0.89,
      "grad_norm": 2.890625,
      "learning_rate": 9.600000000000001e-06,
      "loss": 0.0605,
      "step": 2840
    },
    {
      "epoch": 0.89,
      "grad_norm": 0.0015411376953125,
      "learning_rate": 9.555555555555556e-06,
      "loss": 0.0229,
      "step": 2850
    },
    {
      "epoch": 0.9,
      "grad_norm": 0.0026092529296875,
      "learning_rate": 9.511111111111112e-06,
      "loss": 0.0494,
      "step": 2860
    },
    {
      "epoch": 0.9,
      "grad_norm": 0.00732421875,
      "learning_rate": 9.466666666666667e-06,
      "loss": 0.1074,
      "step": 2870
    },
    {
      "epoch": 0.9,
      "grad_norm": 0.007080078125,
      "learning_rate": 9.422222222222222e-06,
      "loss": 0.0181,
      "step": 2880
    },
    {
      "epoch": 0.91,
      "grad_norm": 0.0040283203125,
      "learning_rate": 9.377777777777779e-06,
      "loss": 0.0481,
      "step": 2890
    },
    {
      "epoch": 0.91,
      "grad_norm": 0.08154296875,
      "learning_rate": 9.333333333333334e-06,
      "loss": 0.0291,
      "step": 2900
    },
    {
      "epoch": 0.91,
      "grad_norm": 0.001861572265625,
      "learning_rate": 9.28888888888889e-06,
      "loss": 0.0827,
      "step": 2910
    },
    {
      "epoch": 0.91,
      "grad_norm": 0.0017852783203125,
      "learning_rate": 9.244444444444445e-06,
      "loss": 0.019,
      "step": 2920
    },
    {
      "epoch": 0.92,
      "grad_norm": 0.00726318359375,
      "learning_rate": 9.200000000000002e-06,
      "loss": 0.0496,
      "step": 2930
    },
    {
      "epoch": 0.92,
      "grad_norm": 0.002044677734375,
      "learning_rate": 9.155555555555557e-06,
      "loss": 0.0302,
      "step": 2940
    },
    {
      "epoch": 0.92,
      "grad_norm": 0.00250244140625,
      "learning_rate": 9.111111111111112e-06,
      "loss": 0.015,
      "step": 2950
    },
    {
      "epoch": 0.93,
      "grad_norm": 0.00193023681640625,
      "learning_rate": 9.066666666666667e-06,
      "loss": 0.0422,
      "step": 2960
    },
    {
      "epoch": 0.93,
      "grad_norm": 6.1875,
      "learning_rate": 9.022222222222223e-06,
      "loss": 0.0594,
      "step": 2970
    },
    {
      "epoch": 0.93,
      "grad_norm": 0.00128936767578125,
      "learning_rate": 8.977777777777778e-06,
      "loss": 0.0027,
      "step": 2980
    },
    {
      "epoch": 0.94,
      "grad_norm": 0.00128936767578125,
      "learning_rate": 8.933333333333333e-06,
      "loss": 0.0445,
      "step": 2990
    },
    {
      "epoch": 0.94,
      "grad_norm": 0.00146484375,
      "learning_rate": 8.888888888888888e-06,
      "loss": 0.0358,
      "step": 3000
    },
    {
      "epoch": 0.94,
      "eval_loss": 0.01655612699687481,
      "eval_runtime": 53.3483,
      "eval_samples_per_second": 18.745,
      "eval_steps_per_second": 18.745,
      "step": 3000
    },
    {
      "epoch": 0.94,
      "grad_norm": 10.6875,
      "learning_rate": 8.844444444444445e-06,
      "loss": 0.137,
      "step": 3010
    },
    {
      "epoch": 0.95,
      "grad_norm": 0.00127410888671875,
      "learning_rate": 8.8e-06,
      "loss": 0.0456,
      "step": 3020
    },
    {
      "epoch": 0.95,
      "grad_norm": 0.003204345703125,
      "learning_rate": 8.755555555555556e-06,
      "loss": 0.0133,
      "step": 3030
    },
    {
      "epoch": 0.95,
      "grad_norm": 8.625,
      "learning_rate": 8.711111111111111e-06,
      "loss": 0.1079,
      "step": 3040
    },
    {
      "epoch": 0.96,
      "grad_norm": 0.004791259765625,
      "learning_rate": 8.666666666666668e-06,
      "loss": 0.0327,
      "step": 3050
    },
    {
      "epoch": 0.96,
      "grad_norm": 0.004150390625,
      "learning_rate": 8.622222222222223e-06,
      "loss": 0.0525,
      "step": 3060
    },
    {
      "epoch": 0.96,
      "grad_norm": 8.9375,
      "learning_rate": 8.577777777777778e-06,
      "loss": 0.0584,
      "step": 3070
    },
    {
      "epoch": 0.96,
      "grad_norm": 0.00131988525390625,
      "learning_rate": 8.533333333333335e-06,
      "loss": 0.0406,
      "step": 3080
    },
    {
      "epoch": 0.97,
      "grad_norm": 14.0,
      "learning_rate": 8.48888888888889e-06,
      "loss": 0.0195,
      "step": 3090
    },
    {
      "epoch": 0.97,
      "grad_norm": 0.001373291015625,
      "learning_rate": 8.444444444444446e-06,
      "loss": 0.0571,
      "step": 3100
    },
    {
      "epoch": 0.97,
      "grad_norm": 0.00469970703125,
      "learning_rate": 8.400000000000001e-06,
      "loss": 0.1144,
      "step": 3110
    },
    {
      "epoch": 0.98,
      "grad_norm": 0.03759765625,
      "learning_rate": 8.355555555555556e-06,
      "loss": 0.0206,
      "step": 3120
    },
    {
      "epoch": 0.98,
      "grad_norm": 0.0027618408203125,
      "learning_rate": 8.311111111111111e-06,
      "loss": 0.0463,
      "step": 3130
    },
    {
      "epoch": 0.98,
      "grad_norm": 0.00156402587890625,
      "learning_rate": 8.266666666666667e-06,
      "loss": 0.079,
      "step": 3140
    },
    {
      "epoch": 0.99,
      "grad_norm": 12.875,
      "learning_rate": 8.222222222222222e-06,
      "loss": 0.0177,
      "step": 3150
    },
    {
      "epoch": 0.99,
      "grad_norm": 0.004547119140625,
      "learning_rate": 8.177777777777779e-06,
      "loss": 0.0287,
      "step": 3160
    },
    {
      "epoch": 0.99,
      "grad_norm": 0.00136566162109375,
      "learning_rate": 8.133333333333334e-06,
      "loss": 0.0063,
      "step": 3170
    },
    {
      "epoch": 1.0,
      "grad_norm": 0.00140380859375,
      "learning_rate": 8.08888888888889e-06,
      "loss": 0.0575,
      "step": 3180
    },
    {
      "epoch": 1.0,
      "grad_norm": 0.00157928466796875,
      "learning_rate": 8.044444444444444e-06,
      "loss": 0.0281,
      "step": 3190
    },
    {
      "epoch": 1.0,
      "grad_norm": 4.34375,
      "learning_rate": 8.000000000000001e-06,
      "loss": 0.0165,
      "step": 3200
    },
    {
      "epoch": 1.01,
      "grad_norm": 1.2109375,
      "learning_rate": 7.955555555555557e-06,
      "loss": 0.0136,
      "step": 3210
    },
    {
      "epoch": 1.01,
      "grad_norm": 28.875,
      "learning_rate": 7.911111111111112e-06,
      "loss": 0.0679,
      "step": 3220
    },
    {
      "epoch": 1.01,
      "grad_norm": 0.001251220703125,
      "learning_rate": 7.866666666666667e-06,
      "loss": 0.0122,
      "step": 3230
    },
    {
      "epoch": 1.02,
      "grad_norm": 3.703125,
      "learning_rate": 7.822222222222224e-06,
      "loss": 0.0085,
      "step": 3240
    },
    {
      "epoch": 1.02,
      "grad_norm": 0.00482177734375,
      "learning_rate": 7.77777777777778e-06,
      "loss": 0.0195,
      "step": 3250
    },
    {
      "epoch": 1.02,
      "grad_norm": 0.06396484375,
      "learning_rate": 7.733333333333334e-06,
      "loss": 0.0234,
      "step": 3260
    },
    {
      "epoch": 1.02,
      "grad_norm": 4.71875,
      "learning_rate": 7.68888888888889e-06,
      "loss": 0.0451,
      "step": 3270
    },
    {
      "epoch": 1.03,
      "grad_norm": 0.00148773193359375,
      "learning_rate": 7.644444444444445e-06,
      "loss": 0.0098,
      "step": 3280
    },
    {
      "epoch": 1.03,
      "grad_norm": 0.002532958984375,
      "learning_rate": 7.600000000000001e-06,
      "loss": 0.0034,
      "step": 3290
    },
    {
      "epoch": 1.03,
      "grad_norm": 0.0027923583984375,
      "learning_rate": 7.555555555555556e-06,
      "loss": 0.01,
      "step": 3300
    },
    {
      "epoch": 1.04,
      "grad_norm": 3.5,
      "learning_rate": 7.511111111111111e-06,
      "loss": 0.0269,
      "step": 3310
    },
    {
      "epoch": 1.04,
      "grad_norm": 0.00124359130859375,
      "learning_rate": 7.4666666666666675e-06,
      "loss": 0.018,
      "step": 3320
    },
    {
      "epoch": 1.04,
      "grad_norm": 8.75,
      "learning_rate": 7.422222222222223e-06,
      "loss": 0.1035,
      "step": 3330
    },
    {
      "epoch": 1.05,
      "grad_norm": 13.875,
      "learning_rate": 7.377777777777778e-06,
      "loss": 0.0368,
      "step": 3340
    },
    {
      "epoch": 1.05,
      "grad_norm": 0.004608154296875,
      "learning_rate": 7.333333333333333e-06,
      "loss": 0.0243,
      "step": 3350
    },
    {
      "epoch": 1.05,
      "grad_norm": 0.00146484375,
      "learning_rate": 7.28888888888889e-06,
      "loss": 0.0177,
      "step": 3360
    },
    {
      "epoch": 1.06,
      "grad_norm": 0.0014190673828125,
      "learning_rate": 7.244444444444445e-06,
      "loss": 0.0426,
      "step": 3370
    },
    {
      "epoch": 1.06,
      "grad_norm": 0.004913330078125,
      "learning_rate": 7.2000000000000005e-06,
      "loss": 0.0506,
      "step": 3380
    },
    {
      "epoch": 1.06,
      "grad_norm": 9.125,
      "learning_rate": 7.155555555555556e-06,
      "loss": 0.0113,
      "step": 3390
    },
    {
      "epoch": 1.07,
      "grad_norm": 0.0015716552734375,
      "learning_rate": 7.111111111111112e-06,
      "loss": 0.0075,
      "step": 3400
    },
    {
      "epoch": 1.07,
      "grad_norm": 9.9375,
      "learning_rate": 7.066666666666667e-06,
      "loss": 0.0188,
      "step": 3410
    },
    {
      "epoch": 1.07,
      "grad_norm": 0.001312255859375,
      "learning_rate": 7.022222222222222e-06,
      "loss": 0.0017,
      "step": 3420
    },
    {
      "epoch": 1.07,
      "grad_norm": 0.001251220703125,
      "learning_rate": 6.977777777777779e-06,
      "loss": 0.0339,
      "step": 3430
    },
    {
      "epoch": 1.08,
      "grad_norm": 0.00262451171875,
      "learning_rate": 6.9333333333333344e-06,
      "loss": 0.0326,
      "step": 3440
    },
    {
      "epoch": 1.08,
      "grad_norm": 0.004547119140625,
      "learning_rate": 6.88888888888889e-06,
      "loss": 0.0145,
      "step": 3450
    },
    {
      "epoch": 1.08,
      "grad_norm": 0.00107574462890625,
      "learning_rate": 6.844444444444445e-06,
      "loss": 0.0126,
      "step": 3460
    },
    {
      "epoch": 1.09,
      "grad_norm": 0.875,
      "learning_rate": 6.800000000000001e-06,
      "loss": 0.0141,
      "step": 3470
    },
    {
      "epoch": 1.09,
      "grad_norm": 7.59375,
      "learning_rate": 6.755555555555556e-06,
      "loss": 0.0301,
      "step": 3480
    },
    {
      "epoch": 1.09,
      "grad_norm": 7.75,
      "learning_rate": 6.711111111111111e-06,
      "loss": 0.0432,
      "step": 3490
    },
    {
      "epoch": 1.1,
      "grad_norm": 0.0010528564453125,
      "learning_rate": 6.666666666666667e-06,
      "loss": 0.0335,
      "step": 3500
    },
    {
      "epoch": 1.1,
      "grad_norm": 0.0048828125,
      "learning_rate": 6.6222222222222236e-06,
      "loss": 0.0616,
      "step": 3510
    },
    {
      "epoch": 1.1,
      "grad_norm": 0.001312255859375,
      "learning_rate": 6.577777777777779e-06,
      "loss": 0.0144,
      "step": 3520
    },
    {
      "epoch": 1.11,
      "grad_norm": 2.5,
      "learning_rate": 6.533333333333334e-06,
      "loss": 0.0415,
      "step": 3530
    },
    {
      "epoch": 1.11,
      "grad_norm": 0.004730224609375,
      "learning_rate": 6.488888888888889e-06,
      "loss": 0.0315,
      "step": 3540
    },
    {
      "epoch": 1.11,
      "grad_norm": 0.00128173828125,
      "learning_rate": 6.444444444444445e-06,
      "loss": 0.0051,
      "step": 3550
    },
    {
      "epoch": 1.12,
      "grad_norm": 0.00115966796875,
      "learning_rate": 6.4000000000000006e-06,
      "loss": 0.0411,
      "step": 3560
    },
    {
      "epoch": 1.12,
      "grad_norm": 0.0010528564453125,
      "learning_rate": 6.355555555555556e-06,
      "loss": 0.0271,
      "step": 3570
    },
    {
      "epoch": 1.12,
      "grad_norm": 0.00482177734375,
      "learning_rate": 6.311111111111111e-06,
      "loss": 0.0172,
      "step": 3580
    },
    {
      "epoch": 1.12,
      "grad_norm": 0.0027008056640625,
      "learning_rate": 6.266666666666668e-06,
      "loss": 0.0062,
      "step": 3590
    },
    {
      "epoch": 1.13,
      "grad_norm": 6.84375,
      "learning_rate": 6.222222222222223e-06,
      "loss": 0.0234,
      "step": 3600
    },
    {
      "epoch": 1.13,
      "grad_norm": 0.0023345947265625,
      "learning_rate": 6.177777777777778e-06,
      "loss": 0.0,
      "step": 3610
    },
    {
      "epoch": 1.13,
      "grad_norm": 0.00115966796875,
      "learning_rate": 6.133333333333334e-06,
      "loss": 0.0194,
      "step": 3620
    },
    {
      "epoch": 1.14,
      "grad_norm": 0.1376953125,
      "learning_rate": 6.08888888888889e-06,
      "loss": 0.0088,
      "step": 3630
    },
    {
      "epoch": 1.14,
      "grad_norm": 0.203125,
      "learning_rate": 6.044444444444445e-06,
      "loss": 0.0339,
      "step": 3640
    },
    {
      "epoch": 1.14,
      "grad_norm": 15.25,
      "learning_rate": 6e-06,
      "loss": 0.0362,
      "step": 3650
    },
    {
      "epoch": 1.15,
      "grad_norm": 4.375,
      "learning_rate": 5.955555555555555e-06,
      "loss": 0.0335,
      "step": 3660
    },
    {
      "epoch": 1.15,
      "grad_norm": 3.5,
      "learning_rate": 5.911111111111112e-06,
      "loss": 0.0254,
      "step": 3670
    },
    {
      "epoch": 1.15,
      "grad_norm": 0.000949859619140625,
      "learning_rate": 5.8666666666666675e-06,
      "loss": 0.0173,
      "step": 3680
    },
    {
      "epoch": 1.16,
      "grad_norm": 5.125,
      "learning_rate": 5.822222222222223e-06,
      "loss": 0.0677,
      "step": 3690
    },
    {
      "epoch": 1.16,
      "grad_norm": 9.75,
      "learning_rate": 5.777777777777778e-06,
      "loss": 0.0162,
      "step": 3700
    },
    {
      "epoch": 1.16,
      "grad_norm": 0.002288818359375,
      "learning_rate": 5.733333333333334e-06,
      "loss": 0.0334,
      "step": 3710
    },
    {
      "epoch": 1.17,
      "grad_norm": 0.0230712890625,
      "learning_rate": 5.688888888888889e-06,
      "loss": 0.0673,
      "step": 3720
    },
    {
      "epoch": 1.17,
      "grad_norm": 0.0010986328125,
      "learning_rate": 5.6444444444444445e-06,
      "loss": 0.0189,
      "step": 3730
    },
    {
      "epoch": 1.17,
      "grad_norm": 0.00095367431640625,
      "learning_rate": 5.600000000000001e-06,
      "loss": 0.0493,
      "step": 3740
    },
    {
      "epoch": 1.17,
      "grad_norm": 0.0093994140625,
      "learning_rate": 5.555555555555557e-06,
      "loss": 0.032,
      "step": 3750
    },
    {
      "epoch": 1.18,
      "grad_norm": 0.00119781494140625,
      "learning_rate": 5.511111111111112e-06,
      "loss": 0.0201,
      "step": 3760
    },
    {
      "epoch": 1.18,
      "grad_norm": 0.00225830078125,
      "learning_rate": 5.466666666666667e-06,
      "loss": 0.0928,
      "step": 3770
    },
    {
      "epoch": 1.18,
      "grad_norm": 0.0045166015625,
      "learning_rate": 5.422222222222223e-06,
      "loss": 0.0048,
      "step": 3780
    },
    {
      "epoch": 1.19,
      "grad_norm": 0.1513671875,
      "learning_rate": 5.3777777777777784e-06,
      "loss": 0.023,
      "step": 3790
    },
    {
      "epoch": 1.19,
      "grad_norm": 2.875,
      "learning_rate": 5.333333333333334e-06,
      "loss": 0.0134,
      "step": 3800
    },
    {
      "epoch": 1.19,
      "grad_norm": 0.00146484375,
      "learning_rate": 5.288888888888889e-06,
      "loss": 0.0063,
      "step": 3810
    },
    {
      "epoch": 1.2,
      "grad_norm": 0.001007080078125,
      "learning_rate": 5.244444444444445e-06,
      "loss": 0.0241,
      "step": 3820
    },
    {
      "epoch": 1.2,
      "grad_norm": 0.01483154296875,
      "learning_rate": 5.2e-06,
      "loss": 0.0124,
      "step": 3830
    },
    {
      "epoch": 1.2,
      "grad_norm": 0.00433349609375,
      "learning_rate": 5.155555555555556e-06,
      "loss": 0.0231,
      "step": 3840
    },
    {
      "epoch": 1.21,
      "grad_norm": 0.00201416015625,
      "learning_rate": 5.1111111111111115e-06,
      "loss": 0.0001,
      "step": 3850
    },
    {
      "epoch": 1.21,
      "grad_norm": 4.375,
      "learning_rate": 5.0666666666666676e-06,
      "loss": 0.0294,
      "step": 3860
    },
    {
      "epoch": 1.21,
      "grad_norm": 0.004547119140625,
      "learning_rate": 5.022222222222223e-06,
      "loss": 0.0379,
      "step": 3870
    },
    {
      "epoch": 1.22,
      "grad_norm": 0.0010223388671875,
      "learning_rate": 4.977777777777778e-06,
      "loss": 0.0284,
      "step": 3880
    },
    {
      "epoch": 1.22,
      "grad_norm": 4.8125,
      "learning_rate": 4.933333333333334e-06,
      "loss": 0.0131,
      "step": 3890
    },
    {
      "epoch": 1.22,
      "grad_norm": 3.9375,
      "learning_rate": 4.888888888888889e-06,
      "loss": 0.0527,
      "step": 3900
    },
    {
      "epoch": 1.23,
      "grad_norm": 0.0010986328125,
      "learning_rate": 4.8444444444444446e-06,
      "loss": 0.0002,
      "step": 3910
    },
    {
      "epoch": 1.23,
      "grad_norm": 0.0035247802734375,
      "learning_rate": 4.800000000000001e-06,
      "loss": 0.0455,
      "step": 3920
    },
    {
      "epoch": 1.23,
      "grad_norm": 8.125,
      "learning_rate": 4.755555555555556e-06,
      "loss": 0.0104,
      "step": 3930
    },
    {
      "epoch": 1.23,
      "grad_norm": 0.00156402587890625,
      "learning_rate": 4.711111111111111e-06,
      "loss": 0.0,
      "step": 3940
    },
    {
      "epoch": 1.24,
      "grad_norm": 0.0026397705078125,
      "learning_rate": 4.666666666666667e-06,
      "loss": 0.0191,
      "step": 3950
    },
    {
      "epoch": 1.24,
      "grad_norm": 0.0037689208984375,
      "learning_rate": 4.622222222222222e-06,
      "loss": 0.0288,
      "step": 3960
    },
    {
      "epoch": 1.24,
      "grad_norm": 0.00135040283203125,
      "learning_rate": 4.5777777777777785e-06,
      "loss": 0.0519,
      "step": 3970
    },
    {
      "epoch": 1.25,
      "grad_norm": 0.00106048583984375,
      "learning_rate": 4.533333333333334e-06,
      "loss": 0.0175,
      "step": 3980
    },
    {
      "epoch": 1.25,
      "grad_norm": 7.0,
      "learning_rate": 4.488888888888889e-06,
      "loss": 0.0066,
      "step": 3990
    },
    {
      "epoch": 1.25,
      "grad_norm": 0.0308837890625,
      "learning_rate": 4.444444444444444e-06,
      "loss": 0.0253,
      "step": 4000
    },
    {
      "epoch": 1.25,
      "eval_loss": 0.019900379702448845,
      "eval_runtime": 53.4572,
      "eval_samples_per_second": 18.707,
      "eval_steps_per_second": 18.707,
      "step": 4000
    },
    {
      "epoch": 1.26,
      "grad_norm": 0.004425048828125,
      "learning_rate": 4.4e-06,
      "loss": 0.0446,
      "step": 4010
    },
    {
      "epoch": 1.26,
      "grad_norm": 3.515625,
      "learning_rate": 4.3555555555555555e-06,
      "loss": 0.0218,
      "step": 4020
    },
    {
      "epoch": 1.26,
      "grad_norm": 40.75,
      "learning_rate": 4.3111111111111115e-06,
      "loss": 0.063,
      "step": 4030
    },
    {
      "epoch": 1.27,
      "grad_norm": 0.00130462646484375,
      "learning_rate": 4.266666666666668e-06,
      "loss": 0.0107,
      "step": 4040
    },
    {
      "epoch": 1.27,
      "grad_norm": 0.004302978515625,
      "learning_rate": 4.222222222222223e-06,
      "loss": 0.1038,
      "step": 4050
    },
    {
      "epoch": 1.27,
      "grad_norm": 2.34375,
      "learning_rate": 4.177777777777778e-06,
      "loss": 0.0249,
      "step": 4060
    },
    {
      "epoch": 1.28,
      "grad_norm": 0.01165771484375,
      "learning_rate": 4.133333333333333e-06,
      "loss": 0.0037,
      "step": 4070
    },
    {
      "epoch": 1.28,
      "grad_norm": 3.65625,
      "learning_rate": 4.088888888888889e-06,
      "loss": 0.0327,
      "step": 4080
    },
    {
      "epoch": 1.28,
      "grad_norm": 0.0019989013671875,
      "learning_rate": 4.044444444444445e-06,
      "loss": 0.0239,
      "step": 4090
    },
    {
      "epoch": 1.28,
      "grad_norm": 0.1796875,
      "learning_rate": 4.000000000000001e-06,
      "loss": 0.0338,
      "step": 4100
    },
    {
      "epoch": 1.29,
      "grad_norm": 0.0020294189453125,
      "learning_rate": 3.955555555555556e-06,
      "loss": 0.0132,
      "step": 4110
    },
    {
      "epoch": 1.29,
      "grad_norm": 5.21875,
      "learning_rate": 3.911111111111112e-06,
      "loss": 0.0517,
      "step": 4120
    },
    {
      "epoch": 1.29,
      "grad_norm": 7.5,
      "learning_rate": 3.866666666666667e-06,
      "loss": 0.031,
      "step": 4130
    },
    {
      "epoch": 1.3,
      "grad_norm": 0.0634765625,
      "learning_rate": 3.8222222222222224e-06,
      "loss": 0.025,
      "step": 4140
    },
    {
      "epoch": 1.3,
      "grad_norm": 0.00112152099609375,
      "learning_rate": 3.777777777777778e-06,
      "loss": 0.0612,
      "step": 4150
    },
    {
      "epoch": 1.3,
      "grad_norm": 10.9375,
      "learning_rate": 3.7333333333333337e-06,
      "loss": 0.0324,
      "step": 4160
    },
    {
      "epoch": 1.31,
      "grad_norm": 3.515625,
      "learning_rate": 3.688888888888889e-06,
      "loss": 0.0083,
      "step": 4170
    },
    {
      "epoch": 1.31,
      "grad_norm": 0.001617431640625,
      "learning_rate": 3.644444444444445e-06,
      "loss": 0.0202,
      "step": 4180
    },
    {
      "epoch": 1.31,
      "grad_norm": 0.000965118408203125,
      "learning_rate": 3.6000000000000003e-06,
      "loss": 0.0053,
      "step": 4190
    },
    {
      "epoch": 1.32,
      "grad_norm": 0.0009613037109375,
      "learning_rate": 3.555555555555556e-06,
      "loss": 0.0071,
      "step": 4200
    },
    {
      "epoch": 1.32,
      "grad_norm": 8.3125,
      "learning_rate": 3.511111111111111e-06,
      "loss": 0.0146,
      "step": 4210
    },
    {
      "epoch": 1.32,
      "grad_norm": 0.00958251953125,
      "learning_rate": 3.4666666666666672e-06,
      "loss": 0.0134,
      "step": 4220
    },
    {
      "epoch": 1.33,
      "grad_norm": 0.004852294921875,
      "learning_rate": 3.4222222222222224e-06,
      "loss": 0.0439,
      "step": 4230
    },
    {
      "epoch": 1.33,
      "grad_norm": 0.00107574462890625,
      "learning_rate": 3.377777777777778e-06,
      "loss": 0.0119,
      "step": 4240
    },
    {
      "epoch": 1.33,
      "grad_norm": 0.000972747802734375,
      "learning_rate": 3.3333333333333333e-06,
      "loss": 0.0148,
      "step": 4250
    },
    {
      "epoch": 1.33,
      "grad_norm": 0.0018157958984375,
      "learning_rate": 3.2888888888888894e-06,
      "loss": 0.018,
      "step": 4260
    },
    {
      "epoch": 1.34,
      "grad_norm": 0.004638671875,
      "learning_rate": 3.2444444444444446e-06,
      "loss": 0.0257,
      "step": 4270
    },
    {
      "epoch": 1.34,
      "grad_norm": 13.1875,
      "learning_rate": 3.2000000000000003e-06,
      "loss": 0.1221,
      "step": 4280
    },
    {
      "epoch": 1.34,
      "grad_norm": 0.0011444091796875,
      "learning_rate": 3.1555555555555555e-06,
      "loss": 0.0103,
      "step": 4290
    },
    {
      "epoch": 1.35,
      "grad_norm": 0.00104522705078125,
      "learning_rate": 3.1111111111111116e-06,
      "loss": 0.0097,
      "step": 4300
    },
    {
      "epoch": 1.35,
      "grad_norm": 0.008544921875,
      "learning_rate": 3.066666666666667e-06,
      "loss": 0.0167,
      "step": 4310
    },
    {
      "epoch": 1.35,
      "grad_norm": 0.004974365234375,
      "learning_rate": 3.0222222222222225e-06,
      "loss": 0.0344,
      "step": 4320
    },
    {
      "epoch": 1.36,
      "grad_norm": 0.0016632080078125,
      "learning_rate": 2.9777777777777777e-06,
      "loss": 0.0225,
      "step": 4330
    },
    {
      "epoch": 1.36,
      "grad_norm": 0.0011749267578125,
      "learning_rate": 2.9333333333333338e-06,
      "loss": 0.0501,
      "step": 4340
    },
    {
      "epoch": 1.36,
      "grad_norm": 0.004302978515625,
      "learning_rate": 2.888888888888889e-06,
      "loss": 0.0353,
      "step": 4350
    },
    {
      "epoch": 1.37,
      "grad_norm": 6.0625,
      "learning_rate": 2.8444444444444446e-06,
      "loss": 0.0499,
      "step": 4360
    },
    {
      "epoch": 1.37,
      "grad_norm": 11.375,
      "learning_rate": 2.8000000000000003e-06,
      "loss": 0.0481,
      "step": 4370
    },
    {
      "epoch": 1.37,
      "grad_norm": 7.40625,
      "learning_rate": 2.755555555555556e-06,
      "loss": 0.046,
      "step": 4380
    },
    {
      "epoch": 1.38,
      "grad_norm": 0.26171875,
      "learning_rate": 2.7111111111111116e-06,
      "loss": 0.0224,
      "step": 4390
    },
    {
      "epoch": 1.38,
      "grad_norm": 22.25,
      "learning_rate": 2.666666666666667e-06,
      "loss": 0.048,
      "step": 4400
    },
    {
      "epoch": 1.38,
      "grad_norm": 0.00112152099609375,
      "learning_rate": 2.6222222222222225e-06,
      "loss": 0.0403,
      "step": 4410
    },
    {
      "epoch": 1.38,
      "grad_norm": 7.78125,
      "learning_rate": 2.577777777777778e-06,
      "loss": 0.0224,
      "step": 4420
    },
    {
      "epoch": 1.39,
      "grad_norm": 0.00116729736328125,
      "learning_rate": 2.5333333333333338e-06,
      "loss": 0.024,
      "step": 4430
    },
    {
      "epoch": 1.39,
      "grad_norm": 6.96875,
      "learning_rate": 2.488888888888889e-06,
      "loss": 0.0603,
      "step": 4440
    },
    {
      "epoch": 1.39,
      "grad_norm": 0.00099945068359375,
      "learning_rate": 2.4444444444444447e-06,
      "loss": 0.0245,
      "step": 4450
    },
    {
      "epoch": 1.4,
      "grad_norm": 0.004119873046875,
      "learning_rate": 2.4000000000000003e-06,
      "loss": 0.0144,
      "step": 4460
    },
    {
      "epoch": 1.4,
      "grad_norm": 6.0,
      "learning_rate": 2.3555555555555555e-06,
      "loss": 0.0329,
      "step": 4470
    },
    {
      "epoch": 1.4,
      "grad_norm": 9.9375,
      "learning_rate": 2.311111111111111e-06,
      "loss": 0.0342,
      "step": 4480
    },
    {
      "epoch": 1.41,
      "grad_norm": 0.2490234375,
      "learning_rate": 2.266666666666667e-06,
      "loss": 0.0128,
      "step": 4490
    },
    {
      "epoch": 1.41,
      "grad_norm": 0.00421142578125,
      "learning_rate": 2.222222222222222e-06,
      "loss": 0.0146,
      "step": 4500
    },
    {
      "epoch": 1.41,
      "grad_norm": 0.00152587890625,
      "learning_rate": 2.1777777777777777e-06,
      "loss": 0.0191,
      "step": 4510
    },
    {
      "epoch": 1.42,
      "grad_norm": 3.84375,
      "learning_rate": 2.133333333333334e-06,
      "loss": 0.0172,
      "step": 4520
    },
    {
      "epoch": 1.42,
      "grad_norm": 0.00092315673828125,
      "learning_rate": 2.088888888888889e-06,
      "loss": 0.0341,
      "step": 4530
    },
    {
      "epoch": 1.42,
      "grad_norm": 0.00156402587890625,
      "learning_rate": 2.0444444444444447e-06,
      "loss": 0.0143,
      "step": 4540
    },
    {
      "epoch": 1.43,
      "grad_norm": 0.00115203857421875,
      "learning_rate": 2.0000000000000003e-06,
      "loss": 0.0096,
      "step": 4550
    },
    {
      "epoch": 1.43,
      "grad_norm": 0.002410888671875,
      "learning_rate": 1.955555555555556e-06,
      "loss": 0.0083,
      "step": 4560
    },
    {
      "epoch": 1.43,
      "grad_norm": 0.00091552734375,
      "learning_rate": 1.9111111111111112e-06,
      "loss": 0.0089,
      "step": 4570
    },
    {
      "epoch": 1.43,
      "grad_norm": 0.0016326904296875,
      "learning_rate": 1.8666666666666669e-06,
      "loss": 0.008,
      "step": 4580
    },
    {
      "epoch": 1.44,
      "grad_norm": 0.004364013671875,
      "learning_rate": 1.8222222222222225e-06,
      "loss": 0.0497,
      "step": 4590
    },
    {
      "epoch": 1.44,
      "grad_norm": 0.004241943359375,
      "learning_rate": 1.777777777777778e-06,
      "loss": 0.0358,
      "step": 4600
    },
    {
      "epoch": 1.44,
      "grad_norm": 0.0009613037109375,
      "learning_rate": 1.7333333333333336e-06,
      "loss": 0.0263,
      "step": 4610
    },
    {
      "epoch": 1.45,
      "grad_norm": 0.00109100341796875,
      "learning_rate": 1.688888888888889e-06,
      "loss": 0.0274,
      "step": 4620
    },
    {
      "epoch": 1.45,
      "grad_norm": 6.03125,
      "learning_rate": 1.6444444444444447e-06,
      "loss": 0.0608,
      "step": 4630
    },
    {
      "epoch": 1.45,
      "grad_norm": 0.00113677978515625,
      "learning_rate": 1.6000000000000001e-06,
      "loss": 0.0,
      "step": 4640
    },
    {
      "epoch": 1.46,
      "grad_norm": 7.1875,
      "learning_rate": 1.5555555555555558e-06,
      "loss": 0.0306,
      "step": 4650
    },
    {
      "epoch": 1.46,
      "grad_norm": 0.1025390625,
      "learning_rate": 1.5111111111111112e-06,
      "loss": 0.0911,
      "step": 4660
    },
    {
      "epoch": 1.46,
      "grad_norm": 0.0012054443359375,
      "learning_rate": 1.4666666666666669e-06,
      "loss": 0.0128,
      "step": 4670
    },
    {
      "epoch": 1.47,
      "grad_norm": 0.0009918212890625,
      "learning_rate": 1.4222222222222223e-06,
      "loss": 0.0017,
      "step": 4680
    },
    {
      "epoch": 1.47,
      "grad_norm": 5.125,
      "learning_rate": 1.377777777777778e-06,
      "loss": 0.0488,
      "step": 4690
    },
    {
      "epoch": 1.47,
      "grad_norm": 4.09375,
      "learning_rate": 1.3333333333333334e-06,
      "loss": 0.0071,
      "step": 4700
    },
    {
      "epoch": 1.48,
      "grad_norm": 0.00164794921875,
      "learning_rate": 1.288888888888889e-06,
      "loss": 0.0103,
      "step": 4710
    },
    {
      "epoch": 1.48,
      "grad_norm": 0.00186920166015625,
      "learning_rate": 1.2444444444444445e-06,
      "loss": 0.0139,
      "step": 4720
    },
    {
      "epoch": 1.48,
      "grad_norm": 2.203125,
      "learning_rate": 1.2000000000000002e-06,
      "loss": 0.043,
      "step": 4730
    },
    {
      "epoch": 1.49,
      "grad_norm": 5.03125,
      "learning_rate": 1.1555555555555556e-06,
      "loss": 0.057,
      "step": 4740
    },
    {
      "epoch": 1.49,
      "grad_norm": 0.00213623046875,
      "learning_rate": 1.111111111111111e-06,
      "loss": 0.0441,
      "step": 4750
    },
    {
      "epoch": 1.49,
      "grad_norm": 0.00433349609375,
      "learning_rate": 1.066666666666667e-06,
      "loss": 0.0042,
      "step": 4760
    },
    {
      "epoch": 1.49,
      "grad_norm": 3.140625,
      "learning_rate": 1.0222222222222223e-06,
      "loss": 0.0328,
      "step": 4770
    },
    {
      "epoch": 1.5,
      "grad_norm": 0.0048828125,
      "learning_rate": 9.77777777777778e-07,
      "loss": 0.0592,
      "step": 4780
    },
    {
      "epoch": 1.5,
      "grad_norm": 0.004364013671875,
      "learning_rate": 9.333333333333334e-07,
      "loss": 0.0307,
      "step": 4790
    },
    {
      "epoch": 1.5,
      "grad_norm": 10.8125,
      "learning_rate": 8.88888888888889e-07,
      "loss": 0.0961,
      "step": 4800
    },
    {
      "epoch": 1.51,
      "grad_norm": 0.00176239013671875,
      "learning_rate": 8.444444444444445e-07,
      "loss": 0.0042,
      "step": 4810
    },
    {
      "epoch": 1.51,
      "grad_norm": 0.00439453125,
      "learning_rate": 8.000000000000001e-07,
      "loss": 0.0354,
      "step": 4820
    },
    {
      "epoch": 1.51,
      "grad_norm": 0.333984375,
      "learning_rate": 7.555555555555556e-07,
      "loss": 0.0438,
      "step": 4830
    },
    {
      "epoch": 1.52,
      "grad_norm": 0.17578125,
      "learning_rate": 7.111111111111112e-07,
      "loss": 0.0102,
      "step": 4840
    },
    {
      "epoch": 1.52,
      "grad_norm": 0.0042724609375,
      "learning_rate": 6.666666666666667e-07,
      "loss": 0.045,
      "step": 4850
    },
    {
      "epoch": 1.52,
      "grad_norm": 0.00116729736328125,
      "learning_rate": 6.222222222222223e-07,
      "loss": 0.006,
      "step": 4860
    },
    {
      "epoch": 1.53,
      "grad_norm": 0.0009918212890625,
      "learning_rate": 5.777777777777778e-07,
      "loss": 0.0579,
      "step": 4870
    },
    {
      "epoch": 1.53,
      "grad_norm": 4.625,
      "learning_rate": 5.333333333333335e-07,
      "loss": 0.0106,
      "step": 4880
    },
    {
      "epoch": 1.53,
      "grad_norm": 19.625,
      "learning_rate": 4.88888888888889e-07,
      "loss": 0.0563,
      "step": 4890
    },
    {
      "epoch": 1.54,
      "grad_norm": 0.0010528564453125,
      "learning_rate": 4.444444444444445e-07,
      "loss": 0.0234,
      "step": 4900
    },
    {
      "epoch": 1.54,
      "grad_norm": 0.00119781494140625,
      "learning_rate": 4.0000000000000003e-07,
      "loss": 0.0747,
      "step": 4910
    },
    {
      "epoch": 1.54,
      "grad_norm": 0.00396728515625,
      "learning_rate": 3.555555555555556e-07,
      "loss": 0.0097,
      "step": 4920
    },
    {
      "epoch": 1.54,
      "grad_norm": 0.001007080078125,
      "learning_rate": 3.111111111111111e-07,
      "loss": 0.0202,
      "step": 4930
    },
    {
      "epoch": 1.55,
      "grad_norm": 0.0011138916015625,
      "learning_rate": 2.666666666666667e-07,
      "loss": 0.0065,
      "step": 4940
    },
    {
      "epoch": 1.55,
      "grad_norm": 0.00115966796875,
      "learning_rate": 2.2222222222222224e-07,
      "loss": 0.0173,
      "step": 4950
    },
    {
      "epoch": 1.55,
      "grad_norm": 0.001800537109375,
      "learning_rate": 1.777777777777778e-07,
      "loss": 0.0358,
      "step": 4960
    },
    {
      "epoch": 1.56,
      "grad_norm": 0.00179290771484375,
      "learning_rate": 1.3333333333333336e-07,
      "loss": 0.0396,
      "step": 4970
    },
    {
      "epoch": 1.56,
      "grad_norm": 0.00118255615234375,
      "learning_rate": 8.88888888888889e-08,
      "loss": 0.0155,
      "step": 4980
    },
    {
      "epoch": 1.56,
      "grad_norm": 0.00494384765625,
      "learning_rate": 4.444444444444445e-08,
      "loss": 0.0573,
      "step": 4990
    },
    {
      "epoch": 1.57,
      "grad_norm": 0.0018463134765625,
      "learning_rate": 0.0,
      "loss": 0.0309,
      "step": 5000
    },
    {
      "epoch": 1.57,
      "eval_loss": 0.019992012530565262,
      "eval_runtime": 53.6022,
      "eval_samples_per_second": 18.656,
      "eval_steps_per_second": 18.656,
      "step": 5000
    }
  ],
  "logging_steps": 10,
  "max_steps": 5000,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 2,
  "save_steps": 1000,
  "total_flos": 7.866849165312e+16,
  "train_batch_size": 1,
  "trial_name": null,
  "trial_params": null
}