{
  "best_metric": 0.753384912959381,
  "best_model_checkpoint": "../models/t5-picard/checkpoint-2368",
  "epoch": 99.98963557338682,
  "global_step": 7400,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.01,
      "learning_rate": 0.0001,
      "loss": 3.3904,
      "step": 1
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0001,
      "loss": 1.8308,
      "step": 4
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0001,
      "loss": 0.8985,
      "step": 8
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0001,
      "loss": 0.7145,
      "step": 12
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0001,
      "loss": 0.4975,
      "step": 16
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0001,
      "loss": 0.4788,
      "step": 20
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0001,
      "loss": 0.4007,
      "step": 24
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0001,
      "loss": 0.3452,
      "step": 28
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0001,
      "loss": 0.3293,
      "step": 32
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0001,
      "loss": 0.3214,
      "step": 36
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0001,
      "loss": 0.3254,
      "step": 40
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0001,
      "loss": 0.2502,
      "step": 44
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0001,
      "loss": 0.261,
      "step": 48
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0001,
      "loss": 0.2192,
      "step": 52
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001,
      "loss": 0.2354,
      "step": 56
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0001,
      "loss": 0.2376,
      "step": 60
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.0001,
      "loss": 0.1848,
      "step": 64
    },
    {
      "epoch": 0.86,
      "eval_exec": 0.528046421663443,
      "eval_loss": 0.1530311554670334,
      "eval_runtime": 253.3827,
      "eval_samples_per_second": 4.081,
      "step": 64
    },
    {
      "epoch": 0.91,
      "learning_rate": 0.0001,
      "loss": 0.1615,
      "step": 68
    },
    {
      "epoch": 0.96,
      "learning_rate": 0.0001,
      "loss": 0.1756,
      "step": 72
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0001,
      "loss": 0.2114,
      "step": 76
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0001,
      "loss": 0.1703,
      "step": 80
    },
    {
      "epoch": 1.13,
      "learning_rate": 0.0001,
      "loss": 0.1485,
      "step": 84
    },
    {
      "epoch": 1.19,
      "learning_rate": 0.0001,
      "loss": 0.119,
      "step": 88
    },
    {
      "epoch": 1.24,
      "learning_rate": 0.0001,
      "loss": 0.1273,
      "step": 92
    },
    {
      "epoch": 1.29,
      "learning_rate": 0.0001,
      "loss": 0.1625,
      "step": 96
    },
    {
      "epoch": 1.35,
      "learning_rate": 0.0001,
      "loss": 0.13,
      "step": 100
    },
    {
      "epoch": 1.4,
      "learning_rate": 0.0001,
      "loss": 0.1265,
      "step": 104
    },
    {
      "epoch": 1.45,
      "learning_rate": 0.0001,
      "loss": 0.1088,
      "step": 108
    },
    {
      "epoch": 1.51,
      "learning_rate": 0.0001,
      "loss": 0.1397,
      "step": 112
    },
    {
      "epoch": 1.56,
      "learning_rate": 0.0001,
      "loss": 0.1404,
      "step": 116
    },
    {
      "epoch": 1.62,
      "learning_rate": 0.0001,
      "loss": 0.1136,
      "step": 120
    },
    {
      "epoch": 1.67,
      "learning_rate": 0.0001,
      "loss": 0.0919,
      "step": 124
    },
    {
      "epoch": 1.72,
      "learning_rate": 0.0001,
      "loss": 0.1114,
      "step": 128
    },
    {
      "epoch": 1.72,
      "eval_exec": 0.5309477756286267,
      "eval_loss": 0.1567097008228302,
      "eval_runtime": 282.0248,
      "eval_samples_per_second": 3.666,
      "step": 128
    },
    {
      "epoch": 1.78,
      "learning_rate": 0.0001,
      "loss": 0.131,
      "step": 132
    },
    {
      "epoch": 1.83,
      "learning_rate": 0.0001,
      "loss": 0.0972,
      "step": 136
    },
    {
      "epoch": 1.88,
      "learning_rate": 0.0001,
      "loss": 0.0924,
      "step": 140
    },
    {
      "epoch": 1.94,
      "learning_rate": 0.0001,
      "loss": 0.0915,
      "step": 144
    },
    {
      "epoch": 1.99,
      "learning_rate": 0.0001,
      "loss": 0.1105,
      "step": 148
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0001,
      "loss": 0.1299,
      "step": 152
    },
    {
      "epoch": 2.11,
      "learning_rate": 0.0001,
      "loss": 0.0861,
      "step": 156
    },
    {
      "epoch": 2.16,
      "learning_rate": 0.0001,
      "loss": 0.0779,
      "step": 160
    },
    {
      "epoch": 2.21,
      "learning_rate": 0.0001,
      "loss": 0.0677,
      "step": 164
    },
    {
      "epoch": 2.27,
      "learning_rate": 0.0001,
      "loss": 0.0879,
      "step": 168
    },
    {
      "epoch": 2.32,
      "learning_rate": 0.0001,
      "loss": 0.0953,
      "step": 172
    },
    {
      "epoch": 2.37,
      "learning_rate": 0.0001,
      "loss": 0.0766,
      "step": 176
    },
    {
      "epoch": 2.43,
      "learning_rate": 0.0001,
      "loss": 0.0621,
      "step": 180
    },
    {
      "epoch": 2.48,
      "learning_rate": 0.0001,
      "loss": 0.0793,
      "step": 184
    },
    {
      "epoch": 2.53,
      "learning_rate": 0.0001,
      "loss": 0.0929,
      "step": 188
    },
    {
      "epoch": 2.59,
      "learning_rate": 0.0001,
      "loss": 0.074,
      "step": 192
    },
    {
      "epoch": 2.59,
      "eval_exec": 0.6876208897485493,
      "eval_loss": 0.1364244967699051,
      "eval_runtime": 267.8144,
      "eval_samples_per_second": 3.861,
      "step": 192
    },
    {
      "epoch": 2.64,
      "learning_rate": 0.0001,
      "loss": 0.0661,
      "step": 196
    },
    {
      "epoch": 2.7,
      "learning_rate": 0.0001,
      "loss": 0.059,
      "step": 200
    },
    {
      "epoch": 2.75,
      "learning_rate": 0.0001,
      "loss": 0.0857,
      "step": 204
    },
    {
      "epoch": 2.8,
      "learning_rate": 0.0001,
      "loss": 0.0743,
      "step": 208
    },
    {
      "epoch": 2.86,
      "learning_rate": 0.0001,
      "loss": 0.0755,
      "step": 212
    },
    {
      "epoch": 2.91,
      "learning_rate": 0.0001,
      "loss": 0.0572,
      "step": 216
    },
    {
      "epoch": 2.96,
      "learning_rate": 0.0001,
      "loss": 0.1111,
      "step": 220
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.0001,
      "loss": 0.0819,
      "step": 224
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0001,
      "loss": 0.0695,
      "step": 228
    },
    {
      "epoch": 3.13,
      "learning_rate": 0.0001,
      "loss": 0.0553,
      "step": 232
    },
    {
      "epoch": 3.19,
      "learning_rate": 0.0001,
      "loss": 0.0473,
      "step": 236
    },
    {
      "epoch": 3.24,
      "learning_rate": 0.0001,
      "loss": 0.051,
      "step": 240
    },
    {
      "epoch": 3.29,
      "learning_rate": 0.0001,
      "loss": 0.065,
      "step": 244
    },
    {
      "epoch": 3.35,
      "learning_rate": 0.0001,
      "loss": 0.0554,
      "step": 248
    },
    {
      "epoch": 3.4,
      "learning_rate": 0.0001,
      "loss": 0.0494,
      "step": 252
    },
    {
      "epoch": 3.45,
      "learning_rate": 0.0001,
      "loss": 0.0468,
      "step": 256
    },
    {
      "epoch": 3.45,
      "eval_exec": 0.6876208897485493,
      "eval_loss": 0.15044596791267395,
      "eval_runtime": 292.7353,
      "eval_samples_per_second": 3.532,
      "step": 256
    },
    {
      "epoch": 3.51,
      "learning_rate": 0.0001,
      "loss": 0.0588,
      "step": 260
    },
    {
      "epoch": 3.56,
      "learning_rate": 0.0001,
      "loss": 0.0591,
      "step": 264
    },
    {
      "epoch": 3.62,
      "learning_rate": 0.0001,
      "loss": 0.0498,
      "step": 268
    },
    {
      "epoch": 3.67,
      "learning_rate": 0.0001,
      "loss": 0.0414,
      "step": 272
    },
    {
      "epoch": 3.72,
      "learning_rate": 0.0001,
      "loss": 0.048,
      "step": 276
    },
    {
      "epoch": 3.78,
      "learning_rate": 0.0001,
      "loss": 0.0612,
      "step": 280
    },
    {
      "epoch": 3.83,
      "learning_rate": 0.0001,
      "loss": 0.051,
      "step": 284
    },
    {
      "epoch": 3.88,
      "learning_rate": 0.0001,
      "loss": 0.042,
      "step": 288
    },
    {
      "epoch": 3.94,
      "learning_rate": 0.0001,
      "loss": 0.2851,
      "step": 292
    },
    {
      "epoch": 3.99,
      "learning_rate": 0.0001,
      "loss": 0.0504,
      "step": 296
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0001,
      "loss": 0.0575,
      "step": 300
    },
    {
      "epoch": 4.11,
      "learning_rate": 0.0001,
      "loss": 0.0389,
      "step": 304
    },
    {
      "epoch": 4.16,
      "learning_rate": 0.0001,
      "loss": 0.0352,
      "step": 308
    },
    {
      "epoch": 4.21,
      "learning_rate": 0.0001,
      "loss": 0.0335,
      "step": 312
    },
    {
      "epoch": 4.27,
      "learning_rate": 0.0001,
      "loss": 0.0431,
      "step": 316
    },
    {
      "epoch": 4.32,
      "learning_rate": 0.0001,
      "loss": 0.0417,
      "step": 320
    },
    {
      "epoch": 4.32,
      "eval_exec": 0.706963249516441,
      "eval_loss": 0.15714390575885773,
      "eval_runtime": 277.5044,
      "eval_samples_per_second": 3.726,
      "step": 320
    },
    {
      "epoch": 4.37,
      "learning_rate": 0.0001,
      "loss": 0.0341,
      "step": 324
    },
    {
      "epoch": 4.43,
      "learning_rate": 0.0001,
      "loss": 0.0426,
      "step": 328
    },
    {
      "epoch": 4.48,
      "learning_rate": 0.0001,
      "loss": 0.0407,
      "step": 332
    },
    {
      "epoch": 4.53,
      "learning_rate": 0.0001,
      "loss": 0.0446,
      "step": 336
    },
    {
      "epoch": 4.59,
      "learning_rate": 0.0001,
      "loss": 0.0408,
      "step": 340
    },
    {
      "epoch": 4.64,
      "learning_rate": 0.0001,
      "loss": 0.0323,
      "step": 344
    },
    {
      "epoch": 4.7,
      "learning_rate": 0.0001,
      "loss": 0.032,
      "step": 348
    },
    {
      "epoch": 4.75,
      "learning_rate": 0.0001,
      "loss": 0.0428,
      "step": 352
    },
    {
      "epoch": 4.8,
      "learning_rate": 0.0001,
      "loss": 0.0452,
      "step": 356
    },
    {
      "epoch": 4.86,
      "learning_rate": 0.0001,
      "loss": 0.033,
      "step": 360
    },
    {
      "epoch": 4.91,
      "learning_rate": 0.0001,
      "loss": 0.0238,
      "step": 364
    },
    {
      "epoch": 4.96,
      "learning_rate": 0.0001,
      "loss": 0.0383,
      "step": 368
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0001,
      "loss": 0.0393,
      "step": 372
    },
    {
      "epoch": 5.08,
      "learning_rate": 0.0001,
      "loss": 0.033,
      "step": 376
    },
    {
      "epoch": 5.13,
      "learning_rate": 0.0001,
      "loss": 0.0261,
      "step": 380
    },
    {
      "epoch": 5.19,
      "learning_rate": 0.0001,
      "loss": 0.0204,
      "step": 384
    },
    {
      "epoch": 5.19,
      "eval_exec": 0.7040618955512572,
      "eval_loss": 0.17861126363277435,
      "eval_runtime": 288.586,
      "eval_samples_per_second": 3.583,
      "step": 384
    },
    {
      "epoch": 5.24,
      "learning_rate": 0.0001,
      "loss": 0.0283,
      "step": 388
    },
    {
      "epoch": 5.29,
      "learning_rate": 0.0001,
      "loss": 0.0328,
      "step": 392
    },
    {
      "epoch": 5.35,
      "learning_rate": 0.0001,
      "loss": 0.0303,
      "step": 396
    },
    {
      "epoch": 5.4,
      "learning_rate": 0.0001,
      "loss": 0.0225,
      "step": 400
    },
    {
      "epoch": 5.45,
      "learning_rate": 0.0001,
      "loss": 0.0243,
      "step": 404
    },
    {
      "epoch": 5.51,
      "learning_rate": 0.0001,
      "loss": 0.0319,
      "step": 408
    },
    {
      "epoch": 5.56,
      "learning_rate": 0.0001,
      "loss": 0.0392,
      "step": 412
    },
    {
      "epoch": 5.62,
      "learning_rate": 0.0001,
      "loss": 0.0265,
      "step": 416
    },
    {
      "epoch": 5.67,
      "learning_rate": 0.0001,
      "loss": 0.0198,
      "step": 420
    },
    {
      "epoch": 5.72,
      "learning_rate": 0.0001,
      "loss": 0.0293,
      "step": 424
    },
    {
      "epoch": 5.78,
      "learning_rate": 0.0001,
      "loss": 0.0326,
      "step": 428
    },
    {
      "epoch": 5.83,
      "learning_rate": 0.0001,
      "loss": 0.0288,
      "step": 432
    },
    {
      "epoch": 5.88,
      "learning_rate": 0.0001,
      "loss": 0.023,
      "step": 436
    },
    {
      "epoch": 5.94,
      "learning_rate": 0.0001,
      "loss": 0.0227,
      "step": 440
    },
    {
      "epoch": 5.99,
      "learning_rate": 0.0001,
      "loss": 0.0273,
      "step": 444
    },
    {
      "epoch": 6.05,
      "learning_rate": 0.0001,
      "loss": 0.029,
      "step": 448
    },
    {
      "epoch": 6.05,
      "eval_exec": 0.6924564796905223,
      "eval_loss": 0.17845195531845093,
      "eval_runtime": 299.4799,
      "eval_samples_per_second": 3.453,
      "step": 448
    },
    {
      "epoch": 6.11,
      "learning_rate": 0.0001,
      "loss": 0.0276,
      "step": 452
    },
    {
      "epoch": 6.16,
      "learning_rate": 0.0001,
      "loss": 0.0227,
      "step": 456
    },
    {
      "epoch": 6.21,
      "learning_rate": 0.0001,
      "loss": 0.0185,
      "step": 460
    },
    {
      "epoch": 6.27,
      "learning_rate": 0.0001,
      "loss": 0.0244,
      "step": 464
    },
    {
      "epoch": 6.32,
      "learning_rate": 0.0001,
      "loss": 0.0281,
      "step": 468
    },
    {
      "epoch": 6.37,
      "learning_rate": 0.0001,
      "loss": 0.0181,
      "step": 472
    },
    {
      "epoch": 6.43,
      "learning_rate": 0.0001,
      "loss": 0.0132,
      "step": 476
    },
    {
      "epoch": 6.48,
      "learning_rate": 0.0001,
      "loss": 0.0205,
      "step": 480
    },
    {
      "epoch": 6.53,
      "learning_rate": 0.0001,
      "loss": 0.0274,
      "step": 484
    },
    {
      "epoch": 6.59,
      "learning_rate": 0.0001,
      "loss": 0.0209,
      "step": 488
    },
    {
      "epoch": 6.64,
      "learning_rate": 0.0001,
      "loss": 0.0161,
      "step": 492
    },
    {
      "epoch": 6.7,
      "learning_rate": 0.0001,
      "loss": 0.0186,
      "step": 496
    },
    {
      "epoch": 6.75,
      "learning_rate": 0.0001,
      "loss": 0.0202,
      "step": 500
    },
    {
      "epoch": 6.8,
      "learning_rate": 0.0001,
      "loss": 0.0246,
      "step": 504
    },
    {
      "epoch": 6.86,
      "learning_rate": 0.0001,
      "loss": 0.0177,
      "step": 508
    },
    {
      "epoch": 6.91,
      "learning_rate": 0.0001,
      "loss": 0.0147,
      "step": 512
    },
    {
      "epoch": 6.91,
      "eval_exec": 0.7214700193423598,
      "eval_loss": 0.1890781968832016,
      "eval_runtime": 285.738,
      "eval_samples_per_second": 3.619,
      "step": 512
    },
    {
      "epoch": 6.96,
      "learning_rate": 0.0001,
      "loss": 0.023,
      "step": 516
    },
    {
      "epoch": 7.03,
      "learning_rate": 0.0001,
      "loss": 0.0231,
      "step": 520
    },
    {
      "epoch": 7.08,
      "learning_rate": 0.0001,
      "loss": 0.0169,
      "step": 524
    },
    {
      "epoch": 7.13,
      "learning_rate": 0.0001,
      "loss": 0.0161,
      "step": 528
    },
    {
      "epoch": 7.19,
      "learning_rate": 0.0001,
      "loss": 0.0115,
      "step": 532
    },
    {
      "epoch": 7.24,
      "learning_rate": 0.0001,
      "loss": 0.0145,
      "step": 536
    },
    {
      "epoch": 7.29,
      "learning_rate": 0.0001,
      "loss": 0.0273,
      "step": 540
    },
    {
      "epoch": 7.35,
      "learning_rate": 0.0001,
      "loss": 0.022,
      "step": 544
    },
    {
      "epoch": 7.4,
      "learning_rate": 0.0001,
      "loss": 0.0154,
      "step": 548
    },
    {
      "epoch": 7.45,
      "learning_rate": 0.0001,
      "loss": 0.0123,
      "step": 552
    },
    {
      "epoch": 7.51,
      "learning_rate": 0.0001,
      "loss": 0.0159,
      "step": 556
    },
    {
      "epoch": 7.56,
      "learning_rate": 0.0001,
      "loss": 0.0198,
      "step": 560
    },
    {
      "epoch": 7.62,
      "learning_rate": 0.0001,
      "loss": 0.0213,
      "step": 564
    },
    {
      "epoch": 7.67,
      "learning_rate": 0.0001,
      "loss": 0.0148,
      "step": 568
    },
    {
      "epoch": 7.72,
      "learning_rate": 0.0001,
      "loss": 0.0137,
      "step": 572
    },
    {
      "epoch": 7.78,
      "learning_rate": 0.0001,
      "loss": 0.0181,
      "step": 576
    },
    {
      "epoch": 7.78,
      "eval_exec": 0.7040618955512572,
      "eval_loss": 0.22579778730869293,
      "eval_runtime": 297.1361,
      "eval_samples_per_second": 3.48,
      "step": 576
    },
    {
      "epoch": 7.83,
      "learning_rate": 0.0001,
      "loss": 0.0196,
      "step": 580
    },
    {
      "epoch": 7.88,
      "learning_rate": 0.0001,
      "loss": 0.0141,
      "step": 584
    },
    {
      "epoch": 7.94,
      "learning_rate": 0.0001,
      "loss": 0.0118,
      "step": 588
    },
    {
      "epoch": 7.99,
      "learning_rate": 0.0001,
      "loss": 0.0192,
      "step": 592
    },
    {
      "epoch": 8.05,
      "learning_rate": 0.0001,
      "loss": 0.0181,
      "step": 596
    },
    {
      "epoch": 8.11,
      "learning_rate": 0.0001,
      "loss": 0.0108,
      "step": 600
    },
    {
      "epoch": 8.16,
      "learning_rate": 0.0001,
      "loss": 0.0112,
      "step": 604
    },
    {
      "epoch": 8.21,
      "learning_rate": 0.0001,
      "loss": 0.0092,
      "step": 608
    },
    {
      "epoch": 8.27,
      "learning_rate": 0.0001,
      "loss": 0.0153,
      "step": 612
    },
    {
      "epoch": 8.32,
      "learning_rate": 0.0001,
      "loss": 0.0169,
      "step": 616
    },
    {
      "epoch": 8.37,
      "learning_rate": 0.0001,
      "loss": 0.011,
      "step": 620
    },
    {
      "epoch": 8.43,
      "learning_rate": 0.0001,
      "loss": 0.0095,
      "step": 624
    },
    {
      "epoch": 8.48,
      "learning_rate": 0.0001,
      "loss": 0.0128,
      "step": 628
    },
    {
      "epoch": 8.53,
      "learning_rate": 0.0001,
      "loss": 0.0161,
      "step": 632
    },
    {
      "epoch": 8.59,
      "learning_rate": 0.0001,
      "loss": 0.0128,
      "step": 636
    },
    {
      "epoch": 8.64,
      "learning_rate": 0.0001,
      "loss": 0.0103,
      "step": 640
    },
    {
      "epoch": 8.64,
      "eval_exec": 0.6895551257253385,
      "eval_loss": 0.2176068276166916,
      "eval_runtime": 294.0528,
      "eval_samples_per_second": 3.516,
      "step": 640
    },
    {
      "epoch": 8.7,
      "learning_rate": 0.0001,
      "loss": 0.0074,
      "step": 644
    },
    {
      "epoch": 8.75,
      "learning_rate": 0.0001,
      "loss": 0.0152,
      "step": 648
    },
    {
      "epoch": 8.8,
      "learning_rate": 0.0001,
      "loss": 0.0152,
      "step": 652
    },
    {
      "epoch": 8.86,
      "learning_rate": 0.0001,
      "loss": 0.0118,
      "step": 656
    },
    {
      "epoch": 8.91,
      "learning_rate": 0.0001,
      "loss": 0.0114,
      "step": 660
    },
    {
      "epoch": 8.96,
      "learning_rate": 0.0001,
      "loss": 0.0116,
      "step": 664
    },
    {
      "epoch": 9.03,
      "learning_rate": 0.0001,
      "loss": 0.015,
      "step": 668
    },
    {
      "epoch": 9.08,
      "learning_rate": 0.0001,
      "loss": 0.0127,
      "step": 672
    },
    {
      "epoch": 9.13,
      "learning_rate": 0.0001,
      "loss": 0.0098,
      "step": 676
    },
    {
      "epoch": 9.19,
      "learning_rate": 0.0001,
      "loss": 0.009,
      "step": 680
    },
    {
      "epoch": 9.24,
      "learning_rate": 0.0001,
      "loss": 0.0101,
      "step": 684
    },
    {
      "epoch": 9.29,
      "learning_rate": 0.0001,
      "loss": 0.0211,
      "step": 688
    },
    {
      "epoch": 9.35,
      "learning_rate": 0.0001,
      "loss": 0.0141,
      "step": 692
    },
    {
      "epoch": 9.4,
      "learning_rate": 0.0001,
      "loss": 0.0089,
      "step": 696
    },
    {
      "epoch": 9.45,
      "learning_rate": 0.0001,
      "loss": 0.008,
      "step": 700
    },
    {
      "epoch": 9.51,
      "learning_rate": 0.0001,
      "loss": 0.0118,
      "step": 704
    },
    {
      "epoch": 9.51,
      "eval_exec": 0.688588007736944,
      "eval_loss": 0.2283668965101242,
      "eval_runtime": 296.2258,
      "eval_samples_per_second": 3.491,
      "step": 704
    },
    {
      "epoch": 9.56,
      "learning_rate": 0.0001,
      "loss": 0.0156,
      "step": 708
    },
    {
      "epoch": 9.62,
      "learning_rate": 0.0001,
      "loss": 0.0082,
      "step": 712
    },
    {
      "epoch": 9.67,
      "learning_rate": 0.0001,
      "loss": 0.0075,
      "step": 716
    },
    {
      "epoch": 9.72,
      "learning_rate": 0.0001,
      "loss": 0.01,
      "step": 720
    },
    {
      "epoch": 9.78,
      "learning_rate": 0.0001,
      "loss": 0.0111,
      "step": 724
    },
    {
      "epoch": 9.83,
      "learning_rate": 0.0001,
      "loss": 0.0103,
      "step": 728
    },
    {
      "epoch": 9.88,
      "learning_rate": 0.0001,
      "loss": 0.0084,
      "step": 732
    },
    {
      "epoch": 9.94,
      "learning_rate": 0.0001,
      "loss": 0.0063,
      "step": 736
    },
    {
      "epoch": 9.99,
      "learning_rate": 0.0001,
      "loss": 0.0091,
      "step": 740
    },
    {
      "epoch": 10.05,
      "learning_rate": 0.0001,
      "loss": 0.0119,
      "step": 744
    },
    {
      "epoch": 10.11,
      "learning_rate": 0.0001,
      "loss": 0.0089,
      "step": 748
    },
    {
      "epoch": 10.16,
      "learning_rate": 0.0001,
      "loss": 0.008,
      "step": 752
    },
    {
      "epoch": 10.21,
      "learning_rate": 0.0001,
      "loss": 0.0079,
      "step": 756
    },
    {
      "epoch": 10.27,
      "learning_rate": 0.0001,
      "loss": 0.0103,
      "step": 760
    },
    {
      "epoch": 10.32,
      "learning_rate": 0.0001,
      "loss": 0.0134,
      "step": 764
    },
    {
      "epoch": 10.37,
      "learning_rate": 0.0001,
      "loss": 0.0088,
      "step": 768
    },
    {
      "epoch": 10.37,
      "eval_exec": 0.7147001934235977,
      "eval_loss": 0.22069412469863892,
      "eval_runtime": 296.2474,
      "eval_samples_per_second": 3.49,
      "step": 768
    },
    {
      "epoch": 10.43,
      "learning_rate": 0.0001,
      "loss": 0.0079,
      "step": 772
    },
    {
      "epoch": 10.48,
      "learning_rate": 0.0001,
      "loss": 0.0079,
      "step": 776
    },
    {
      "epoch": 10.53,
      "learning_rate": 0.0001,
      "loss": 0.0126,
      "step": 780
    },
    {
      "epoch": 10.59,
      "learning_rate": 0.0001,
      "loss": 0.0084,
      "step": 784
    },
    {
      "epoch": 10.64,
      "learning_rate": 0.0001,
      "loss": 0.0084,
      "step": 788
    },
    {
      "epoch": 10.7,
      "learning_rate": 0.0001,
      "loss": 0.0056,
      "step": 792
    },
    {
      "epoch": 10.75,
      "learning_rate": 0.0001,
      "loss": 0.0122,
      "step": 796
    },
    {
      "epoch": 10.8,
      "learning_rate": 0.0001,
      "loss": 0.0145,
      "step": 800
    },
    {
      "epoch": 10.86,
      "learning_rate": 0.0001,
      "loss": 0.0076,
      "step": 804
    },
    {
      "epoch": 10.91,
      "learning_rate": 0.0001,
      "loss": 0.0071,
      "step": 808
    },
    {
      "epoch": 10.96,
      "learning_rate": 0.0001,
      "loss": 0.0107,
      "step": 812
    },
    {
      "epoch": 11.03,
      "learning_rate": 0.0001,
      "loss": 0.0146,
      "step": 816
    },
    {
      "epoch": 11.08,
      "learning_rate": 0.0001,
      "loss": 0.008,
      "step": 820
    },
    {
      "epoch": 11.13,
      "learning_rate": 0.0001,
      "loss": 0.0057,
      "step": 824
    },
    {
      "epoch": 11.19,
      "learning_rate": 0.0001,
      "loss": 0.0054,
      "step": 828
    },
    {
      "epoch": 11.24,
      "learning_rate": 0.0001,
      "loss": 0.0059,
      "step": 832
    },
    {
      "epoch": 11.24,
      "eval_exec": 0.7098646034816247,
      "eval_loss": 0.25100022554397583,
      "eval_runtime": 292.4465,
      "eval_samples_per_second": 3.536,
      "step": 832
    },
    {
      "epoch": 11.29,
      "learning_rate": 0.0001,
      "loss": 0.0096,
      "step": 836
    },
    {
      "epoch": 11.35,
      "learning_rate": 0.0001,
      "loss": 0.0079,
      "step": 840
    },
    {
      "epoch": 11.4,
      "learning_rate": 0.0001,
      "loss": 0.0059,
      "step": 844
    },
    {
      "epoch": 11.45,
      "learning_rate": 0.0001,
      "loss": 0.0099,
      "step": 848
    },
    {
      "epoch": 11.51,
      "learning_rate": 0.0001,
      "loss": 0.0133,
      "step": 852
    },
    {
      "epoch": 11.56,
      "learning_rate": 0.0001,
      "loss": 0.0087,
      "step": 856
    },
    {
      "epoch": 11.62,
      "learning_rate": 0.0001,
      "loss": 0.0073,
      "step": 860
    },
    {
      "epoch": 11.67,
      "learning_rate": 0.0001,
      "loss": 0.0048,
      "step": 864
    },
    {
      "epoch": 11.72,
      "learning_rate": 0.0001,
      "loss": 0.007,
      "step": 868
    },
    {
      "epoch": 11.78,
      "learning_rate": 0.0001,
      "loss": 0.0098,
      "step": 872
    },
    {
      "epoch": 11.83,
      "learning_rate": 0.0001,
      "loss": 0.0119,
      "step": 876
    },
    {
      "epoch": 11.88,
      "learning_rate": 0.0001,
      "loss": 0.0121,
      "step": 880
    },
    {
      "epoch": 11.94,
      "learning_rate": 0.0001,
      "loss": 0.0059,
      "step": 884
    },
    {
      "epoch": 11.99,
      "learning_rate": 0.0001,
      "loss": 0.0103,
      "step": 888
    },
    {
      "epoch": 12.05,
      "learning_rate": 0.0001,
      "loss": 0.0072,
      "step": 892
    },
    {
      "epoch": 12.11,
      "learning_rate": 0.0001,
      "loss": 0.0055,
      "step": 896
    },
    {
      "epoch": 12.11,
      "eval_exec": 0.7098646034816247,
      "eval_loss": 0.24743221700191498,
      "eval_runtime": 295.2581,
      "eval_samples_per_second": 3.502,
      "step": 896
    },
    {
      "epoch": 12.16,
      "learning_rate": 0.0001,
      "loss": 0.0068,
      "step": 900
    },
    {
      "epoch": 12.21,
      "learning_rate": 0.0001,
      "loss": 0.0034,
      "step": 904
    },
    {
      "epoch": 12.27,
      "learning_rate": 0.0001,
      "loss": 0.0081,
      "step": 908
    },
    {
      "epoch": 12.32,
      "learning_rate": 0.0001,
      "loss": 0.0095,
      "step": 912
    },
    {
      "epoch": 12.37,
      "learning_rate": 0.0001,
      "loss": 0.0069,
      "step": 916
    },
    {
      "epoch": 12.43,
      "learning_rate": 0.0001,
      "loss": 0.0091,
      "step": 920
    },
    {
      "epoch": 12.48,
      "learning_rate": 0.0001,
      "loss": 0.0062,
      "step": 924
    },
    {
      "epoch": 12.53,
      "learning_rate": 0.0001,
      "loss": 0.0084,
      "step": 928
    },
    {
      "epoch": 12.59,
      "learning_rate": 0.0001,
      "loss": 0.0064,
      "step": 932
    },
    {
      "epoch": 12.64,
      "learning_rate": 0.0001,
      "loss": 0.0057,
      "step": 936
    },
    {
      "epoch": 12.7,
      "learning_rate": 0.0001,
      "loss": 0.0075,
      "step": 940
    },
    {
      "epoch": 12.75,
      "learning_rate": 0.0001,
      "loss": 0.0091,
      "step": 944
    },
    {
      "epoch": 12.8,
      "learning_rate": 0.0001,
      "loss": 0.0078,
      "step": 948
    },
    {
      "epoch": 12.86,
      "learning_rate": 0.0001,
      "loss": 0.0061,
      "step": 952
    },
    {
      "epoch": 12.91,
      "learning_rate": 0.0001,
      "loss": 0.0098,
      "step": 956
    },
    {
      "epoch": 12.96,
      "learning_rate": 0.0001,
      "loss": 0.0072,
      "step": 960
    },
    {
      "epoch": 12.96,
      "eval_exec": 0.6924564796905223,
      "eval_loss": 0.25289151072502136,
      "eval_runtime": 288.8532,
      "eval_samples_per_second": 3.58,
      "step": 960
    },
    {
      "epoch": 13.03,
      "learning_rate": 0.0001,
      "loss": 0.0111,
      "step": 964
    },
    {
      "epoch": 13.08,
      "learning_rate": 0.0001,
      "loss": 0.0048,
      "step": 968
    },
    {
      "epoch": 13.13,
      "learning_rate": 0.0001,
      "loss": 0.0054,
      "step": 972
    },
    {
      "epoch": 13.19,
      "learning_rate": 0.0001,
      "loss": 0.0043,
      "step": 976
    },
    {
      "epoch": 13.24,
      "learning_rate": 0.0001,
      "loss": 0.0118,
      "step": 980
    },
    {
      "epoch": 13.29,
      "learning_rate": 0.0001,
      "loss": 0.0069,
      "step": 984
    },
    {
      "epoch": 13.35,
      "learning_rate": 0.0001,
      "loss": 0.0057,
      "step": 988
    },
    {
      "epoch": 13.4,
      "learning_rate": 0.0001,
      "loss": 0.0054,
      "step": 992
    },
    {
      "epoch": 13.45,
      "learning_rate": 0.0001,
      "loss": 0.0035,
      "step": 996
    },
    {
      "epoch": 13.51,
      "learning_rate": 0.0001,
      "loss": 0.0041,
      "step": 1000
    },
    {
      "epoch": 13.56,
      "learning_rate": 0.0001,
      "loss": 0.0052,
      "step": 1004
    },
    {
      "epoch": 13.62,
      "learning_rate": 0.0001,
      "loss": 0.0068,
      "step": 1008
    },
    {
      "epoch": 13.67,
      "learning_rate": 0.0001,
      "loss": 0.0039,
      "step": 1012
    },
    {
      "epoch": 13.72,
      "learning_rate": 0.0001,
      "loss": 0.0052,
      "step": 1016
    },
    {
      "epoch": 13.78,
      "learning_rate": 0.0001,
      "loss": 0.0075,
      "step": 1020
    },
    {
      "epoch": 13.83,
      "learning_rate": 0.0001,
      "loss": 0.0078,
      "step": 1024
    },
    {
      "epoch": 13.83,
      "eval_exec": 0.7311411992263056,
      "eval_loss": 0.26111745834350586,
      "eval_runtime": 286.014,
      "eval_samples_per_second": 3.615,
      "step": 1024
    },
    {
      "epoch": 13.88,
      "learning_rate": 0.0001,
      "loss": 0.0059,
      "step": 1028
    },
    {
      "epoch": 13.94,
      "learning_rate": 0.0001,
      "loss": 0.0046,
      "step": 1032
    },
    {
      "epoch": 13.99,
      "learning_rate": 0.0001,
      "loss": 0.007,
      "step": 1036
    },
    {
      "epoch": 14.05,
      "learning_rate": 0.0001,
      "loss": 0.0079,
      "step": 1040
    },
    {
      "epoch": 14.11,
      "learning_rate": 0.0001,
      "loss": 0.0056,
      "step": 1044
    },
    {
      "epoch": 14.16,
      "learning_rate": 0.0001,
      "loss": 0.005,
      "step": 1048
    },
    {
      "epoch": 14.21,
      "learning_rate": 0.0001,
      "loss": 0.0039,
      "step": 1052
    },
    {
      "epoch": 14.27,
      "learning_rate": 0.0001,
      "loss": 0.0043,
      "step": 1056
    },
    {
      "epoch": 14.32,
      "learning_rate": 0.0001,
      "loss": 0.005,
      "step": 1060
    },
    {
      "epoch": 14.37,
      "learning_rate": 0.0001,
      "loss": 0.0047,
      "step": 1064
    },
    {
      "epoch": 14.43,
      "learning_rate": 0.0001,
      "loss": 0.0026,
      "step": 1068
    },
    {
      "epoch": 14.48,
      "learning_rate": 0.0001,
      "loss": 0.005,
      "step": 1072
    },
    {
      "epoch": 14.53,
      "learning_rate": 0.0001,
      "loss": 0.0073,
      "step": 1076
    },
    {
      "epoch": 14.59,
      "learning_rate": 0.0001,
      "loss": 0.0069,
      "step": 1080
    },
    {
      "epoch": 14.64,
      "learning_rate": 0.0001,
      "loss": 0.0043,
      "step": 1084
    },
    {
      "epoch": 14.7,
      "learning_rate": 0.0001,
      "loss": 0.0035,
      "step": 1088
    },
    {
      "epoch": 14.7,
      "eval_exec": 0.7205029013539652,
      "eval_loss": 0.2667447030544281,
      "eval_runtime": 283.0779,
      "eval_samples_per_second": 3.653,
      "step": 1088
    },
    {
      "epoch": 14.75,
      "learning_rate": 0.0001,
      "loss": 0.0076,
      "step": 1092
    },
    {
      "epoch": 14.8,
      "learning_rate": 0.0001,
      "loss": 0.0083,
      "step": 1096
    },
    {
      "epoch": 14.86,
      "learning_rate": 0.0001,
      "loss": 0.0064,
      "step": 1100
    },
    {
      "epoch": 14.91,
      "learning_rate": 0.0001,
      "loss": 0.0042,
      "step": 1104
    },
    {
      "epoch": 14.96,
      "learning_rate": 0.0001,
      "loss": 0.0066,
      "step": 1108
    },
    {
      "epoch": 15.03,
      "learning_rate": 0.0001,
      "loss": 0.0053,
      "step": 1112
    },
    {
      "epoch": 15.08,
      "learning_rate": 0.0001,
      "loss": 0.005,
      "step": 1116
    },
    {
      "epoch": 15.13,
      "learning_rate": 0.0001,
      "loss": 0.004,
      "step": 1120
    },
    {
      "epoch": 15.19,
      "learning_rate": 0.0001,
      "loss": 0.0038,
      "step": 1124
    },
    {
      "epoch": 15.24,
      "learning_rate": 0.0001,
      "loss": 0.005,
      "step": 1128
    },
    {
      "epoch": 15.29,
      "learning_rate": 0.0001,
      "loss": 0.0064,
      "step": 1132
    },
    {
      "epoch": 15.35,
      "learning_rate": 0.0001,
      "loss": 0.0048,
      "step": 1136
    },
    {
      "epoch": 15.4,
      "learning_rate": 0.0001,
      "loss": 0.0034,
      "step": 1140
    },
    {
      "epoch": 15.45,
      "learning_rate": 0.0001,
      "loss": 0.0034,
      "step": 1144
    },
    {
      "epoch": 15.51,
      "learning_rate": 0.0001,
      "loss": 0.005,
      "step": 1148
    },
    {
      "epoch": 15.56,
      "learning_rate": 0.0001,
      "loss": 0.0078,
      "step": 1152
    },
    {
      "epoch": 15.56,
      "eval_exec": 0.730174081237911,
      "eval_loss": 0.24596308171749115,
      "eval_runtime": 290.0696,
      "eval_samples_per_second": 3.565,
      "step": 1152
    },
    {
      "epoch": 15.62,
      "learning_rate": 0.0001,
      "loss": 0.0032,
      "step": 1156
    },
    {
      "epoch": 15.67,
      "learning_rate": 0.0001,
      "loss": 0.0025,
      "step": 1160
    },
    {
      "epoch": 15.72,
      "learning_rate": 0.0001,
      "loss": 0.0041,
      "step": 1164
    },
    {
      "epoch": 15.78,
      "learning_rate": 0.0001,
      "loss": 0.0058,
      "step": 1168
    },
    {
      "epoch": 15.83,
      "learning_rate": 0.0001,
      "loss": 0.0082,
      "step": 1172
    },
    {
      "epoch": 15.88,
      "learning_rate": 0.0001,
      "loss": 0.0037,
      "step": 1176
    },
    {
      "epoch": 15.94,
      "learning_rate": 0.0001,
      "loss": 0.0021,
      "step": 1180
    },
    {
      "epoch": 15.99,
      "learning_rate": 0.0001,
      "loss": 0.0046,
      "step": 1184
    },
    {
      "epoch": 16.05,
      "learning_rate": 0.0001,
      "loss": 0.0045,
      "step": 1188
    },
    {
      "epoch": 16.11,
      "learning_rate": 0.0001,
      "loss": 0.0038,
      "step": 1192
    },
    {
      "epoch": 16.16,
      "learning_rate": 0.0001,
      "loss": 0.0032,
      "step": 1196
    },
    {
      "epoch": 16.21,
      "learning_rate": 0.0001,
      "loss": 0.0048,
      "step": 1200
    },
    {
      "epoch": 16.27,
      "learning_rate": 0.0001,
      "loss": 0.0035,
      "step": 1204
    },
    {
      "epoch": 16.32,
      "learning_rate": 0.0001,
      "loss": 0.0044,
      "step": 1208
    },
    {
      "epoch": 16.37,
      "learning_rate": 0.0001,
      "loss": 0.0035,
      "step": 1212
    },
    {
      "epoch": 16.43,
      "learning_rate": 0.0001,
      "loss": 0.0014,
      "step": 1216
    },
    {
      "epoch": 16.43,
      "eval_exec": 0.7050290135396519,
      "eval_loss": 0.28547874093055725,
      "eval_runtime": 289.8855,
      "eval_samples_per_second": 3.567,
      "step": 1216
    },
    {
      "epoch": 16.48,
      "learning_rate": 0.0001,
      "loss": 0.003,
      "step": 1220
    },
    {
      "epoch": 16.53,
      "learning_rate": 0.0001,
      "loss": 0.0044,
      "step": 1224
    },
    {
      "epoch": 16.59,
      "learning_rate": 0.0001,
      "loss": 0.0046,
      "step": 1228
    },
    {
      "epoch": 16.64,
      "learning_rate": 0.0001,
      "loss": 0.0037,
      "step": 1232
    },
    {
      "epoch": 16.7,
      "learning_rate": 0.0001,
      "loss": 0.0025,
      "step": 1236
    },
    {
      "epoch": 16.75,
      "learning_rate": 0.0001,
      "loss": 0.0052,
      "step": 1240
    },
    {
      "epoch": 16.8,
      "learning_rate": 0.0001,
      "loss": 0.0054,
      "step": 1244
    },
    {
      "epoch": 16.86,
      "learning_rate": 0.0001,
      "loss": 0.0045,
      "step": 1248
    },
    {
      "epoch": 16.91,
      "learning_rate": 0.0001,
      "loss": 0.004,
      "step": 1252
    },
    {
      "epoch": 16.96,
      "learning_rate": 0.0001,
      "loss": 0.0049,
      "step": 1256
    },
    {
      "epoch": 17.03,
      "learning_rate": 0.0001,
      "loss": 0.0049,
      "step": 1260
    },
    {
      "epoch": 17.08,
      "learning_rate": 0.0001,
      "loss": 0.0049,
      "step": 1264
    },
    {
      "epoch": 17.13,
      "learning_rate": 0.0001,
      "loss": 0.0038,
      "step": 1268
    },
    {
      "epoch": 17.19,
      "learning_rate": 0.0001,
      "loss": 0.0033,
      "step": 1272
    },
    {
      "epoch": 17.24,
      "learning_rate": 0.0001,
      "loss": 0.0029,
      "step": 1276
    },
    {
      "epoch": 17.29,
      "learning_rate": 0.0001,
      "loss": 0.0043,
      "step": 1280
    },
    {
      "epoch": 17.29,
      "eval_exec": 0.695357833655706,
      "eval_loss": 0.2833567261695862,
      "eval_runtime": 296.5011,
      "eval_samples_per_second": 3.487,
      "step": 1280
    },
    {
      "epoch": 17.35,
      "learning_rate": 0.0001,
      "loss": 0.004,
      "step": 1284
    },
    {
      "epoch": 17.4,
      "learning_rate": 0.0001,
      "loss": 0.0065,
      "step": 1288
    },
    {
      "epoch": 17.45,
      "learning_rate": 0.0001,
      "loss": 0.0024,
      "step": 1292
    },
    {
      "epoch": 17.51,
      "learning_rate": 0.0001,
      "loss": 0.0046,
      "step": 1296
    },
    {
      "epoch": 17.56,
      "learning_rate": 0.0001,
      "loss": 0.0047,
      "step": 1300
    },
    {
      "epoch": 17.62,
      "learning_rate": 0.0001,
      "loss": 0.0048,
      "step": 1304
    },
    {
      "epoch": 17.67,
      "learning_rate": 0.0001,
      "loss": 0.0029,
      "step": 1308
    },
    {
      "epoch": 17.72,
      "learning_rate": 0.0001,
      "loss": 0.0035,
      "step": 1312
    },
    {
      "epoch": 17.78,
      "learning_rate": 0.0001,
      "loss": 0.0054,
      "step": 1316
    },
    {
      "epoch": 17.83,
      "learning_rate": 0.0001,
      "loss": 0.006,
      "step": 1320
    },
    {
      "epoch": 17.88,
      "learning_rate": 0.0001,
      "loss": 0.0042,
      "step": 1324
    },
    {
      "epoch": 17.94,
      "learning_rate": 0.0001,
      "loss": 0.0024,
      "step": 1328
    },
    {
      "epoch": 17.99,
      "learning_rate": 0.0001,
      "loss": 0.0053,
      "step": 1332
    },
    {
      "epoch": 18.05,
      "learning_rate": 0.0001,
      "loss": 0.0199,
      "step": 1336
    },
    {
      "epoch": 18.11,
      "learning_rate": 0.0001,
      "loss": 0.008,
      "step": 1340
    },
    {
      "epoch": 18.16,
      "learning_rate": 0.0001,
      "loss": 0.0046,
      "step": 1344
    },
    {
      "epoch": 18.16,
      "eval_exec": 0.7021276595744681,
      "eval_loss": 0.2530258595943451,
      "eval_runtime": 289.9942,
      "eval_samples_per_second": 3.566,
      "step": 1344
    },
    {
      "epoch": 18.21,
      "learning_rate": 0.0001,
      "loss": 0.0028,
      "step": 1348
    },
    {
      "epoch": 18.27,
      "learning_rate": 0.0001,
      "loss": 0.0039,
      "step": 1352
    },
    {
      "epoch": 18.32,
      "learning_rate": 0.0001,
      "loss": 0.0045,
      "step": 1356
    },
    {
      "epoch": 18.37,
      "learning_rate": 0.0001,
      "loss": 0.0032,
      "step": 1360
    },
    {
      "epoch": 18.43,
      "learning_rate": 0.0001,
      "loss": 0.003,
      "step": 1364
    },
    {
      "epoch": 18.48,
      "learning_rate": 0.0001,
      "loss": 0.0031,
      "step": 1368
    },
    {
      "epoch": 18.53,
      "learning_rate": 0.0001,
      "loss": 0.0054,
      "step": 1372
    },
    {
      "epoch": 18.59,
      "learning_rate": 0.0001,
      "loss": 0.0037,
      "step": 1376
    },
    {
      "epoch": 18.64,
      "learning_rate": 0.0001,
      "loss": 0.0039,
      "step": 1380
    },
    {
      "epoch": 18.7,
      "learning_rate": 0.0001,
      "loss": 0.0029,
      "step": 1384
    },
    {
      "epoch": 18.75,
      "learning_rate": 0.0001,
      "loss": 0.0037,
      "step": 1388
    },
    {
      "epoch": 18.8,
      "learning_rate": 0.0001,
      "loss": 0.0043,
      "step": 1392
    },
    {
      "epoch": 18.86,
      "learning_rate": 0.0001,
      "loss": 0.0026,
      "step": 1396
    },
    {
      "epoch": 18.91,
      "learning_rate": 0.0001,
      "loss": 0.0027,
      "step": 1400
    },
    {
      "epoch": 18.96,
      "learning_rate": 0.0001,
      "loss": 0.0036,
      "step": 1404
    },
    {
      "epoch": 19.03,
      "learning_rate": 0.0001,
      "loss": 0.004,
      "step": 1408
    },
    {
      "epoch": 19.03,
      "eval_exec": 0.7330754352030948,
      "eval_loss": 0.2696512043476105,
      "eval_runtime": 291.6539,
      "eval_samples_per_second": 3.545,
      "step": 1408
    },
    {
      "epoch": 19.08,
      "learning_rate": 0.0001,
      "loss": 0.0049,
      "step": 1412
    },
    {
      "epoch": 19.13,
      "learning_rate": 0.0001,
      "loss": 0.0015,
      "step": 1416
    },
    {
      "epoch": 19.19,
      "learning_rate": 0.0001,
      "loss": 0.0018,
      "step": 1420
    },
    {
      "epoch": 19.24,
      "learning_rate": 0.0001,
      "loss": 0.0024,
      "step": 1424
    },
    {
      "epoch": 19.29,
      "learning_rate": 0.0001,
      "loss": 0.0046,
      "step": 1428
    },
    {
      "epoch": 19.35,
      "learning_rate": 0.0001,
      "loss": 0.0035,
      "step": 1432
    },
    {
      "epoch": 19.4,
      "learning_rate": 0.0001,
      "loss": 0.0019,
      "step": 1436
    },
    {
      "epoch": 19.45,
      "learning_rate": 0.0001,
      "loss": 0.0018,
      "step": 1440
    },
    {
      "epoch": 19.51,
      "learning_rate": 0.0001,
      "loss": 0.0045,
      "step": 1444
    },
    {
      "epoch": 19.56,
      "learning_rate": 0.0001,
      "loss": 0.0036,
      "step": 1448
    },
    {
      "epoch": 19.62,
      "learning_rate": 0.0001,
      "loss": 0.002,
      "step": 1452
    },
    {
      "epoch": 19.67,
      "learning_rate": 0.0001,
      "loss": 0.0024,
      "step": 1456
    },
    {
      "epoch": 19.72,
      "learning_rate": 0.0001,
      "loss": 0.01,
      "step": 1460
    },
    {
      "epoch": 19.78,
      "learning_rate": 0.0001,
      "loss": 0.0051,
      "step": 1464
    },
    {
      "epoch": 19.83,
      "learning_rate": 0.0001,
      "loss": 0.0039,
      "step": 1468
    },
    {
      "epoch": 19.88,
      "learning_rate": 0.0001,
      "loss": 0.0036,
      "step": 1472
    },
    {
      "epoch": 19.88,
      "eval_exec": 0.7224371373307543,
      "eval_loss": 0.2764066755771637,
      "eval_runtime": 304.8058,
      "eval_samples_per_second": 3.392,
      "step": 1472
    },
    {
      "epoch": 19.94,
      "learning_rate": 0.0001,
      "loss": 0.0024,
      "step": 1476
    },
    {
      "epoch": 19.99,
      "learning_rate": 0.0001,
      "loss": 0.0038,
      "step": 1480
    },
    {
      "epoch": 20.05,
      "learning_rate": 0.0001,
      "loss": 0.0041,
      "step": 1484
    },
    {
      "epoch": 20.11,
      "learning_rate": 0.0001,
      "loss": 0.004,
      "step": 1488
    },
    {
      "epoch": 20.16,
      "learning_rate": 0.0001,
      "loss": 0.002,
      "step": 1492
    },
    {
      "epoch": 20.21,
      "learning_rate": 0.0001,
      "loss": 0.0036,
      "step": 1496
    },
    {
      "epoch": 20.27,
      "learning_rate": 0.0001,
      "loss": 0.0031,
      "step": 1500
    },
    {
      "epoch": 20.32,
      "learning_rate": 0.0001,
      "loss": 0.0031,
      "step": 1504
    },
    {
      "epoch": 20.37,
      "learning_rate": 0.0001,
      "loss": 0.0025,
      "step": 1508
    },
    {
      "epoch": 20.43,
      "learning_rate": 0.0001,
      "loss": 0.0042,
      "step": 1512
    },
    {
      "epoch": 20.48,
      "learning_rate": 0.0001,
      "loss": 0.0035,
      "step": 1516
    },
    {
      "epoch": 20.53,
      "learning_rate": 0.0001,
      "loss": 0.0074,
      "step": 1520
    },
    {
      "epoch": 20.59,
      "learning_rate": 0.0001,
      "loss": 0.0039,
      "step": 1524
    },
    {
      "epoch": 20.64,
      "learning_rate": 0.0001,
      "loss": 0.0025,
      "step": 1528
    },
    {
      "epoch": 20.7,
      "learning_rate": 0.0001,
      "loss": 0.0031,
      "step": 1532
    },
    {
      "epoch": 20.75,
      "learning_rate": 0.0001,
      "loss": 0.0035,
      "step": 1536
    },
    {
      "epoch": 20.75,
      "eval_exec": 0.7156673114119922,
      "eval_loss": 0.2632952034473419,
      "eval_runtime": 295.0708,
      "eval_samples_per_second": 3.504,
      "step": 1536
    },
    {
      "epoch": 20.8,
      "learning_rate": 0.0001,
      "loss": 0.0031,
      "step": 1540
    },
    {
      "epoch": 20.86,
      "learning_rate": 0.0001,
      "loss": 0.0032,
      "step": 1544
    },
    {
      "epoch": 20.91,
      "learning_rate": 0.0001,
      "loss": 0.0027,
      "step": 1548
    },
    {
      "epoch": 20.96,
      "learning_rate": 0.0001,
      "loss": 0.0038,
      "step": 1552
    },
    {
      "epoch": 21.03,
      "learning_rate": 0.0001,
      "loss": 0.0034,
      "step": 1556
    },
    {
      "epoch": 21.08,
      "learning_rate": 0.0001,
      "loss": 0.0034,
      "step": 1560
    },
    {
      "epoch": 21.13,
      "learning_rate": 0.0001,
      "loss": 0.0053,
      "step": 1564
    },
    {
      "epoch": 21.19,
      "learning_rate": 0.0001,
      "loss": 0.0015,
      "step": 1568
    },
    {
      "epoch": 21.24,
      "learning_rate": 0.0001,
      "loss": 0.0022,
      "step": 1572
    },
    {
      "epoch": 21.29,
      "learning_rate": 0.0001,
      "loss": 0.003,
      "step": 1576
    },
    {
      "epoch": 21.35,
      "learning_rate": 0.0001,
      "loss": 0.0026,
      "step": 1580
    },
    {
      "epoch": 21.4,
      "learning_rate": 0.0001,
      "loss": 0.0028,
      "step": 1584
    },
    {
      "epoch": 21.45,
      "learning_rate": 0.0001,
      "loss": 0.0038,
      "step": 1588
    },
    {
      "epoch": 21.51,
      "learning_rate": 0.0001,
      "loss": 0.0029,
      "step": 1592
    },
    {
      "epoch": 21.56,
      "learning_rate": 0.0001,
      "loss": 0.0039,
      "step": 1596
    },
    {
      "epoch": 21.62,
      "learning_rate": 0.0001,
      "loss": 0.0047,
      "step": 1600
    },
    {
      "epoch": 21.62,
      "eval_exec": 0.7224371373307543,
      "eval_loss": 0.2584507465362549,
      "eval_runtime": 296.7113,
      "eval_samples_per_second": 3.485,
      "step": 1600
    },
    {
      "epoch": 21.67,
      "learning_rate": 0.0001,
      "loss": 0.0019,
      "step": 1604
    },
    {
      "epoch": 21.72,
      "learning_rate": 0.0001,
      "loss": 0.0027,
      "step": 1608
    },
    {
      "epoch": 21.78,
      "learning_rate": 0.0001,
      "loss": 0.0031,
      "step": 1612
    },
    {
      "epoch": 21.83,
      "learning_rate": 0.0001,
      "loss": 0.0023,
      "step": 1616
    },
    {
      "epoch": 21.88,
      "learning_rate": 0.0001,
      "loss": 0.0023,
      "step": 1620
    },
    {
      "epoch": 21.94,
      "learning_rate": 0.0001,
      "loss": 0.0041,
      "step": 1624
    },
    {
      "epoch": 21.99,
      "learning_rate": 0.0001,
      "loss": 0.0027,
      "step": 1628
    },
    {
      "epoch": 22.05,
      "learning_rate": 0.0001,
      "loss": 0.0027,
      "step": 1632
    },
    {
      "epoch": 22.11,
      "learning_rate": 0.0001,
      "loss": 0.0019,
      "step": 1636
    },
    {
      "epoch": 22.16,
      "learning_rate": 0.0001,
      "loss": 0.0024,
      "step": 1640
    },
    {
      "epoch": 22.21,
      "learning_rate": 0.0001,
      "loss": 0.002,
      "step": 1644
    },
    {
      "epoch": 22.27,
      "learning_rate": 0.0001,
      "loss": 0.0027,
      "step": 1648
    },
    {
      "epoch": 22.32,
      "learning_rate": 0.0001,
      "loss": 0.0033,
      "step": 1652
    },
    {
      "epoch": 22.37,
      "learning_rate": 0.0001,
      "loss": 0.0026,
      "step": 1656
    },
    {
      "epoch": 22.43,
      "learning_rate": 0.0001,
      "loss": 0.0027,
      "step": 1660
    },
    {
      "epoch": 22.48,
      "learning_rate": 0.0001,
      "loss": 0.0042,
      "step": 1664
    },
    {
      "epoch": 22.48,
      "eval_exec": 0.7263056092843327,
      "eval_loss": 0.262117475271225,
      "eval_runtime": 293.7879,
      "eval_samples_per_second": 3.52,
      "step": 1664
    },
    {
      "epoch": 22.53,
      "learning_rate": 0.0001,
      "loss": 0.0045,
      "step": 1668
    },
    {
      "epoch": 22.59,
      "learning_rate": 0.0001,
      "loss": 0.0038,
      "step": 1672
    },
    {
      "epoch": 22.64,
      "learning_rate": 0.0001,
      "loss": 0.0029,
      "step": 1676
    },
    {
      "epoch": 22.7,
      "learning_rate": 0.0001,
      "loss": 0.0019,
      "step": 1680
    },
    {
      "epoch": 22.75,
      "learning_rate": 0.0001,
      "loss": 0.0041,
      "step": 1684
    },
    {
      "epoch": 22.8,
      "learning_rate": 0.0001,
      "loss": 0.0178,
      "step": 1688
    },
    {
      "epoch": 22.86,
      "learning_rate": 0.0001,
      "loss": 0.0028,
      "step": 1692
    },
    {
      "epoch": 22.91,
      "learning_rate": 0.0001,
      "loss": 0.0032,
      "step": 1696
    },
    {
      "epoch": 22.96,
      "learning_rate": 0.0001,
      "loss": 0.0035,
      "step": 1700
    },
    {
      "epoch": 23.03,
      "learning_rate": 0.0001,
      "loss": 0.0037,
      "step": 1704
    },
    {
      "epoch": 23.08,
      "learning_rate": 0.0001,
      "loss": 0.0026,
      "step": 1708
    },
    {
      "epoch": 23.13,
      "learning_rate": 0.0001,
      "loss": 0.002,
      "step": 1712
    },
    {
      "epoch": 23.19,
      "learning_rate": 0.0001,
      "loss": 0.0023,
      "step": 1716
    },
    {
      "epoch": 23.24,
      "learning_rate": 0.0001,
      "loss": 0.0017,
      "step": 1720
    },
    {
      "epoch": 23.29,
      "learning_rate": 0.0001,
      "loss": 0.0026,
      "step": 1724
    },
    {
      "epoch": 23.35,
      "learning_rate": 0.0001,
      "loss": 0.0026,
      "step": 1728
    },
    {
      "epoch": 23.35,
      "eval_exec": 0.7205029013539652,
      "eval_loss": 0.2860746383666992,
      "eval_runtime": 301.1885,
      "eval_samples_per_second": 3.433,
      "step": 1728
    },
    {
      "epoch": 23.4,
      "learning_rate": 0.0001,
      "loss": 0.0043,
      "step": 1732
    },
    {
      "epoch": 23.45,
      "learning_rate": 0.0001,
      "loss": 0.0024,
      "step": 1736
    },
    {
      "epoch": 23.51,
      "learning_rate": 0.0001,
      "loss": 0.0026,
      "step": 1740
    },
    {
      "epoch": 23.56,
      "learning_rate": 0.0001,
      "loss": 0.0032,
      "step": 1744
    },
    {
      "epoch": 23.62,
      "learning_rate": 0.0001,
      "loss": 0.0026,
      "step": 1748
    },
    {
      "epoch": 23.67,
      "learning_rate": 0.0001,
      "loss": 0.0022,
      "step": 1752
    },
    {
      "epoch": 23.72,
      "learning_rate": 0.0001,
      "loss": 0.0018,
      "step": 1756
    },
    {
      "epoch": 23.78,
      "learning_rate": 0.0001,
      "loss": 0.0024,
      "step": 1760
    },
    {
      "epoch": 23.83,
      "learning_rate": 0.0001,
      "loss": 0.0022,
      "step": 1764
    },
    {
      "epoch": 23.88,
      "learning_rate": 0.0001,
      "loss": 0.0032,
      "step": 1768
    },
    {
      "epoch": 23.94,
      "learning_rate": 0.0001,
      "loss": 0.0021,
      "step": 1772
    },
    {
      "epoch": 23.99,
      "learning_rate": 0.0001,
      "loss": 0.0033,
      "step": 1776
    },
    {
      "epoch": 24.05,
      "learning_rate": 0.0001,
      "loss": 0.0034,
      "step": 1780
    },
    {
      "epoch": 24.11,
      "learning_rate": 0.0001,
      "loss": 0.0019,
      "step": 1784
    },
    {
      "epoch": 24.16,
      "learning_rate": 0.0001,
      "loss": 0.0029,
      "step": 1788
    },
    {
      "epoch": 24.21,
      "learning_rate": 0.0001,
      "loss": 0.0022,
      "step": 1792
    },
    {
      "epoch": 24.21,
      "eval_exec": 0.7243713733075435,
      "eval_loss": 0.27234283089637756,
      "eval_runtime": 296.6131,
      "eval_samples_per_second": 3.486,
      "step": 1792
    },
    {
      "epoch": 24.27,
      "learning_rate": 0.0001,
      "loss": 0.0018,
      "step": 1796
    },
    {
      "epoch": 24.32,
      "learning_rate": 0.0001,
      "loss": 0.0022,
      "step": 1800
    },
    {
      "epoch": 24.37,
      "learning_rate": 0.0001,
      "loss": 0.0018,
      "step": 1804
    },
    {
      "epoch": 24.43,
      "learning_rate": 0.0001,
      "loss": 0.0024,
      "step": 1808
    },
    {
      "epoch": 24.48,
      "learning_rate": 0.0001,
      "loss": 0.002,
      "step": 1812
    },
    {
      "epoch": 24.53,
      "learning_rate": 0.0001,
      "loss": 0.0025,
      "step": 1816
    },
    {
      "epoch": 24.59,
      "learning_rate": 0.0001,
      "loss": 0.0024,
      "step": 1820
    },
    {
      "epoch": 24.64,
      "learning_rate": 0.0001,
      "loss": 0.0035,
      "step": 1824
    },
    {
      "epoch": 24.7,
      "learning_rate": 0.0001,
      "loss": 0.0023,
      "step": 1828
    },
    {
      "epoch": 24.75,
      "learning_rate": 0.0001,
      "loss": 0.0016,
      "step": 1832
    },
    {
      "epoch": 24.8,
      "learning_rate": 0.0001,
      "loss": 0.0033,
      "step": 1836
    },
    {
      "epoch": 24.86,
      "learning_rate": 0.0001,
      "loss": 0.0035,
      "step": 1840
    },
    {
      "epoch": 24.91,
      "learning_rate": 0.0001,
      "loss": 0.0027,
      "step": 1844
    },
    {
      "epoch": 24.96,
      "learning_rate": 0.0001,
      "loss": 0.0033,
      "step": 1848
    },
    {
      "epoch": 25.03,
      "learning_rate": 0.0001,
      "loss": 0.0029,
      "step": 1852
    },
    {
      "epoch": 25.08,
      "learning_rate": 0.0001,
      "loss": 0.0024,
      "step": 1856
    },
    {
      "epoch": 25.08,
      "eval_exec": 0.718568665377176,
      "eval_loss": 0.2882576584815979,
      "eval_runtime": 293.4007,
      "eval_samples_per_second": 3.524,
      "step": 1856
    },
    {
      "epoch": 25.13,
      "learning_rate": 0.0001,
      "loss": 0.0026,
      "step": 1860
    },
    {
      "epoch": 25.19,
      "learning_rate": 0.0001,
      "loss": 0.0027,
      "step": 1864
    },
    {
      "epoch": 25.24,
      "learning_rate": 0.0001,
      "loss": 0.0033,
      "step": 1868
    },
    {
      "epoch": 25.29,
      "learning_rate": 0.0001,
      "loss": 0.0037,
      "step": 1872
    },
    {
      "epoch": 25.35,
      "learning_rate": 0.0001,
      "loss": 0.0022,
      "step": 1876
    },
    {
      "epoch": 25.4,
      "learning_rate": 0.0001,
      "loss": 0.0028,
      "step": 1880
    },
    {
      "epoch": 25.45,
      "learning_rate": 0.0001,
      "loss": 0.0037,
      "step": 1884
    },
    {
      "epoch": 25.51,
      "learning_rate": 0.0001,
      "loss": 0.004,
      "step": 1888
    },
    {
      "epoch": 25.56,
      "learning_rate": 0.0001,
      "loss": 0.0031,
      "step": 1892
    },
    {
      "epoch": 25.62,
      "learning_rate": 0.0001,
      "loss": 0.0025,
      "step": 1896
    },
    {
      "epoch": 25.67,
      "learning_rate": 0.0001,
      "loss": 0.0044,
      "step": 1900
    },
    {
      "epoch": 25.72,
      "learning_rate": 0.0001,
      "loss": 0.0033,
      "step": 1904
    },
    {
      "epoch": 25.78,
      "learning_rate": 0.0001,
      "loss": 0.0035,
      "step": 1908
    },
    {
      "epoch": 25.83,
      "learning_rate": 0.0001,
      "loss": 0.002,
      "step": 1912
    },
    {
      "epoch": 25.88,
      "learning_rate": 0.0001,
      "loss": 0.0032,
      "step": 1916
    },
    {
      "epoch": 25.94,
      "learning_rate": 0.0001,
      "loss": 0.002,
      "step": 1920
    },
    {
      "epoch": 25.94,
      "eval_exec": 0.6963249516441006,
      "eval_loss": 0.2879628837108612,
      "eval_runtime": 294.5537,
      "eval_samples_per_second": 3.51,
      "step": 1920
    },
    {
      "epoch": 25.99,
      "learning_rate": 0.0001,
      "loss": 0.0028,
      "step": 1924
    },
    {
      "epoch": 26.05,
      "learning_rate": 0.0001,
      "loss": 0.0032,
      "step": 1928
    },
    {
      "epoch": 26.11,
      "learning_rate": 0.0001,
      "loss": 0.0015,
      "step": 1932
    },
    {
      "epoch": 26.16,
      "learning_rate": 0.0001,
      "loss": 0.0022,
      "step": 1936
    },
    {
      "epoch": 26.21,
      "learning_rate": 0.0001,
      "loss": 0.0022,
      "step": 1940
    },
    {
      "epoch": 26.27,
      "learning_rate": 0.0001,
      "loss": 0.0025,
      "step": 1944
    },
    {
      "epoch": 26.32,
      "learning_rate": 0.0001,
      "loss": 0.0056,
      "step": 1948
    },
    {
      "epoch": 26.37,
      "learning_rate": 0.0001,
      "loss": 0.0033,
      "step": 1952
    },
    {
      "epoch": 26.43,
      "learning_rate": 0.0001,
      "loss": 0.0022,
      "step": 1956
    },
    {
      "epoch": 26.48,
      "learning_rate": 0.0001,
      "loss": 0.0023,
      "step": 1960
    },
    {
      "epoch": 26.53,
      "learning_rate": 0.0001,
      "loss": 0.003,
      "step": 1964
    },
    {
      "epoch": 26.59,
      "learning_rate": 0.0001,
      "loss": 0.0022,
      "step": 1968
    },
    {
      "epoch": 26.64,
      "learning_rate": 0.0001,
      "loss": 0.0026,
      "step": 1972
    },
    {
      "epoch": 26.7,
      "learning_rate": 0.0001,
      "loss": 0.0024,
      "step": 1976
    },
    {
      "epoch": 26.75,
      "learning_rate": 0.0001,
      "loss": 0.0033,
      "step": 1980
    },
    {
      "epoch": 26.8,
      "learning_rate": 0.0001,
      "loss": 0.0019,
      "step": 1984
    },
    {
      "epoch": 26.8,
      "eval_exec": 0.7272727272727273,
      "eval_loss": 0.29901865124702454,
      "eval_runtime": 296.6574,
      "eval_samples_per_second": 3.486,
      "step": 1984
    },
    {
      "epoch": 26.86,
      "learning_rate": 0.0001,
      "loss": 0.0029,
      "step": 1988
    },
    {
      "epoch": 26.91,
      "learning_rate": 0.0001,
      "loss": 0.002,
      "step": 1992
    },
    {
      "epoch": 26.96,
      "learning_rate": 0.0001,
      "loss": 0.0048,
      "step": 1996
    },
    {
      "epoch": 27.03,
      "learning_rate": 0.0001,
      "loss": 0.0031,
      "step": 2000
    },
    {
      "epoch": 27.08,
      "learning_rate": 0.0001,
      "loss": 0.0019,
      "step": 2004
    },
    {
      "epoch": 27.13,
      "learning_rate": 0.0001,
      "loss": 0.0012,
      "step": 2008
    },
    {
      "epoch": 27.19,
      "learning_rate": 0.0001,
      "loss": 0.0015,
      "step": 2012
    },
    {
      "epoch": 27.24,
      "learning_rate": 0.0001,
      "loss": 0.0012,
      "step": 2016
    },
    {
      "epoch": 27.29,
      "learning_rate": 0.0001,
      "loss": 0.0028,
      "step": 2020
    },
    {
      "epoch": 27.35,
      "learning_rate": 0.0001,
      "loss": 0.0021,
      "step": 2024
    },
    {
      "epoch": 27.4,
      "learning_rate": 0.0001,
      "loss": 0.0019,
      "step": 2028
    },
    {
      "epoch": 27.45,
      "learning_rate": 0.0001,
      "loss": 0.0014,
      "step": 2032
    },
    {
      "epoch": 27.51,
      "learning_rate": 0.0001,
      "loss": 0.0017,
      "step": 2036
    },
    {
      "epoch": 27.56,
      "learning_rate": 0.0001,
      "loss": 0.0016,
      "step": 2040
    },
    {
      "epoch": 27.62,
      "learning_rate": 0.0001,
      "loss": 0.0009,
      "step": 2044
    },
    {
      "epoch": 27.67,
      "learning_rate": 0.0001,
      "loss": 0.0021,
      "step": 2048
    },
    {
      "epoch": 27.67,
      "eval_exec": 0.7195357833655706,
      "eval_loss": 0.3057432770729065,
      "eval_runtime": 300.1856,
      "eval_samples_per_second": 3.445,
      "step": 2048
    },
    {
      "epoch": 27.72,
      "learning_rate": 0.0001,
      "loss": 0.0027,
      "step": 2052
    },
    {
      "epoch": 27.78,
      "learning_rate": 0.0001,
      "loss": 0.0023,
      "step": 2056
    },
    {
      "epoch": 27.83,
      "learning_rate": 0.0001,
      "loss": 0.0018,
      "step": 2060
    },
    {
      "epoch": 27.88,
      "learning_rate": 0.0001,
      "loss": 0.0022,
      "step": 2064
    },
    {
      "epoch": 27.94,
      "learning_rate": 0.0001,
      "loss": 0.0019,
      "step": 2068
    },
    {
      "epoch": 27.99,
      "learning_rate": 0.0001,
      "loss": 0.002,
      "step": 2072
    },
    {
      "epoch": 28.05,
      "learning_rate": 0.0001,
      "loss": 0.0032,
      "step": 2076
    },
    {
      "epoch": 28.11,
      "learning_rate": 0.0001,
      "loss": 0.0031,
      "step": 2080
    },
    {
      "epoch": 28.16,
      "learning_rate": 0.0001,
      "loss": 0.0018,
      "step": 2084
    },
    {
      "epoch": 28.21,
      "learning_rate": 0.0001,
      "loss": 0.0017,
      "step": 2088
    },
    {
      "epoch": 28.27,
      "learning_rate": 0.0001,
      "loss": 0.0015,
      "step": 2092
    },
    {
      "epoch": 28.32,
      "learning_rate": 0.0001,
      "loss": 0.0029,
      "step": 2096
    },
    {
      "epoch": 28.37,
      "learning_rate": 0.0001,
      "loss": 0.0023,
      "step": 2100
    },
    {
      "epoch": 28.43,
      "learning_rate": 0.0001,
      "loss": 0.0013,
      "step": 2104
    },
    {
      "epoch": 28.48,
      "learning_rate": 0.0001,
      "loss": 0.002,
      "step": 2108
    },
    {
      "epoch": 28.53,
      "learning_rate": 0.0001,
      "loss": 0.0021,
      "step": 2112
    },
    {
      "epoch": 28.53,
      "eval_exec": 0.7282398452611218,
      "eval_loss": 0.28940409421920776,
      "eval_runtime": 294.4149,
      "eval_samples_per_second": 3.512,
      "step": 2112
    },
    {
      "epoch": 28.59,
      "learning_rate": 0.0001,
      "loss": 0.0016,
      "step": 2116
    },
    {
      "epoch": 28.64,
      "learning_rate": 0.0001,
      "loss": 0.002,
      "step": 2120
    },
    {
      "epoch": 28.7,
      "learning_rate": 0.0001,
      "loss": 0.0023,
      "step": 2124
    },
    {
      "epoch": 28.75,
      "learning_rate": 0.0001,
      "loss": 0.0031,
      "step": 2128
    },
    {
      "epoch": 28.8,
      "learning_rate": 0.0001,
      "loss": 0.0024,
      "step": 2132
    },
    {
      "epoch": 28.86,
      "learning_rate": 0.0001,
      "loss": 0.0034,
      "step": 2136
    },
    {
      "epoch": 28.91,
      "learning_rate": 0.0001,
      "loss": 0.0021,
      "step": 2140
    },
    {
      "epoch": 28.96,
      "learning_rate": 0.0001,
      "loss": 0.0031,
      "step": 2144
    },
    {
      "epoch": 29.03,
      "learning_rate": 0.0001,
      "loss": 0.0019,
      "step": 2148
    },
    {
      "epoch": 29.08,
      "learning_rate": 0.0001,
      "loss": 0.0028,
      "step": 2152
    },
    {
      "epoch": 29.13,
      "learning_rate": 0.0001,
      "loss": 0.0015,
      "step": 2156
    },
    {
      "epoch": 29.19,
      "learning_rate": 0.0001,
      "loss": 0.0019,
      "step": 2160
    },
    {
      "epoch": 29.24,
      "learning_rate": 0.0001,
      "loss": 0.001,
      "step": 2164
    },
    {
      "epoch": 29.29,
      "learning_rate": 0.0001,
      "loss": 0.0025,
      "step": 2168
    },
    {
      "epoch": 29.35,
      "learning_rate": 0.0001,
      "loss": 0.003,
      "step": 2172
    },
    {
      "epoch": 29.4,
      "learning_rate": 0.0001,
      "loss": 0.0014,
      "step": 2176
    },
    {
      "epoch": 29.4,
      "eval_exec": 0.7166344294003868,
      "eval_loss": 0.308432400226593,
      "eval_runtime": 289.3433,
      "eval_samples_per_second": 3.574,
      "step": 2176
    },
    {
      "epoch": 29.45,
      "learning_rate": 0.0001,
      "loss": 0.0012,
      "step": 2180
    },
    {
      "epoch": 29.51,
      "learning_rate": 0.0001,
      "loss": 0.0026,
      "step": 2184
    },
    {
      "epoch": 29.56,
      "learning_rate": 0.0001,
      "loss": 0.0044,
      "step": 2188
    },
    {
      "epoch": 29.62,
      "learning_rate": 0.0001,
      "loss": 0.0019,
      "step": 2192
    },
    {
      "epoch": 29.67,
      "learning_rate": 0.0001,
      "loss": 0.0013,
      "step": 2196
    },
    {
      "epoch": 29.72,
      "learning_rate": 0.0001,
      "loss": 0.0021,
      "step": 2200
    },
    {
      "epoch": 29.78,
      "learning_rate": 0.0001,
      "loss": 0.002,
      "step": 2204
    },
    {
      "epoch": 29.83,
      "learning_rate": 0.0001,
      "loss": 0.003,
      "step": 2208
    },
    {
      "epoch": 29.88,
      "learning_rate": 0.0001,
      "loss": 0.0019,
      "step": 2212
    },
    {
      "epoch": 29.94,
      "learning_rate": 0.0001,
      "loss": 0.0012,
      "step": 2216
    },
    {
      "epoch": 29.99,
      "learning_rate": 0.0001,
      "loss": 0.0025,
      "step": 2220
    },
    {
      "epoch": 30.05,
      "learning_rate": 0.0001,
      "loss": 0.0024,
      "step": 2224
    },
    {
      "epoch": 30.11,
      "learning_rate": 0.0001,
      "loss": 0.0047,
      "step": 2228
    },
    {
      "epoch": 30.16,
      "learning_rate": 0.0001,
      "loss": 0.0027,
      "step": 2232
    },
    {
      "epoch": 30.21,
      "learning_rate": 0.0001,
      "loss": 0.0008,
      "step": 2236
    },
    {
      "epoch": 30.27,
      "learning_rate": 0.0001,
      "loss": 0.0035,
      "step": 2240
    },
    {
      "epoch": 30.27,
      "eval_exec": 0.7176015473887815,
      "eval_loss": 0.28124359250068665,
      "eval_runtime": 292.5108,
      "eval_samples_per_second": 3.535,
      "step": 2240
    },
    {
      "epoch": 30.32,
      "learning_rate": 0.0001,
      "loss": 0.0022,
      "step": 2244
    },
    {
      "epoch": 30.37,
      "learning_rate": 0.0001,
      "loss": 0.0023,
      "step": 2248
    },
    {
      "epoch": 30.43,
      "learning_rate": 0.0001,
      "loss": 0.0034,
      "step": 2252
    },
    {
      "epoch": 30.48,
      "learning_rate": 0.0001,
      "loss": 0.0024,
      "step": 2256
    },
    {
      "epoch": 30.53,
      "learning_rate": 0.0001,
      "loss": 0.0038,
      "step": 2260
    },
    {
      "epoch": 30.59,
      "learning_rate": 0.0001,
      "loss": 0.0024,
      "step": 2264
    },
    {
      "epoch": 30.64,
      "learning_rate": 0.0001,
      "loss": 0.004,
      "step": 2268
    },
    {
      "epoch": 30.7,
      "learning_rate": 0.0001,
      "loss": 0.0012,
      "step": 2272
    },
    {
      "epoch": 30.75,
      "learning_rate": 0.0001,
      "loss": 0.0023,
      "step": 2276
    },
    {
      "epoch": 30.8,
      "learning_rate": 0.0001,
      "loss": 0.0021,
      "step": 2280
    },
    {
      "epoch": 30.86,
      "learning_rate": 0.0001,
      "loss": 0.0017,
      "step": 2284
    },
    {
      "epoch": 30.91,
      "learning_rate": 0.0001,
      "loss": 0.001,
      "step": 2288
    },
    {
      "epoch": 30.96,
      "learning_rate": 0.0001,
      "loss": 0.0011,
      "step": 2292
    },
    {
      "epoch": 31.03,
      "learning_rate": 0.0001,
      "loss": 0.0027,
      "step": 2296
    },
    {
      "epoch": 31.08,
      "learning_rate": 0.0001,
      "loss": 0.0013,
      "step": 2300
    },
    {
      "epoch": 31.13,
      "learning_rate": 0.0001,
      "loss": 0.0011,
      "step": 2304
    },
    {
      "epoch": 31.13,
      "eval_exec": 0.7214700193423598,
      "eval_loss": 0.30306151509284973,
      "eval_runtime": 285.8576,
      "eval_samples_per_second": 3.617,
      "step": 2304
    },
    {
      "epoch": 31.19,
      "learning_rate": 0.0001,
      "loss": 0.0007,
      "step": 2308
    },
    {
      "epoch": 31.24,
      "learning_rate": 0.0001,
      "loss": 0.0014,
      "step": 2312
    },
    {
      "epoch": 31.29,
      "learning_rate": 0.0001,
      "loss": 0.003,
      "step": 2316
    },
    {
      "epoch": 31.35,
      "learning_rate": 0.0001,
      "loss": 0.003,
      "step": 2320
    },
    {
      "epoch": 31.4,
      "learning_rate": 0.0001,
      "loss": 0.002,
      "step": 2324
    },
    {
      "epoch": 31.45,
      "learning_rate": 0.0001,
      "loss": 0.0013,
      "step": 2328
    },
    {
      "epoch": 31.51,
      "learning_rate": 0.0001,
      "loss": 0.0022,
      "step": 2332
    },
    {
      "epoch": 31.56,
      "learning_rate": 0.0001,
      "loss": 0.0022,
      "step": 2336
    },
    {
      "epoch": 31.62,
      "learning_rate": 0.0001,
      "loss": 0.0023,
      "step": 2340
    },
    {
      "epoch": 31.67,
      "learning_rate": 0.0001,
      "loss": 0.0028,
      "step": 2344
    },
    {
      "epoch": 31.72,
      "learning_rate": 0.0001,
      "loss": 0.0016,
      "step": 2348
    },
    {
      "epoch": 31.78,
      "learning_rate": 0.0001,
      "loss": 0.0019,
      "step": 2352
    },
    {
      "epoch": 31.83,
      "learning_rate": 0.0001,
      "loss": 0.0007,
      "step": 2356
    },
    {
      "epoch": 31.88,
      "learning_rate": 0.0001,
      "loss": 0.0009,
      "step": 2360
    },
    {
      "epoch": 31.94,
      "learning_rate": 0.0001,
      "loss": 0.0014,
      "step": 2364
    },
    {
      "epoch": 31.99,
      "learning_rate": 0.0001,
      "loss": 0.0025,
      "step": 2368
    },
    {
      "epoch": 31.99,
      "eval_exec": 0.753384912959381,
      "eval_loss": 0.2954598069190979,
      "eval_runtime": 301.2457,
      "eval_samples_per_second": 3.432,
      "step": 2368
    },
    {
      "epoch": 32.05,
      "learning_rate": 0.0001,
      "loss": 0.0016,
      "step": 2372
    },
    {
      "epoch": 32.11,
      "learning_rate": 0.0001,
      "loss": 0.0021,
      "step": 2376
    },
    {
      "epoch": 32.16,
      "learning_rate": 0.0001,
      "loss": 0.0025,
      "step": 2380
    },
    {
      "epoch": 32.21,
      "learning_rate": 0.0001,
      "loss": 0.0045,
      "step": 2384
    },
    {
      "epoch": 32.27,
      "learning_rate": 0.0001,
      "loss": 0.0028,
      "step": 2388
    },
    {
      "epoch": 32.32,
      "learning_rate": 0.0001,
      "loss": 0.0023,
      "step": 2392
    },
    {
      "epoch": 32.37,
      "learning_rate": 0.0001,
      "loss": 0.0014,
      "step": 2396
    },
    {
      "epoch": 32.43,
      "learning_rate": 0.0001,
      "loss": 0.0014,
      "step": 2400
    },
    {
      "epoch": 32.48,
      "learning_rate": 0.0001,
      "loss": 0.0019,
      "step": 2404
    },
    {
      "epoch": 32.53,
      "learning_rate": 0.0001,
      "loss": 0.0014,
      "step": 2408
    },
    {
      "epoch": 32.59,
      "learning_rate": 0.0001,
      "loss": 0.0008,
      "step": 2412
    },
    {
      "epoch": 32.64,
      "learning_rate": 0.0001,
      "loss": 0.0007,
      "step": 2416
    },
    {
      "epoch": 32.7,
      "learning_rate": 0.0001,
      "loss": 0.004,
      "step": 2420
    },
    {
      "epoch": 32.75,
      "learning_rate": 0.0001,
      "loss": 0.003,
      "step": 2424
    },
    {
      "epoch": 32.8,
      "learning_rate": 0.0001,
      "loss": 0.002,
      "step": 2428
    },
    {
      "epoch": 32.86,
      "learning_rate": 0.0001,
      "loss": 0.0008,
      "step": 2432
    },
    {
      "epoch": 32.86,
      "eval_exec": 0.730174081237911,
      "eval_loss": 0.2934824824333191,
      "eval_runtime": 296.4371,
      "eval_samples_per_second": 3.488,
      "step": 2432
    },
    {
      "epoch": 32.91,
      "learning_rate": 0.0001,
      "loss": 0.0009,
      "step": 2436
    },
    {
      "epoch": 32.96,
      "learning_rate": 0.0001,
      "loss": 0.0023,
      "step": 2440
    },
    {
      "epoch": 33.03,
      "learning_rate": 0.0001,
      "loss": 0.0024,
      "step": 2444
    },
    {
      "epoch": 33.08,
      "learning_rate": 0.0001,
      "loss": 0.0017,
      "step": 2448
    },
    {
      "epoch": 33.13,
      "learning_rate": 0.0001,
      "loss": 0.0032,
      "step": 2452
    },
    {
      "epoch": 33.19,
      "learning_rate": 0.0001,
      "loss": 0.0014,
      "step": 2456
    },
    {
      "epoch": 33.24,
      "learning_rate": 0.0001,
      "loss": 0.0016,
      "step": 2460
    },
    {
      "epoch": 33.29,
      "learning_rate": 0.0001,
      "loss": 0.0016,
      "step": 2464
    },
    {
      "epoch": 33.35,
      "learning_rate": 0.0001,
      "loss": 0.001,
      "step": 2468
    },
    {
      "epoch": 33.4,
      "learning_rate": 0.0001,
      "loss": 0.0011,
      "step": 2472
    },
    {
      "epoch": 33.45,
      "learning_rate": 0.0001,
      "loss": 0.0008,
      "step": 2476
    },
    {
      "epoch": 33.51,
      "learning_rate": 0.0001,
      "loss": 0.0025,
      "step": 2480
    },
    {
      "epoch": 33.56,
      "learning_rate": 0.0001,
      "loss": 0.0027,
      "step": 2484
    },
    {
      "epoch": 33.62,
      "learning_rate": 0.0001,
      "loss": 0.0017,
      "step": 2488
    },
    {
      "epoch": 33.67,
      "learning_rate": 0.0001,
      "loss": 0.0013,
      "step": 2492
    },
    {
      "epoch": 33.72,
      "learning_rate": 0.0001,
      "loss": 0.0024,
      "step": 2496
    },
    {
      "epoch": 33.72,
      "eval_exec": 0.695357833655706,
      "eval_loss": 0.28229427337646484,
      "eval_runtime": 288.5686,
      "eval_samples_per_second": 3.583,
      "step": 2496
    },
    {
      "epoch": 33.78,
      "learning_rate": 0.0001,
      "loss": 0.0041,
      "step": 2500
    },
    {
      "epoch": 33.83,
      "learning_rate": 0.0001,
      "loss": 0.0014,
      "step": 2504
    },
    {
      "epoch": 33.88,
      "learning_rate": 0.0001,
      "loss": 0.0017,
      "step": 2508
    },
    {
      "epoch": 33.94,
      "learning_rate": 0.0001,
      "loss": 0.0011,
      "step": 2512
    },
    {
      "epoch": 33.99,
      "learning_rate": 0.0001,
      "loss": 0.0029,
      "step": 2516
    },
    {
      "epoch": 34.05,
      "learning_rate": 0.0001,
      "loss": 0.002,
      "step": 2520
    },
    {
      "epoch": 34.11,
      "learning_rate": 0.0001,
      "loss": 0.0015,
      "step": 2524
    },
    {
      "epoch": 34.16,
      "learning_rate": 0.0001,
      "loss": 0.0015,
      "step": 2528
    },
    {
      "epoch": 34.21,
      "learning_rate": 0.0001,
      "loss": 0.0021,
      "step": 2532
    },
    {
      "epoch": 34.27,
      "learning_rate": 0.0001,
      "loss": 0.0029,
      "step": 2536
    },
    {
      "epoch": 34.32,
      "learning_rate": 0.0001,
      "loss": 0.0017,
      "step": 2540
    },
    {
      "epoch": 34.37,
      "learning_rate": 0.0001,
      "loss": 0.0015,
      "step": 2544
    },
    {
      "epoch": 34.43,
      "learning_rate": 0.0001,
      "loss": 0.0025,
      "step": 2548
    },
    {
      "epoch": 34.48,
      "learning_rate": 0.0001,
      "loss": 0.0015,
      "step": 2552
    },
    {
      "epoch": 34.53,
      "learning_rate": 0.0001,
      "loss": 0.002,
      "step": 2556
    },
    {
      "epoch": 34.59,
      "learning_rate": 0.0001,
      "loss": 0.0019,
      "step": 2560
    },
    {
      "epoch": 34.59,
      "eval_exec": 0.718568665377176,
      "eval_loss": 0.2955181896686554,
      "eval_runtime": 294.421,
      "eval_samples_per_second": 3.512,
      "step": 2560
    },
    {
      "epoch": 34.64,
      "learning_rate": 0.0001,
      "loss": 0.0017,
      "step": 2564
    },
    {
      "epoch": 34.7,
      "learning_rate": 0.0001,
      "loss": 0.0008,
      "step": 2568
    },
    {
      "epoch": 34.75,
      "learning_rate": 0.0001,
      "loss": 0.003,
      "step": 2572
    },
    {
      "epoch": 34.8,
      "learning_rate": 0.0001,
      "loss": 0.0018,
      "step": 2576
    },
    {
      "epoch": 34.86,
      "learning_rate": 0.0001,
      "loss": 0.0022,
      "step": 2580
    },
    {
      "epoch": 34.91,
      "learning_rate": 0.0001,
      "loss": 0.0018,
      "step": 2584
    },
    {
      "epoch": 34.96,
      "learning_rate": 0.0001,
      "loss": 0.001,
      "step": 2588
    },
    {
      "epoch": 35.03,
      "learning_rate": 0.0001,
      "loss": 0.0016,
      "step": 2592
    },
    {
      "epoch": 35.08,
      "learning_rate": 0.0001,
      "loss": 0.0026,
      "step": 2596
    },
    {
      "epoch": 35.13,
      "learning_rate": 0.0001,
      "loss": 0.001,
      "step": 2600
    },
    {
      "epoch": 35.19,
      "learning_rate": 0.0001,
      "loss": 0.0021,
      "step": 2604
    },
    {
      "epoch": 35.24,
      "learning_rate": 0.0001,
      "loss": 0.0018,
      "step": 2608
    },
    {
      "epoch": 35.29,
      "learning_rate": 0.0001,
      "loss": 0.0014,
      "step": 2612
    },
    {
      "epoch": 35.35,
      "learning_rate": 0.0001,
      "loss": 0.0018,
      "step": 2616
    },
    {
      "epoch": 35.4,
      "learning_rate": 0.0001,
      "loss": 0.0014,
      "step": 2620
    },
    {
      "epoch": 35.45,
      "learning_rate": 0.0001,
      "loss": 0.0015,
      "step": 2624
    },
    {
      "epoch": 35.45,
      "eval_exec": 0.7243713733075435,
      "eval_loss": 0.3102652430534363,
      "eval_runtime": 296.911,
      "eval_samples_per_second": 3.483,
      "step": 2624
    },
    {
      "epoch": 35.51,
      "learning_rate": 0.0001,
      "loss": 0.0034,
      "step": 2628
    },
    {
      "epoch": 35.56,
      "learning_rate": 0.0001,
      "loss": 0.0023,
      "step": 2632
    },
    {
      "epoch": 35.62,
      "learning_rate": 0.0001,
      "loss": 0.0012,
      "step": 2636
    },
    {
      "epoch": 35.67,
      "learning_rate": 0.0001,
      "loss": 0.0019,
      "step": 2640
    },
    {
      "epoch": 35.72,
      "learning_rate": 0.0001,
      "loss": 0.0023,
      "step": 2644
    },
    {
      "epoch": 35.78,
      "learning_rate": 0.0001,
      "loss": 0.0014,
      "step": 2648
    },
    {
      "epoch": 35.83,
      "learning_rate": 0.0001,
      "loss": 0.0028,
      "step": 2652
    },
    {
      "epoch": 35.88,
      "learning_rate": 0.0001,
      "loss": 0.0018,
      "step": 2656
    },
    {
      "epoch": 35.94,
      "learning_rate": 0.0001,
      "loss": 0.0028,
      "step": 2660
    },
    {
      "epoch": 35.99,
      "learning_rate": 0.0001,
      "loss": 0.0015,
      "step": 2664
    },
    {
      "epoch": 36.05,
      "learning_rate": 0.0001,
      "loss": 0.0027,
      "step": 2668
    },
    {
      "epoch": 36.11,
      "learning_rate": 0.0001,
      "loss": 0.0017,
      "step": 2672
    },
    {
      "epoch": 36.16,
      "learning_rate": 0.0001,
      "loss": 0.0014,
      "step": 2676
    },
    {
      "epoch": 36.21,
      "learning_rate": 0.0001,
      "loss": 0.0009,
      "step": 2680
    },
    {
      "epoch": 36.27,
      "learning_rate": 0.0001,
      "loss": 0.0017,
      "step": 2684
    },
    {
      "epoch": 36.32,
      "learning_rate": 0.0001,
      "loss": 0.003,
      "step": 2688
    },
    {
      "epoch": 36.32,
      "eval_exec": 0.7030947775628626,
      "eval_loss": 0.2786959409713745,
      "eval_runtime": 286.0516,
      "eval_samples_per_second": 3.615,
      "step": 2688
    },
    {
      "epoch": 36.37,
      "learning_rate": 0.0001,
      "loss": 0.0022,
      "step": 2692
    },
    {
      "epoch": 36.43,
      "learning_rate": 0.0001,
      "loss": 0.0018,
      "step": 2696
    },
    {
      "epoch": 36.48,
      "learning_rate": 0.0001,
      "loss": 0.001,
      "step": 2700
    },
    {
      "epoch": 36.53,
      "learning_rate": 0.0001,
      "loss": 0.0026,
      "step": 2704
    },
    {
      "epoch": 36.59,
      "learning_rate": 0.0001,
      "loss": 0.002,
      "step": 2708
    },
    {
      "epoch": 36.64,
      "learning_rate": 0.0001,
      "loss": 0.0022,
      "step": 2712
    },
    {
      "epoch": 36.7,
      "learning_rate": 0.0001,
      "loss": 0.0009,
      "step": 2716
    },
    {
      "epoch": 36.75,
      "learning_rate": 0.0001,
      "loss": 0.0021,
      "step": 2720
    },
    {
      "epoch": 36.8,
      "learning_rate": 0.0001,
      "loss": 0.002,
      "step": 2724
    },
    {
      "epoch": 36.86,
      "learning_rate": 0.0001,
      "loss": 0.0012,
      "step": 2728
    },
    {
      "epoch": 36.91,
      "learning_rate": 0.0001,
      "loss": 0.0018,
      "step": 2732
    },
    {
      "epoch": 36.96,
      "learning_rate": 0.0001,
      "loss": 0.0013,
      "step": 2736
    },
    {
      "epoch": 37.03,
      "learning_rate": 0.0001,
      "loss": 0.0016,
      "step": 2740
    },
    {
      "epoch": 37.08,
      "learning_rate": 0.0001,
      "loss": 0.0012,
      "step": 2744
    },
    {
      "epoch": 37.13,
      "learning_rate": 0.0001,
      "loss": 0.0013,
      "step": 2748
    },
    {
      "epoch": 37.19,
      "learning_rate": 0.0001,
      "loss": 0.001,
      "step": 2752
    },
    {
      "epoch": 37.19,
      "eval_exec": 0.7195357833655706,
      "eval_loss": 0.3104759156703949,
      "eval_runtime": 302.4048,
      "eval_samples_per_second": 3.419,
      "step": 2752
    },
    {
      "epoch": 37.24,
      "learning_rate": 0.0001,
      "loss": 0.0016,
      "step": 2756
    },
    {
      "epoch": 37.29,
      "learning_rate": 0.0001,
      "loss": 0.0015,
      "step": 2760
    },
    {
      "epoch": 37.35,
      "learning_rate": 0.0001,
      "loss": 0.001,
      "step": 2764
    },
    {
      "epoch": 37.4,
      "learning_rate": 0.0001,
      "loss": 0.0018,
      "step": 2768
    },
    {
      "epoch": 37.45,
      "learning_rate": 0.0001,
      "loss": 0.0022,
      "step": 2772
    },
    {
      "epoch": 37.51,
      "learning_rate": 0.0001,
      "loss": 0.001,
      "step": 2776
    },
    {
      "epoch": 37.56,
      "learning_rate": 0.0001,
      "loss": 0.0015,
      "step": 2780
    },
    {
      "epoch": 37.62,
      "learning_rate": 0.0001,
      "loss": 0.0015,
      "step": 2784
    },
    {
      "epoch": 37.67,
      "learning_rate": 0.0001,
      "loss": 0.0013,
      "step": 2788
    },
    {
      "epoch": 37.72,
      "learning_rate": 0.0001,
      "loss": 0.0009,
      "step": 2792
    },
    {
      "epoch": 37.78,
      "learning_rate": 0.0001,
      "loss": 0.0029,
      "step": 2796
    },
    {
      "epoch": 37.83,
      "learning_rate": 0.0001,
      "loss": 0.0016,
      "step": 2800
    },
    {
      "epoch": 37.88,
      "learning_rate": 0.0001,
      "loss": 0.0009,
      "step": 2804
    },
    {
      "epoch": 37.94,
      "learning_rate": 0.0001,
      "loss": 0.0013,
      "step": 2808
    },
    {
      "epoch": 37.99,
      "learning_rate": 0.0001,
      "loss": 0.0021,
      "step": 2812
    },
    {
      "epoch": 38.05,
      "learning_rate": 0.0001,
      "loss": 0.0038,
      "step": 2816
    },
    {
      "epoch": 38.05,
      "eval_exec": 0.7011605415860735,
      "eval_loss": 0.3018852472305298,
      "eval_runtime": 286.3546,
      "eval_samples_per_second": 3.611,
      "step": 2816
    },
    {
      "epoch": 38.11,
      "learning_rate": 0.0001,
      "loss": 0.0018,
      "step": 2820
    },
    {
      "epoch": 38.16,
      "learning_rate": 0.0001,
      "loss": 0.0013,
      "step": 2824
    },
    {
      "epoch": 38.21,
      "learning_rate": 0.0001,
      "loss": 0.0016,
      "step": 2828
    },
    {
      "epoch": 38.27,
      "learning_rate": 0.0001,
      "loss": 0.0017,
      "step": 2832
    },
    {
      "epoch": 38.32,
      "learning_rate": 0.0001,
      "loss": 0.0016,
      "step": 2836
    },
    {
      "epoch": 38.37,
      "learning_rate": 0.0001,
      "loss": 0.0018,
      "step": 2840
    },
    {
      "epoch": 38.43,
      "learning_rate": 0.0001,
      "loss": 0.0006,
      "step": 2844
    },
    {
      "epoch": 38.48,
      "learning_rate": 0.0001,
      "loss": 0.001,
      "step": 2848
    },
    {
      "epoch": 38.53,
      "learning_rate": 0.0001,
      "loss": 0.0017,
      "step": 2852
    },
    {
      "epoch": 38.59,
      "learning_rate": 0.0001,
      "loss": 0.0022,
      "step": 2856
    },
    {
      "epoch": 38.64,
      "learning_rate": 0.0001,
      "loss": 0.0012,
      "step": 2860
    },
    {
      "epoch": 38.7,
      "learning_rate": 0.0001,
      "loss": 0.0024,
      "step": 2864
    },
    {
      "epoch": 38.75,
      "learning_rate": 0.0001,
      "loss": 0.0012,
      "step": 2868
    },
    {
      "epoch": 38.8,
      "learning_rate": 0.0001,
      "loss": 0.0027,
      "step": 2872
    },
    {
      "epoch": 38.86,
      "learning_rate": 0.0001,
      "loss": 0.001,
      "step": 2876
    },
    {
      "epoch": 38.91,
      "learning_rate": 0.0001,
      "loss": 0.0012,
      "step": 2880
    },
    {
      "epoch": 38.91,
      "eval_exec": 0.7098646034816247,
      "eval_loss": 0.295253187417984,
      "eval_runtime": 292.7859,
      "eval_samples_per_second": 3.532,
      "step": 2880
    },
    {
      "epoch": 38.96,
      "learning_rate": 0.0001,
      "loss": 0.0011,
      "step": 2884
    },
    {
      "epoch": 39.03,
      "learning_rate": 0.0001,
      "loss": 0.0012,
      "step": 2888
    },
    {
      "epoch": 39.08,
      "learning_rate": 0.0001,
      "loss": 0.0012,
      "step": 2892
    },
    {
      "epoch": 39.13,
      "learning_rate": 0.0001,
      "loss": 0.0005,
      "step": 2896
    },
    {
      "epoch": 39.19,
      "learning_rate": 0.0001,
      "loss": 0.0018,
      "step": 2900
    },
    {
      "epoch": 39.24,
      "learning_rate": 0.0001,
      "loss": 0.0015,
      "step": 2904
    },
    {
      "epoch": 39.29,
      "learning_rate": 0.0001,
      "loss": 0.0008,
      "step": 2908
    },
    {
      "epoch": 39.35,
      "learning_rate": 0.0001,
      "loss": 0.0007,
      "step": 2912
    },
    {
      "epoch": 39.4,
      "learning_rate": 0.0001,
      "loss": 0.0009,
      "step": 2916
    },
    {
      "epoch": 39.45,
      "learning_rate": 0.0001,
      "loss": 0.0004,
      "step": 2920
    },
    {
      "epoch": 39.51,
      "learning_rate": 0.0001,
      "loss": 0.0012,
      "step": 2924
    },
    {
      "epoch": 39.56,
      "learning_rate": 0.0001,
      "loss": 0.0007,
      "step": 2928
    },
    {
      "epoch": 39.62,
      "learning_rate": 0.0001,
      "loss": 0.0011,
      "step": 2932
    },
    {
      "epoch": 39.67,
      "learning_rate": 0.0001,
      "loss": 0.0012,
      "step": 2936
    },
    {
      "epoch": 39.72,
      "learning_rate": 0.0001,
      "loss": 0.0008,
      "step": 2940
    },
    {
      "epoch": 39.78,
      "learning_rate": 0.0001,
      "loss": 0.0014,
      "step": 2944
    },
    {
      "epoch": 39.78,
      "eval_exec": 0.7156673114119922,
      "eval_loss": 0.3224295973777771,
      "eval_runtime": 296.2398,
      "eval_samples_per_second": 3.49,
      "step": 2944
    },
    {
      "epoch": 39.83,
      "learning_rate": 0.0001,
      "loss": 0.002,
      "step": 2948
    },
    {
      "epoch": 39.88,
      "learning_rate": 0.0001,
      "loss": 0.0011,
      "step": 2952
    },
    {
      "epoch": 39.94,
      "learning_rate": 0.0001,
      "loss": 0.001,
      "step": 2956
    },
    {
      "epoch": 39.99,
      "learning_rate": 0.0001,
      "loss": 0.0019,
      "step": 2960
    },
    {
      "epoch": 40.05,
      "learning_rate": 0.0001,
      "loss": 0.0027,
      "step": 2964
    },
    {
      "epoch": 40.11,
      "learning_rate": 0.0001,
      "loss": 0.0024,
      "step": 2968
    },
    {
      "epoch": 40.16,
      "learning_rate": 0.0001,
      "loss": 0.0027,
      "step": 2972
    },
    {
      "epoch": 40.21,
      "learning_rate": 0.0001,
      "loss": 0.0012,
      "step": 2976
    },
    {
      "epoch": 40.27,
      "learning_rate": 0.0001,
      "loss": 0.0022,
      "step": 2980
    },
    {
      "epoch": 40.32,
      "learning_rate": 0.0001,
      "loss": 0.0019,
      "step": 2984
    },
    {
      "epoch": 40.37,
      "learning_rate": 0.0001,
      "loss": 0.0022,
      "step": 2988
    },
    {
      "epoch": 40.43,
      "learning_rate": 0.0001,
      "loss": 0.0007,
      "step": 2992
    },
    {
      "epoch": 40.48,
      "learning_rate": 0.0001,
      "loss": 0.0017,
      "step": 2996
    },
    {
      "epoch": 40.53,
      "learning_rate": 0.0001,
      "loss": 0.0013,
      "step": 3000
    },
    {
      "epoch": 40.59,
      "learning_rate": 0.0001,
      "loss": 0.0038,
      "step": 3004
    },
    {
      "epoch": 40.64,
      "learning_rate": 0.0001,
      "loss": 0.0011,
      "step": 3008
    },
    {
      "epoch": 40.64,
      "eval_exec": 0.7040618955512572,
      "eval_loss": 0.3050314486026764,
      "eval_runtime": 289.6078,
      "eval_samples_per_second": 3.57,
      "step": 3008
    },
    {
      "epoch": 40.7,
      "learning_rate": 0.0001,
      "loss": 0.0026,
      "step": 3012
    },
    {
      "epoch": 40.75,
      "learning_rate": 0.0001,
      "loss": 0.0021,
      "step": 3016
    },
    {
      "epoch": 40.8,
      "learning_rate": 0.0001,
      "loss": 0.0015,
      "step": 3020
    },
    {
      "epoch": 40.86,
      "learning_rate": 0.0001,
      "loss": 0.0021,
      "step": 3024
    },
    {
      "epoch": 40.91,
      "learning_rate": 0.0001,
      "loss": 0.0012,
      "step": 3028
    },
    {
      "epoch": 40.96,
      "learning_rate": 0.0001,
      "loss": 0.0032,
      "step": 3032
    },
    {
      "epoch": 41.03,
      "learning_rate": 0.0001,
      "loss": 0.0018,
      "step": 3036
    },
    {
      "epoch": 41.08,
      "learning_rate": 0.0001,
      "loss": 0.0014,
      "step": 3040
    },
    {
      "epoch": 41.13,
      "learning_rate": 0.0001,
      "loss": 0.0009,
      "step": 3044
    },
    {
      "epoch": 41.19,
      "learning_rate": 0.0001,
      "loss": 0.0019,
      "step": 3048
    },
    {
      "epoch": 41.24,
      "learning_rate": 0.0001,
      "loss": 0.0029,
      "step": 3052
    },
    {
      "epoch": 41.29,
      "learning_rate": 0.0001,
      "loss": 0.0022,
      "step": 3056
    },
    {
      "epoch": 41.35,
      "learning_rate": 0.0001,
      "loss": 0.0011,
      "step": 3060
    },
    {
      "epoch": 41.4,
      "learning_rate": 0.0001,
      "loss": 0.0018,
      "step": 3064
    },
    {
      "epoch": 41.45,
      "learning_rate": 0.0001,
      "loss": 0.0005,
      "step": 3068
    },
    {
      "epoch": 41.51,
      "learning_rate": 0.0001,
      "loss": 0.0016,
      "step": 3072
    },
    {
      "epoch": 41.51,
      "eval_exec": 0.7088974854932302,
      "eval_loss": 0.30448201298713684,
      "eval_runtime": 294.3306,
      "eval_samples_per_second": 3.513,
      "step": 3072
    },
    {
      "epoch": 41.56,
      "learning_rate": 0.0001,
      "loss": 0.0025,
      "step": 3076
    },
    {
      "epoch": 41.62,
      "learning_rate": 0.0001,
      "loss": 0.0011,
      "step": 3080
    },
    {
      "epoch": 41.67,
      "learning_rate": 0.0001,
      "loss": 0.0005,
      "step": 3084
    },
    {
      "epoch": 41.72,
      "learning_rate": 0.0001,
      "loss": 0.0006,
      "step": 3088
    },
    {
      "epoch": 41.78,
      "learning_rate": 0.0001,
      "loss": 0.0012,
      "step": 3092
    },
    {
      "epoch": 41.83,
      "learning_rate": 0.0001,
      "loss": 0.0019,
      "step": 3096
    },
    {
      "epoch": 41.88,
      "learning_rate": 0.0001,
      "loss": 0.0013,
      "step": 3100
    },
    {
      "epoch": 41.94,
      "learning_rate": 0.0001,
      "loss": 0.0007,
      "step": 3104
    },
    {
      "epoch": 41.99,
      "learning_rate": 0.0001,
      "loss": 0.0033,
      "step": 3108
    },
    {
      "epoch": 42.05,
      "learning_rate": 0.0001,
      "loss": 0.0021,
      "step": 3112
    },
    {
      "epoch": 42.11,
      "learning_rate": 0.0001,
      "loss": 0.001,
      "step": 3116
    },
    {
      "epoch": 42.16,
      "learning_rate": 0.0001,
      "loss": 0.0023,
      "step": 3120
    },
    {
      "epoch": 42.21,
      "learning_rate": 0.0001,
      "loss": 0.001,
      "step": 3124
    },
    {
      "epoch": 42.27,
      "learning_rate": 0.0001,
      "loss": 0.0015,
      "step": 3128
    },
    {
      "epoch": 42.32,
      "learning_rate": 0.0001,
      "loss": 0.0012,
      "step": 3132
    },
    {
      "epoch": 42.37,
      "learning_rate": 0.0001,
      "loss": 0.0007,
      "step": 3136
    },
    {
      "epoch": 42.37,
      "eval_exec": 0.7243713733075435,
      "eval_loss": 0.3050415515899658,
      "eval_runtime": 293.6495,
      "eval_samples_per_second": 3.521,
      "step": 3136
    },
    {
      "epoch": 42.43,
      "learning_rate": 0.0001,
      "loss": 0.0022,
      "step": 3140
    },
    {
      "epoch": 42.48,
      "learning_rate": 0.0001,
      "loss": 0.0012,
      "step": 3144
    },
    {
      "epoch": 42.53,
      "learning_rate": 0.0001,
      "loss": 0.0015,
      "step": 3148
    },
    {
      "epoch": 42.59,
      "learning_rate": 0.0001,
      "loss": 0.0037,
      "step": 3152
    },
    {
      "epoch": 42.64,
      "learning_rate": 0.0001,
      "loss": 0.0024,
      "step": 3156
    },
    {
      "epoch": 42.7,
      "learning_rate": 0.0001,
      "loss": 0.0012,
      "step": 3160
    },
    {
      "epoch": 42.75,
      "learning_rate": 0.0001,
      "loss": 0.0016,
      "step": 3164
    },
    {
      "epoch": 42.8,
      "learning_rate": 0.0001,
      "loss": 0.0031,
      "step": 3168
    },
    {
      "epoch": 42.86,
      "learning_rate": 0.0001,
      "loss": 0.0013,
      "step": 3172
    },
    {
      "epoch": 42.91,
      "learning_rate": 0.0001,
      "loss": 0.0006,
      "step": 3176
    },
    {
      "epoch": 42.96,
      "learning_rate": 0.0001,
      "loss": 0.0015,
      "step": 3180
    },
    {
      "epoch": 43.03,
      "learning_rate": 0.0001,
      "loss": 0.0014,
      "step": 3184
    },
    {
      "epoch": 43.08,
      "learning_rate": 0.0001,
      "loss": 0.0009,
      "step": 3188
    },
    {
      "epoch": 43.13,
      "learning_rate": 0.0001,
      "loss": 0.0008,
      "step": 3192
    },
    {
      "epoch": 43.19,
      "learning_rate": 0.0001,
      "loss": 0.0013,
      "step": 3196
    },
    {
      "epoch": 43.24,
      "learning_rate": 0.0001,
      "loss": 0.0012,
      "step": 3200
    },
    {
      "epoch": 43.24,
      "eval_exec": 0.7040618955512572,
      "eval_loss": 0.3184911012649536,
      "eval_runtime": 290.8614,
      "eval_samples_per_second": 3.555,
      "step": 3200
    },
    {
      "epoch": 43.29,
      "learning_rate": 0.0001,
      "loss": 0.0015,
      "step": 3204
    },
    {
      "epoch": 43.35,
      "learning_rate": 0.0001,
      "loss": 0.0027,
      "step": 3208
    },
    {
      "epoch": 43.4,
      "learning_rate": 0.0001,
      "loss": 0.0025,
      "step": 3212
    },
    {
      "epoch": 43.45,
      "learning_rate": 0.0001,
      "loss": 0.0006,
      "step": 3216
    },
    {
      "epoch": 43.51,
      "learning_rate": 0.0001,
      "loss": 0.0027,
      "step": 3220
    },
    {
      "epoch": 43.56,
      "learning_rate": 0.0001,
      "loss": 0.0016,
      "step": 3224
    },
    {
      "epoch": 43.62,
      "learning_rate": 0.0001,
      "loss": 0.0011,
      "step": 3228
    },
    {
      "epoch": 43.67,
      "learning_rate": 0.0001,
      "loss": 0.0007,
      "step": 3232
    },
    {
      "epoch": 43.72,
      "learning_rate": 0.0001,
      "loss": 0.0028,
      "step": 3236
    },
    {
      "epoch": 43.78,
      "learning_rate": 0.0001,
      "loss": 0.0011,
      "step": 3240
    },
    {
      "epoch": 43.83,
      "learning_rate": 0.0001,
      "loss": 0.0008,
      "step": 3244
    },
    {
      "epoch": 43.88,
      "learning_rate": 0.0001,
      "loss": 0.0014,
      "step": 3248
    },
    {
      "epoch": 43.94,
      "learning_rate": 0.0001,
      "loss": 0.0028,
      "step": 3252
    },
    {
      "epoch": 43.99,
      "learning_rate": 0.0001,
      "loss": 0.0008,
      "step": 3256
    },
    {
      "epoch": 44.05,
      "learning_rate": 0.0001,
      "loss": 0.0026,
      "step": 3260
    },
    {
      "epoch": 44.11,
      "learning_rate": 0.0001,
      "loss": 0.0014,
      "step": 3264
    },
    {
      "epoch": 44.11,
      "eval_exec": 0.7108317214700194,
      "eval_loss": 0.2903190851211548,
      "eval_runtime": 296.1114,
      "eval_samples_per_second": 3.492,
      "step": 3264
    },
    {
      "epoch": 44.16,
      "learning_rate": 0.0001,
      "loss": 0.0008,
      "step": 3268
    },
    {
      "epoch": 44.21,
      "learning_rate": 0.0001,
      "loss": 0.0008,
      "step": 3272
    },
    {
      "epoch": 44.27,
      "learning_rate": 0.0001,
      "loss": 0.0016,
      "step": 3276
    },
    {
      "epoch": 44.32,
      "learning_rate": 0.0001,
      "loss": 0.0017,
      "step": 3280
    },
    {
      "epoch": 44.37,
      "learning_rate": 0.0001,
      "loss": 0.0011,
      "step": 3284
    },
    {
      "epoch": 44.43,
      "learning_rate": 0.0001,
      "loss": 0.0009,
      "step": 3288
    },
    {
      "epoch": 44.48,
      "learning_rate": 0.0001,
      "loss": 0.0006,
      "step": 3292
    },
    {
      "epoch": 44.53,
      "learning_rate": 0.0001,
      "loss": 0.0014,
      "step": 3296
    },
    {
      "epoch": 44.59,
      "learning_rate": 0.0001,
      "loss": 0.0021,
      "step": 3300
    },
    {
      "epoch": 44.64,
      "learning_rate": 0.0001,
      "loss": 0.0008,
      "step": 3304
    },
    {
      "epoch": 44.7,
      "learning_rate": 0.0001,
      "loss": 0.0014,
      "step": 3308
    },
    {
      "epoch": 44.75,
      "learning_rate": 0.0001,
      "loss": 0.0035,
      "step": 3312
    },
    {
      "epoch": 44.8,
      "learning_rate": 0.0001,
      "loss": 0.001,
      "step": 3316
    },
    {
      "epoch": 44.86,
      "learning_rate": 0.0001,
      "loss": 0.0013,
      "step": 3320
    },
    {
      "epoch": 44.91,
      "learning_rate": 0.0001,
      "loss": 0.0007,
      "step": 3324
    },
    {
      "epoch": 44.96,
      "learning_rate": 0.0001,
      "loss": 0.0032,
      "step": 3328
    },
    {
      "epoch": 44.96,
      "eval_exec": 0.7166344294003868,
      "eval_loss": 0.29221683740615845,
      "eval_runtime": 297.9619,
      "eval_samples_per_second": 3.47,
      "step": 3328
    },
    {
      "epoch": 45.03,
      "learning_rate": 0.0001,
      "loss": 0.0013,
      "step": 3332
    },
    {
      "epoch": 45.08,
      "learning_rate": 0.0001,
      "loss": 0.0015,
      "step": 3336
    },
    {
      "epoch": 45.13,
      "learning_rate": 0.0001,
      "loss": 0.0013,
      "step": 3340
    },
    {
      "epoch": 45.19,
      "learning_rate": 0.0001,
      "loss": 0.0012,
      "step": 3344
    },
    {
      "epoch": 45.24,
      "learning_rate": 0.0001,
      "loss": 0.0013,
      "step": 3348
    },
    {
      "epoch": 45.29,
      "learning_rate": 0.0001,
      "loss": 0.0014,
      "step": 3352
    },
    {
      "epoch": 45.35,
      "learning_rate": 0.0001,
      "loss": 0.0006,
      "step": 3356
    },
    {
      "epoch": 45.4,
      "learning_rate": 0.0001,
      "loss": 0.0017,
      "step": 3360
    },
    {
      "epoch": 45.45,
      "learning_rate": 0.0001,
      "loss": 0.0003,
      "step": 3364
    },
    {
      "epoch": 45.51,
      "learning_rate": 0.0001,
      "loss": 0.0008,
      "step": 3368
    },
    {
      "epoch": 45.56,
      "learning_rate": 0.0001,
      "loss": 0.0009,
      "step": 3372
    },
    {
      "epoch": 45.62,
      "learning_rate": 0.0001,
      "loss": 0.0022,
      "step": 3376
    },
    {
      "epoch": 45.67,
      "learning_rate": 0.0001,
      "loss": 0.0009,
      "step": 3380
    },
    {
      "epoch": 45.72,
      "learning_rate": 0.0001,
      "loss": 0.0002,
      "step": 3384
    },
    {
      "epoch": 45.78,
      "learning_rate": 0.0001,
      "loss": 0.0021,
      "step": 3388
    },
    {
      "epoch": 45.83,
      "learning_rate": 0.0001,
      "loss": 0.0023,
      "step": 3392
    },
    {
      "epoch": 45.83,
      "eval_exec": 0.7166344294003868,
      "eval_loss": 0.30226799845695496,
      "eval_runtime": 296.3235,
      "eval_samples_per_second": 3.489,
      "step": 3392
    },
    {
      "epoch": 45.88,
      "learning_rate": 0.0001,
      "loss": 0.0021,
      "step": 3396
    },
    {
      "epoch": 45.94,
      "learning_rate": 0.0001,
      "loss": 0.0023,
      "step": 3400
    },
    {
      "epoch": 45.99,
      "learning_rate": 0.0001,
      "loss": 0.0016,
      "step": 3404
    },
    {
      "epoch": 46.05,
      "learning_rate": 0.0001,
      "loss": 0.0015,
      "step": 3408
    },
    {
      "epoch": 46.11,
      "learning_rate": 0.0001,
      "loss": 0.0007,
      "step": 3412
    },
    {
      "epoch": 46.16,
      "learning_rate": 0.0001,
      "loss": 0.0014,
      "step": 3416
    },
    {
      "epoch": 46.21,
      "learning_rate": 0.0001,
      "loss": 0.0005,
      "step": 3420
    },
    {
      "epoch": 46.27,
      "learning_rate": 0.0001,
      "loss": 0.0014,
      "step": 3424
    },
    {
      "epoch": 46.32,
      "learning_rate": 0.0001,
      "loss": 0.0016,
      "step": 3428
    },
    {
      "epoch": 46.37,
      "learning_rate": 0.0001,
      "loss": 0.0012,
      "step": 3432
    },
    {
      "epoch": 46.43,
      "learning_rate": 0.0001,
      "loss": 0.0007,
      "step": 3436
    },
    {
      "epoch": 46.48,
      "learning_rate": 0.0001,
      "loss": 0.0027,
      "step": 3440
    },
    {
      "epoch": 46.53,
      "learning_rate": 0.0001,
      "loss": 0.0008,
      "step": 3444
    },
    {
      "epoch": 46.59,
      "learning_rate": 0.0001,
      "loss": 0.0012,
      "step": 3448
    },
    {
      "epoch": 46.64,
      "learning_rate": 0.0001,
      "loss": 0.0006,
      "step": 3452
    },
    {
      "epoch": 46.7,
      "learning_rate": 0.0001,
      "loss": 0.0004,
      "step": 3456
    },
    {
      "epoch": 46.7,
      "eval_exec": 0.7437137330754352,
      "eval_loss": 0.30926698446273804,
      "eval_runtime": 297.8667,
      "eval_samples_per_second": 3.471,
      "step": 3456
    },
    {
      "epoch": 46.75,
      "learning_rate": 0.0001,
      "loss": 0.0023,
      "step": 3460
    },
    {
      "epoch": 46.8,
      "learning_rate": 0.0001,
      "loss": 0.0011,
      "step": 3464
    },
    {
      "epoch": 46.86,
      "learning_rate": 0.0001,
      "loss": 0.0011,
      "step": 3468
    },
    {
      "epoch": 46.91,
      "learning_rate": 0.0001,
      "loss": 0.0012,
      "step": 3472
    },
    {
      "epoch": 46.96,
      "learning_rate": 0.0001,
      "loss": 0.0013,
      "step": 3476
    },
    {
      "epoch": 47.03,
      "learning_rate": 0.0001,
      "loss": 0.0009,
      "step": 3480
    },
    {
      "epoch": 47.08,
      "learning_rate": 0.0001,
      "loss": 0.0006,
      "step": 3484
    },
    {
      "epoch": 47.13,
      "learning_rate": 0.0001,
      "loss": 0.0015,
      "step": 3488
    },
    {
      "epoch": 47.19,
      "learning_rate": 0.0001,
      "loss": 0.0003,
      "step": 3492
    },
    {
      "epoch": 47.24,
      "learning_rate": 0.0001,
      "loss": 0.0021,
      "step": 3496
    },
    {
      "epoch": 47.29,
      "learning_rate": 0.0001,
      "loss": 0.0012,
      "step": 3500
    },
    {
      "epoch": 47.35,
      "learning_rate": 0.0001,
      "loss": 0.0006,
      "step": 3504
    },
    {
      "epoch": 47.4,
      "learning_rate": 0.0001,
      "loss": 0.0004,
      "step": 3508
    },
    {
      "epoch": 47.45,
      "learning_rate": 0.0001,
      "loss": 0.0006,
      "step": 3512
    },
    {
      "epoch": 47.51,
      "learning_rate": 0.0001,
      "loss": 0.0025,
      "step": 3516
    },
    {
      "epoch": 47.56,
      "learning_rate": 0.0001,
      "loss": 0.0054,
      "step": 3520
    },
    {
      "epoch": 47.56,
      "eval_exec": 0.7350096711798839,
      "eval_loss": 0.27321064472198486,
      "eval_runtime": 285.4659,
      "eval_samples_per_second": 3.622,
      "step": 3520
    },
    {
      "epoch": 47.62,
      "learning_rate": 0.0001,
      "loss": 0.001,
      "step": 3524
    },
    {
      "epoch": 47.67,
      "learning_rate": 0.0001,
      "loss": 0.0009,
      "step": 3528
    },
    {
      "epoch": 47.72,
      "learning_rate": 0.0001,
      "loss": 0.0013,
      "step": 3532
    },
    {
      "epoch": 47.78,
      "learning_rate": 0.0001,
      "loss": 0.0016,
      "step": 3536
    },
    {
      "epoch": 47.83,
      "learning_rate": 0.0001,
      "loss": 0.0008,
      "step": 3540
    },
    {
      "epoch": 47.88,
      "learning_rate": 0.0001,
      "loss": 0.0006,
      "step": 3544
    },
    {
      "epoch": 47.94,
      "learning_rate": 0.0001,
      "loss": 0.0008,
      "step": 3548
    },
    {
      "epoch": 47.99,
      "learning_rate": 0.0001,
      "loss": 0.0013,
      "step": 3552
    },
    {
      "epoch": 48.05,
      "learning_rate": 0.0001,
      "loss": 0.0014,
      "step": 3556
    },
    {
      "epoch": 48.11,
      "learning_rate": 0.0001,
      "loss": 0.0012,
      "step": 3560
    },
    {
      "epoch": 48.16,
      "learning_rate": 0.0001,
      "loss": 0.0013,
      "step": 3564
    },
    {
      "epoch": 48.21,
      "learning_rate": 0.0001,
      "loss": 0.0003,
      "step": 3568
    },
    {
      "epoch": 48.27,
      "learning_rate": 0.0001,
      "loss": 0.0018,
      "step": 3572
    },
    {
      "epoch": 48.32,
      "learning_rate": 0.0001,
      "loss": 0.0005,
      "step": 3576
    },
    {
      "epoch": 48.37,
      "learning_rate": 0.0001,
      "loss": 0.0029,
      "step": 3580
    },
    {
      "epoch": 48.43,
      "learning_rate": 0.0001,
      "loss": 0.0014,
      "step": 3584
    },
    {
      "epoch": 48.43,
      "eval_exec": 0.746615087040619,
      "eval_loss": 0.2905672490596771,
      "eval_runtime": 293.155,
      "eval_samples_per_second": 3.527,
      "step": 3584
    },
    {
      "epoch": 48.48,
      "learning_rate": 0.0001,
      "loss": 0.0008,
      "step": 3588
    },
    {
      "epoch": 48.53,
      "learning_rate": 0.0001,
      "loss": 0.0012,
      "step": 3592
    },
    {
      "epoch": 48.59,
      "learning_rate": 0.0001,
      "loss": 0.0022,
      "step": 3596
    },
    {
      "epoch": 48.64,
      "learning_rate": 0.0001,
      "loss": 0.0013,
      "step": 3600
    },
    {
      "epoch": 48.7,
      "learning_rate": 0.0001,
      "loss": 0.001,
      "step": 3604
    },
    {
      "epoch": 48.75,
      "learning_rate": 0.0001,
      "loss": 0.0014,
      "step": 3608
    },
    {
      "epoch": 48.8,
      "learning_rate": 0.0001,
      "loss": 0.0011,
      "step": 3612
    },
    {
      "epoch": 48.86,
      "learning_rate": 0.0001,
      "loss": 0.0057,
      "step": 3616
    },
    {
      "epoch": 48.91,
      "learning_rate": 0.0001,
      "loss": 0.0008,
      "step": 3620
    },
    {
      "epoch": 48.96,
      "learning_rate": 0.0001,
      "loss": 0.0012,
      "step": 3624
    },
    {
      "epoch": 49.03,
      "learning_rate": 0.0001,
      "loss": 0.002,
      "step": 3628
    },
    {
      "epoch": 49.08,
      "learning_rate": 0.0001,
      "loss": 0.0022,
      "step": 3632
    },
    {
      "epoch": 49.13,
      "learning_rate": 0.0001,
      "loss": 0.0012,
      "step": 3636
    },
    {
      "epoch": 49.19,
      "learning_rate": 0.0001,
      "loss": 0.002,
      "step": 3640
    },
    {
      "epoch": 49.24,
      "learning_rate": 0.0001,
      "loss": 0.0007,
      "step": 3644
    },
    {
      "epoch": 49.29,
      "learning_rate": 0.0001,
      "loss": 0.0006,
      "step": 3648
    },
    {
      "epoch": 49.29,
      "eval_exec": 0.7388781431334622,
      "eval_loss": 0.2956254184246063,
      "eval_runtime": 295.8069,
      "eval_samples_per_second": 3.496,
      "step": 3648
    },
    {
      "epoch": 49.35,
      "learning_rate": 0.0001,
      "loss": 0.001,
      "step": 3652
    },
    {
      "epoch": 49.4,
      "learning_rate": 0.0001,
      "loss": 0.0014,
      "step": 3656
    },
    {
      "epoch": 49.45,
      "learning_rate": 0.0001,
      "loss": 0.0007,
      "step": 3660
    },
    {
      "epoch": 49.51,
      "learning_rate": 0.0001,
      "loss": 0.003,
      "step": 3664
    },
    {
      "epoch": 49.56,
      "learning_rate": 0.0001,
      "loss": 0.0011,
      "step": 3668
    },
    {
      "epoch": 49.62,
      "learning_rate": 0.0001,
      "loss": 0.0025,
      "step": 3672
    },
    {
      "epoch": 49.67,
      "learning_rate": 0.0001,
      "loss": 0.0009,
      "step": 3676
    },
    {
      "epoch": 49.72,
      "learning_rate": 0.0001,
      "loss": 0.0007,
      "step": 3680
    },
    {
      "epoch": 49.78,
      "learning_rate": 0.0001,
      "loss": 0.0015,
      "step": 3684
    },
    {
      "epoch": 49.83,
      "learning_rate": 0.0001,
      "loss": 0.0007,
      "step": 3688
    },
    {
      "epoch": 49.88,
      "learning_rate": 0.0001,
      "loss": 0.0007,
      "step": 3692
    },
    {
      "epoch": 49.94,
      "learning_rate": 0.0001,
      "loss": 0.0006,
      "step": 3696
    },
    {
      "epoch": 49.99,
      "learning_rate": 0.0001,
      "loss": 0.0014,
      "step": 3700
    },
    {
      "epoch": 50.05,
      "learning_rate": 0.0001,
      "loss": 0.0014,
      "step": 3704
    },
    {
      "epoch": 50.11,
      "learning_rate": 0.0001,
      "loss": 0.0011,
      "step": 3708
    },
    {
      "epoch": 50.16,
      "learning_rate": 0.0001,
      "loss": 0.0015,
      "step": 3712
    },
    {
      "epoch": 50.16,
      "eval_exec": 0.7437137330754352,
      "eval_loss": 0.30492928624153137,
      "eval_runtime": 291.3729,
      "eval_samples_per_second": 3.549,
      "step": 3712
    },
    {
      "epoch": 50.21,
      "learning_rate": 0.0001,
      "loss": 0.0003,
      "step": 3716
    },
    {
      "epoch": 50.27,
      "learning_rate": 0.0001,
      "loss": 0.0008,
      "step": 3720
    },
    {
      "epoch": 50.32,
      "learning_rate": 0.0001,
      "loss": 0.0006,
      "step": 3724
    },
    {
      "epoch": 50.37,
      "learning_rate": 0.0001,
      "loss": 0.0004,
      "step": 3728
    },
    {
      "epoch": 50.43,
      "learning_rate": 0.0001,
      "loss": 0.0004,
      "step": 3732
    },
    {
      "epoch": 50.48,
      "learning_rate": 0.0001,
      "loss": 0.0005,
      "step": 3736
    },
    {
      "epoch": 50.53,
      "learning_rate": 0.0001,
      "loss": 0.0022,
      "step": 3740
    },
    {
      "epoch": 50.59,
      "learning_rate": 0.0001,
      "loss": 0.0011,
      "step": 3744
    },
    {
      "epoch": 50.64,
      "learning_rate": 0.0001,
      "loss": 0.0003,
      "step": 3748
    },
    {
      "epoch": 50.7,
      "learning_rate": 0.0001,
      "loss": 0.0004,
      "step": 3752
    },
    {
      "epoch": 50.75,
      "learning_rate": 0.0001,
      "loss": 0.0015,
      "step": 3756
    },
    {
      "epoch": 50.8,
      "learning_rate": 0.0001,
      "loss": 0.0013,
      "step": 3760
    },
    {
      "epoch": 50.86,
      "learning_rate": 0.0001,
      "loss": 0.0004,
      "step": 3764
    },
    {
      "epoch": 50.91,
      "learning_rate": 0.0001,
      "loss": 0.0004,
      "step": 3768
    },
    {
      "epoch": 50.96,
      "learning_rate": 0.0001,
      "loss": 0.0015,
      "step": 3772
    },
    {
      "epoch": 51.03,
      "learning_rate": 0.0001,
      "loss": 0.0015,
      "step": 3776
    },
    {
      "epoch": 51.03,
      "eval_exec": 0.7495164410058027,
      "eval_loss": 0.32364675402641296,
      "eval_runtime": 300.6402,
      "eval_samples_per_second": 3.439,
      "step": 3776
    },
    {
      "epoch": 51.08,
      "learning_rate": 0.0001,
      "loss": 0.0012,
      "step": 3780
    },
    {
      "epoch": 51.13,
      "learning_rate": 0.0001,
      "loss": 0.0004,
      "step": 3784
    },
    {
      "epoch": 51.19,
      "learning_rate": 0.0001,
      "loss": 0.0005,
      "step": 3788
    },
    {
      "epoch": 51.24,
      "learning_rate": 0.0001,
      "loss": 0.0012,
      "step": 3792
    },
    {
      "epoch": 51.29,
      "learning_rate": 0.0001,
      "loss": 0.0007,
      "step": 3796
    },
    {
      "epoch": 51.35,
      "learning_rate": 0.0001,
      "loss": 0.0018,
      "step": 3800
    },
    {
      "epoch": 51.4,
      "learning_rate": 0.0001,
      "loss": 0.001,
      "step": 3804
    },
    {
      "epoch": 51.45,
      "learning_rate": 0.0001,
      "loss": 0.0003,
      "step": 3808
    },
    {
      "epoch": 51.51,
      "learning_rate": 0.0001,
      "loss": 0.0028,
      "step": 3812
    },
    {
      "epoch": 51.56,
      "learning_rate": 0.0001,
      "loss": 0.0009,
      "step": 3816
    },
    {
      "epoch": 51.62,
      "learning_rate": 0.0001,
      "loss": 0.0013,
      "step": 3820
    },
    {
      "epoch": 51.67,
      "learning_rate": 0.0001,
      "loss": 0.0005,
      "step": 3824
    },
    {
      "epoch": 51.72,
      "learning_rate": 0.0001,
      "loss": 0.0008,
      "step": 3828
    },
    {
      "epoch": 51.78,
      "learning_rate": 0.0001,
      "loss": 0.0014,
      "step": 3832
    },
    {
      "epoch": 51.83,
      "learning_rate": 0.0001,
      "loss": 0.0016,
      "step": 3836
    },
    {
      "epoch": 51.88,
      "learning_rate": 0.0001,
      "loss": 0.0016,
      "step": 3840
    },
    {
      "epoch": 51.88,
      "eval_exec": 0.7253384912959381,
      "eval_loss": 0.30479735136032104,
      "eval_runtime": 283.847,
      "eval_samples_per_second": 3.643,
      "step": 3840
    },
    {
      "epoch": 51.94,
      "learning_rate": 0.0001,
      "loss": 0.0004,
      "step": 3844
    },
    {
      "epoch": 51.99,
      "learning_rate": 0.0001,
      "loss": 0.0032,
      "step": 3848
    },
    {
      "epoch": 52.05,
      "learning_rate": 0.0001,
      "loss": 0.0029,
      "step": 3852
    },
    {
      "epoch": 52.11,
      "learning_rate": 0.0001,
      "loss": 0.0012,
      "step": 3856
    },
    {
      "epoch": 52.16,
      "learning_rate": 0.0001,
      "loss": 0.0015,
      "step": 3860
    },
    {
      "epoch": 52.21,
      "learning_rate": 0.0001,
      "loss": 0.0006,
      "step": 3864
    },
    {
      "epoch": 52.27,
      "learning_rate": 0.0001,
      "loss": 0.001,
      "step": 3868
    },
    {
      "epoch": 52.32,
      "learning_rate": 0.0001,
      "loss": 0.0018,
      "step": 3872
    },
    {
      "epoch": 52.37,
      "learning_rate": 0.0001,
      "loss": 0.0011,
      "step": 3876
    },
    {
      "epoch": 52.43,
      "learning_rate": 0.0001,
      "loss": 0.0004,
      "step": 3880
    },
    {
      "epoch": 52.48,
      "learning_rate": 0.0001,
      "loss": 0.0011,
      "step": 3884
    },
    {
      "epoch": 52.53,
      "learning_rate": 0.0001,
      "loss": 0.0024,
      "step": 3888
    },
    {
      "epoch": 52.59,
      "learning_rate": 0.0001,
      "loss": 0.0015,
      "step": 3892
    },
    {
      "epoch": 52.64,
      "learning_rate": 0.0001,
      "loss": 0.0007,
      "step": 3896
    },
    {
      "epoch": 52.7,
      "learning_rate": 0.0001,
      "loss": 0.0009,
      "step": 3900
    },
    {
      "epoch": 52.75,
      "learning_rate": 0.0001,
      "loss": 0.001,
      "step": 3904
    },
    {
      "epoch": 52.75,
      "eval_exec": 0.7214700193423598,
      "eval_loss": 0.3040502965450287,
      "eval_runtime": 296.1224,
      "eval_samples_per_second": 3.492,
      "step": 3904
    },
    {
      "epoch": 52.8,
      "learning_rate": 0.0001,
      "loss": 0.0022,
      "step": 3908
    },
    {
      "epoch": 52.86,
      "learning_rate": 0.0001,
      "loss": 0.0007,
      "step": 3912
    },
    {
      "epoch": 52.91,
      "learning_rate": 0.0001,
      "loss": 0.0006,
      "step": 3916
    },
    {
      "epoch": 52.96,
      "learning_rate": 0.0001,
      "loss": 0.0008,
      "step": 3920
    },
    {
      "epoch": 53.03,
      "learning_rate": 0.0001,
      "loss": 0.0016,
      "step": 3924
    },
    {
      "epoch": 53.08,
      "learning_rate": 0.0001,
      "loss": 0.0014,
      "step": 3928
    },
    {
      "epoch": 53.13,
      "learning_rate": 0.0001,
      "loss": 0.0012,
      "step": 3932
    },
    {
      "epoch": 53.19,
      "learning_rate": 0.0001,
      "loss": 0.0007,
      "step": 3936
    },
    {
      "epoch": 53.24,
      "learning_rate": 0.0001,
      "loss": 0.0004,
      "step": 3940
    },
    {
      "epoch": 53.29,
      "learning_rate": 0.0001,
      "loss": 0.0011,
      "step": 3944
    },
    {
      "epoch": 53.35,
      "learning_rate": 0.0001,
      "loss": 0.0008,
      "step": 3948
    },
    {
      "epoch": 53.4,
      "learning_rate": 0.0001,
      "loss": 0.0005,
      "step": 3952
    },
    {
      "epoch": 53.45,
      "learning_rate": 0.0001,
      "loss": 0.0009,
      "step": 3956
    },
    {
      "epoch": 53.51,
      "learning_rate": 0.0001,
      "loss": 0.0007,
      "step": 3960
    },
    {
      "epoch": 53.56,
      "learning_rate": 0.0001,
      "loss": 0.0008,
      "step": 3964
    },
    {
      "epoch": 53.62,
      "learning_rate": 0.0001,
      "loss": 0.0004,
      "step": 3968
    },
    {
      "epoch": 53.62,
      "eval_exec": 0.7330754352030948,
      "eval_loss": 0.3174877464771271,
      "eval_runtime": 296.7562,
      "eval_samples_per_second": 3.484,
      "step": 3968
    },
    {
      "epoch": 53.67,
      "learning_rate": 0.0001,
      "loss": 0.0009,
      "step": 3972
    },
    {
      "epoch": 53.72,
      "learning_rate": 0.0001,
      "loss": 0.0003,
      "step": 3976
    },
    {
      "epoch": 53.78,
      "learning_rate": 0.0001,
      "loss": 0.0019,
      "step": 3980
    },
    {
      "epoch": 53.83,
      "learning_rate": 0.0001,
      "loss": 0.0019,
      "step": 3984
    },
    {
      "epoch": 53.88,
      "learning_rate": 0.0001,
      "loss": 0.0009,
      "step": 3988
    },
    {
      "epoch": 53.94,
      "learning_rate": 0.0001,
      "loss": 0.0006,
      "step": 3992
    },
    {
      "epoch": 53.99,
      "learning_rate": 0.0001,
      "loss": 0.0008,
      "step": 3996
    },
    {
      "epoch": 54.05,
      "learning_rate": 0.0001,
      "loss": 0.0018,
      "step": 4000
    },
    {
      "epoch": 54.11,
      "learning_rate": 0.0001,
      "loss": 0.0005,
      "step": 4004
    },
    {
      "epoch": 54.16,
      "learning_rate": 0.0001,
      "loss": 0.0006,
      "step": 4008
    },
    {
      "epoch": 54.21,
      "learning_rate": 0.0001,
      "loss": 0.0005,
      "step": 4012
    },
    {
      "epoch": 54.27,
      "learning_rate": 0.0001,
      "loss": 0.0011,
      "step": 4016
    },
    {
      "epoch": 54.32,
      "learning_rate": 0.0001,
      "loss": 0.0018,
      "step": 4020
    },
    {
      "epoch": 54.37,
      "learning_rate": 0.0001,
      "loss": 0.0018,
      "step": 4024
    },
    {
      "epoch": 54.43,
      "learning_rate": 0.0001,
      "loss": 0.0012,
      "step": 4028
    },
    {
      "epoch": 54.48,
      "learning_rate": 0.0001,
      "loss": 0.0002,
      "step": 4032
    },
    {
      "epoch": 54.48,
      "eval_exec": 0.7340425531914894,
      "eval_loss": 0.3035335838794708,
      "eval_runtime": 288.5745,
      "eval_samples_per_second": 3.583,
      "step": 4032
    },
    {
      "epoch": 54.53,
      "learning_rate": 0.0001,
      "loss": 0.0019,
      "step": 4036
    },
    {
      "epoch": 54.59,
      "learning_rate": 0.0001,
      "loss": 0.0017,
      "step": 4040
    },
    {
      "epoch": 54.64,
      "learning_rate": 0.0001,
      "loss": 0.0033,
      "step": 4044
    },
    {
      "epoch": 54.7,
      "learning_rate": 0.0001,
      "loss": 0.0004,
      "step": 4048
    },
    {
      "epoch": 54.75,
      "learning_rate": 0.0001,
      "loss": 0.0012,
      "step": 4052
    },
    {
      "epoch": 54.8,
      "learning_rate": 0.0001,
      "loss": 0.0013,
      "step": 4056
    },
    {
      "epoch": 54.86,
      "learning_rate": 0.0001,
      "loss": 0.0024,
      "step": 4060
    },
    {
      "epoch": 54.91,
      "learning_rate": 0.0001,
      "loss": 0.0004,
      "step": 4064
    },
    {
      "epoch": 54.96,
      "learning_rate": 0.0001,
      "loss": 0.0004,
      "step": 4068
    },
    {
      "epoch": 55.03,
      "learning_rate": 0.0001,
      "loss": 0.0017,
      "step": 4072
    },
    {
      "epoch": 55.08,
      "learning_rate": 0.0001,
      "loss": 0.0014,
      "step": 4076
    },
    {
      "epoch": 55.13,
      "learning_rate": 0.0001,
      "loss": 0.0007,
      "step": 4080
    },
    {
      "epoch": 55.19,
      "learning_rate": 0.0001,
      "loss": 0.0005,
      "step": 4084
    },
    {
      "epoch": 55.24,
      "learning_rate": 0.0001,
      "loss": 0.0029,
      "step": 4088
    },
    {
      "epoch": 55.29,
      "learning_rate": 0.0001,
      "loss": 0.0018,
      "step": 4092
    },
    {
      "epoch": 55.35,
      "learning_rate": 0.0001,
      "loss": 0.0015,
      "step": 4096
    },
    {
      "epoch": 55.35,
      "eval_exec": 0.7321083172147002,
      "eval_loss": 0.3005565404891968,
      "eval_runtime": 293.428,
      "eval_samples_per_second": 3.524,
      "step": 4096
    },
    {
      "epoch": 55.4,
      "learning_rate": 0.0001,
      "loss": 0.0009,
      "step": 4100
    },
    {
      "epoch": 55.45,
      "learning_rate": 0.0001,
      "loss": 0.0007,
      "step": 4104
    },
    {
      "epoch": 55.51,
      "learning_rate": 0.0001,
      "loss": 0.0012,
      "step": 4108
    },
    {
      "epoch": 55.56,
      "learning_rate": 0.0001,
      "loss": 0.0024,
      "step": 4112
    },
    {
      "epoch": 55.62,
      "learning_rate": 0.0001,
      "loss": 0.0009,
      "step": 4116
    },
    {
      "epoch": 55.67,
      "learning_rate": 0.0001,
      "loss": 0.0006,
      "step": 4120
    },
    {
      "epoch": 55.72,
      "learning_rate": 0.0001,
      "loss": 0.0005,
      "step": 4124
    },
    {
      "epoch": 55.78,
      "learning_rate": 0.0001,
      "loss": 0.0013,
      "step": 4128
    },
    {
      "epoch": 55.83,
      "learning_rate": 0.0001,
      "loss": 0.0011,
      "step": 4132
    },
    {
      "epoch": 55.88,
      "learning_rate": 0.0001,
      "loss": 0.001,
      "step": 4136
    },
    {
      "epoch": 55.94,
      "learning_rate": 0.0001,
      "loss": 0.0008,
      "step": 4140
    },
    {
      "epoch": 55.99,
      "learning_rate": 0.0001,
      "loss": 0.0005,
      "step": 4144
    },
    {
      "epoch": 56.05,
      "learning_rate": 0.0001,
      "loss": 0.0013,
      "step": 4148
    },
    {
      "epoch": 56.11,
      "learning_rate": 0.0001,
      "loss": 0.0009,
      "step": 4152
    },
    {
      "epoch": 56.16,
      "learning_rate": 0.0001,
      "loss": 0.0008,
      "step": 4156
    },
    {
      "epoch": 56.21,
      "learning_rate": 0.0001,
      "loss": 0.0011,
      "step": 4160
    },
    {
      "epoch": 56.21,
      "eval_exec": 0.723404255319149,
      "eval_loss": 0.3162212371826172,
      "eval_runtime": 294.7911,
      "eval_samples_per_second": 3.508,
      "step": 4160
    },
    {
      "epoch": 56.27,
      "learning_rate": 0.0001,
      "loss": 0.0012,
      "step": 4164
    },
    {
      "epoch": 56.32,
      "learning_rate": 0.0001,
      "loss": 0.0055,
      "step": 4168
    },
    {
      "epoch": 56.37,
      "learning_rate": 0.0001,
      "loss": 0.0005,
      "step": 4172
    },
    {
      "epoch": 56.43,
      "learning_rate": 0.0001,
      "loss": 0.0003,
      "step": 4176
    },
    {
      "epoch": 56.48,
      "learning_rate": 0.0001,
      "loss": 0.0008,
      "step": 4180
    },
    {
      "epoch": 56.53,
      "learning_rate": 0.0001,
      "loss": 0.0013,
      "step": 4184
    },
    {
      "epoch": 56.59,
      "learning_rate": 0.0001,
      "loss": 0.0006,
      "step": 4188
    },
    {
      "epoch": 56.64,
      "learning_rate": 0.0001,
      "loss": 0.0025,
      "step": 4192
    },
    {
      "epoch": 56.7,
      "learning_rate": 0.0001,
      "loss": 0.0003,
      "step": 4196
    },
    {
      "epoch": 56.75,
      "learning_rate": 0.0001,
      "loss": 0.0007,
      "step": 4200
    },
    {
      "epoch": 56.8,
      "learning_rate": 0.0001,
      "loss": 0.0005,
      "step": 4204
    },
    {
      "epoch": 56.86,
      "learning_rate": 0.0001,
      "loss": 0.0007,
      "step": 4208
    },
    {
      "epoch": 56.91,
      "learning_rate": 0.0001,
      "loss": 0.0009,
      "step": 4212
    },
    {
      "epoch": 56.96,
      "learning_rate": 0.0001,
      "loss": 0.0008,
      "step": 4216
    },
    {
      "epoch": 57.03,
      "learning_rate": 0.0001,
      "loss": 0.0014,
      "step": 4220
    },
    {
      "epoch": 57.08,
      "learning_rate": 0.0001,
      "loss": 0.003,
      "step": 4224
    },
    {
      "epoch": 57.08,
      "eval_exec": 0.706963249516441,
      "eval_loss": 0.3172769248485565,
      "eval_runtime": 287.3743,
      "eval_samples_per_second": 3.598,
      "step": 4224
    },
    {
      "epoch": 57.13,
      "learning_rate": 0.0001,
      "loss": 0.0007,
      "step": 4228
    },
    {
      "epoch": 57.19,
      "learning_rate": 0.0001,
      "loss": 0.0012,
      "step": 4232
    },
    {
      "epoch": 57.24,
      "learning_rate": 0.0001,
      "loss": 0.0004,
      "step": 4236
    },
    {
      "epoch": 57.29,
      "learning_rate": 0.0001,
      "loss": 0.0013,
      "step": 4240
    },
    {
      "epoch": 57.35,
      "learning_rate": 0.0001,
      "loss": 0.0019,
      "step": 4244
    },
    {
      "epoch": 57.4,
      "learning_rate": 0.0001,
      "loss": 0.0008,
      "step": 4248
    },
    {
      "epoch": 57.45,
      "learning_rate": 0.0001,
      "loss": 0.0011,
      "step": 4252
    },
    {
      "epoch": 57.51,
      "learning_rate": 0.0001,
      "loss": 0.0015,
      "step": 4256
    },
    {
      "epoch": 57.56,
      "learning_rate": 0.0001,
      "loss": 0.001,
      "step": 4260
    },
    {
      "epoch": 57.62,
      "learning_rate": 0.0001,
      "loss": 0.0019,
      "step": 4264
    },
    {
      "epoch": 57.67,
      "learning_rate": 0.0001,
      "loss": 0.0013,
      "step": 4268
    },
    {
      "epoch": 57.72,
      "learning_rate": 0.0001,
      "loss": 0.0007,
      "step": 4272
    },
    {
      "epoch": 57.78,
      "learning_rate": 0.0001,
      "loss": 0.0013,
      "step": 4276
    },
    {
      "epoch": 57.83,
      "learning_rate": 0.0001,
      "loss": 0.0011,
      "step": 4280
    },
    {
      "epoch": 57.88,
      "learning_rate": 0.0001,
      "loss": 0.0012,
      "step": 4284
    },
    {
      "epoch": 57.94,
      "learning_rate": 0.0001,
      "loss": 0.0021,
      "step": 4288
    },
    {
      "epoch": 57.94,
      "eval_exec": 0.7224371373307543,
      "eval_loss": 0.2908557057380676,
      "eval_runtime": 293.2594,
      "eval_samples_per_second": 3.526,
      "step": 4288
    },
    {
      "epoch": 57.99,
      "learning_rate": 0.0001,
      "loss": 0.0016,
      "step": 4292
    },
    {
      "epoch": 58.05,
      "learning_rate": 0.0001,
      "loss": 0.0012,
      "step": 4296
    },
    {
      "epoch": 58.11,
      "learning_rate": 0.0001,
      "loss": 0.0009,
      "step": 4300
    },
    {
      "epoch": 58.16,
      "learning_rate": 0.0001,
      "loss": 0.0032,
      "step": 4304
    },
    {
      "epoch": 58.21,
      "learning_rate": 0.0001,
      "loss": 0.0003,
      "step": 4308
    },
    {
      "epoch": 58.27,
      "learning_rate": 0.0001,
      "loss": 0.0009,
      "step": 4312
    },
    {
      "epoch": 58.32,
      "learning_rate": 0.0001,
      "loss": 0.0006,
      "step": 4316
    },
    {
      "epoch": 58.37,
      "learning_rate": 0.0001,
      "loss": 0.0006,
      "step": 4320
    },
    {
      "epoch": 58.43,
      "learning_rate": 0.0001,
      "loss": 0.0012,
      "step": 4324
    },
    {
      "epoch": 58.48,
      "learning_rate": 0.0001,
      "loss": 0.0003,
      "step": 4328
    },
    {
      "epoch": 58.53,
      "learning_rate": 0.0001,
      "loss": 0.0009,
      "step": 4332
    },
    {
      "epoch": 58.59,
      "learning_rate": 0.0001,
      "loss": 0.0007,
      "step": 4336
    },
    {
      "epoch": 58.64,
      "learning_rate": 0.0001,
      "loss": 0.0018,
      "step": 4340
    },
    {
      "epoch": 58.7,
      "learning_rate": 0.0001,
      "loss": 0.0013,
      "step": 4344
    },
    {
      "epoch": 58.75,
      "learning_rate": 0.0001,
      "loss": 0.0008,
      "step": 4348
    },
    {
      "epoch": 58.8,
      "learning_rate": 0.0001,
      "loss": 0.0008,
      "step": 4352
    },
    {
      "epoch": 58.8,
      "eval_exec": 0.7166344294003868,
      "eval_loss": 0.31116729974746704,
      "eval_runtime": 295.9051,
      "eval_samples_per_second": 3.494,
      "step": 4352
    },
    {
      "epoch": 58.86,
      "learning_rate": 0.0001,
      "loss": 0.0018,
      "step": 4356
    },
    {
      "epoch": 58.91,
      "learning_rate": 0.0001,
      "loss": 0.0025,
      "step": 4360
    },
    {
      "epoch": 58.96,
      "learning_rate": 0.0001,
      "loss": 0.0006,
      "step": 4364
    },
    {
      "epoch": 59.03,
      "learning_rate": 0.0001,
      "loss": 0.001,
      "step": 4368
    },
    {
      "epoch": 59.08,
      "learning_rate": 0.0001,
      "loss": 0.001,
      "step": 4372
    },
    {
      "epoch": 59.13,
      "learning_rate": 0.0001,
      "loss": 0.0012,
      "step": 4376
    },
    {
      "epoch": 59.19,
      "learning_rate": 0.0001,
      "loss": 0.0011,
      "step": 4380
    },
    {
      "epoch": 59.24,
      "learning_rate": 0.0001,
      "loss": 0.0014,
      "step": 4384
    },
    {
      "epoch": 59.29,
      "learning_rate": 0.0001,
      "loss": 0.0009,
      "step": 4388
    },
    {
      "epoch": 59.35,
      "learning_rate": 0.0001,
      "loss": 0.0009,
      "step": 4392
    },
    {
      "epoch": 59.4,
      "learning_rate": 0.0001,
      "loss": 0.0005,
      "step": 4396
    },
    {
      "epoch": 59.45,
      "learning_rate": 0.0001,
      "loss": 0.0007,
      "step": 4400
    },
    {
      "epoch": 59.51,
      "learning_rate": 0.0001,
      "loss": 0.0007,
      "step": 4404
    },
    {
      "epoch": 59.56,
      "learning_rate": 0.0001,
      "loss": 0.0014,
      "step": 4408
    },
    {
      "epoch": 59.62,
      "learning_rate": 0.0001,
      "loss": 0.0006,
      "step": 4412
    },
    {
      "epoch": 59.67,
      "learning_rate": 0.0001,
      "loss": 0.0012,
      "step": 4416
    },
    {
      "epoch": 59.67,
      "eval_exec": 0.7388781431334622,
      "eval_loss": 0.3091637194156647,
      "eval_runtime": 297.2444,
      "eval_samples_per_second": 3.479,
      "step": 4416
    },
    {
      "epoch": 59.72,
      "learning_rate": 0.0001,
      "loss": 0.0023,
      "step": 4420
    },
    {
      "epoch": 59.78,
      "learning_rate": 0.0001,
      "loss": 0.0033,
      "step": 4424
    },
    {
      "epoch": 59.83,
      "learning_rate": 0.0001,
      "loss": 0.0024,
      "step": 4428
    },
    {
      "epoch": 59.88,
      "learning_rate": 0.0001,
      "loss": 0.0016,
      "step": 4432
    },
    {
      "epoch": 59.94,
      "learning_rate": 0.0001,
      "loss": 0.0014,
      "step": 4436
    },
    {
      "epoch": 59.99,
      "learning_rate": 0.0001,
      "loss": 0.0009,
      "step": 4440
    },
    {
      "epoch": 60.05,
      "learning_rate": 0.0001,
      "loss": 0.0014,
      "step": 4444
    },
    {
      "epoch": 60.11,
      "learning_rate": 0.0001,
      "loss": 0.0015,
      "step": 4448
    },
    {
      "epoch": 60.16,
      "learning_rate": 0.0001,
      "loss": 0.0004,
      "step": 4452
    },
    {
      "epoch": 60.21,
      "learning_rate": 0.0001,
      "loss": 0.0009,
      "step": 4456
    },
    {
      "epoch": 60.27,
      "learning_rate": 0.0001,
      "loss": 0.0004,
      "step": 4460
    },
    {
      "epoch": 60.32,
      "learning_rate": 0.0001,
      "loss": 0.0011,
      "step": 4464
    },
    {
      "epoch": 60.37,
      "learning_rate": 0.0001,
      "loss": 0.0007,
      "step": 4468
    },
    {
      "epoch": 60.43,
      "learning_rate": 0.0001,
      "loss": 0.0008,
      "step": 4472
    },
    {
      "epoch": 60.48,
      "learning_rate": 0.0001,
      "loss": 0.0014,
      "step": 4476
    },
    {
      "epoch": 60.53,
      "learning_rate": 0.0001,
      "loss": 0.0021,
      "step": 4480
    },
    {
      "epoch": 60.53,
      "eval_exec": 0.7243713733075435,
      "eval_loss": 0.3012126684188843,
      "eval_runtime": 289.2665,
      "eval_samples_per_second": 3.575,
      "step": 4480
    },
    {
      "epoch": 60.59,
      "learning_rate": 0.0001,
      "loss": 0.001,
      "step": 4484
    },
    {
      "epoch": 60.64,
      "learning_rate": 0.0001,
      "loss": 0.0013,
      "step": 4488
    },
    {
      "epoch": 60.7,
      "learning_rate": 0.0001,
      "loss": 0.0006,
      "step": 4492
    },
    {
      "epoch": 60.75,
      "learning_rate": 0.0001,
      "loss": 0.0006,
      "step": 4496
    },
    {
      "epoch": 60.8,
      "learning_rate": 0.0001,
      "loss": 0.001,
      "step": 4500
    },
    {
      "epoch": 60.86,
      "learning_rate": 0.0001,
      "loss": 0.0031,
      "step": 4504
    },
    {
      "epoch": 60.91,
      "learning_rate": 0.0001,
      "loss": 0.0009,
      "step": 4508
    },
    {
      "epoch": 60.96,
      "learning_rate": 0.0001,
      "loss": 0.0009,
      "step": 4512
    },
    {
      "epoch": 61.03,
      "learning_rate": 0.0001,
      "loss": 0.0012,
      "step": 4516
    },
    {
      "epoch": 61.08,
      "learning_rate": 0.0001,
      "loss": 0.0006,
      "step": 4520
    },
    {
      "epoch": 61.13,
      "learning_rate": 0.0001,
      "loss": 0.0007,
      "step": 4524
    },
    {
      "epoch": 61.19,
      "learning_rate": 0.0001,
      "loss": 0.001,
      "step": 4528
    },
    {
      "epoch": 61.24,
      "learning_rate": 0.0001,
      "loss": 0.0004,
      "step": 4532
    },
    {
      "epoch": 61.29,
      "learning_rate": 0.0001,
      "loss": 0.0005,
      "step": 4536
    },
    {
      "epoch": 61.35,
      "learning_rate": 0.0001,
      "loss": 0.0029,
      "step": 4540
    },
    {
      "epoch": 61.4,
      "learning_rate": 0.0001,
      "loss": 0.0007,
      "step": 4544
    },
    {
      "epoch": 61.4,
      "eval_exec": 0.7272727272727273,
      "eval_loss": 0.30137544870376587,
      "eval_runtime": 286.7428,
      "eval_samples_per_second": 3.606,
      "step": 4544
    },
    {
      "epoch": 61.45,
      "learning_rate": 0.0001,
      "loss": 0.0015,
      "step": 4548
    },
    {
      "epoch": 61.51,
      "learning_rate": 0.0001,
      "loss": 0.0017,
      "step": 4552
    },
    {
      "epoch": 61.56,
      "learning_rate": 0.0001,
      "loss": 0.0008,
      "step": 4556
    },
    {
      "epoch": 61.62,
      "learning_rate": 0.0001,
      "loss": 0.0009,
      "step": 4560
    },
    {
      "epoch": 61.67,
      "learning_rate": 0.0001,
      "loss": 0.001,
      "step": 4564
    },
    {
      "epoch": 61.72,
      "learning_rate": 0.0001,
      "loss": 0.0009,
      "step": 4568
    },
    {
      "epoch": 61.78,
      "learning_rate": 0.0001,
      "loss": 0.0011,
      "step": 4572
    },
    {
      "epoch": 61.83,
      "learning_rate": 0.0001,
      "loss": 0.0007,
      "step": 4576
    },
    {
      "epoch": 61.88,
      "learning_rate": 0.0001,
      "loss": 0.001,
      "step": 4580
    },
    {
      "epoch": 61.94,
      "learning_rate": 0.0001,
      "loss": 0.001,
      "step": 4584
    },
    {
      "epoch": 61.99,
      "learning_rate": 0.0001,
      "loss": 0.001,
      "step": 4588
    },
    {
      "epoch": 62.05,
      "learning_rate": 0.0001,
      "loss": 0.0026,
      "step": 4592
    },
    {
      "epoch": 62.11,
      "learning_rate": 0.0001,
      "loss": 0.0007,
      "step": 4596
    },
    {
      "epoch": 62.16,
      "learning_rate": 0.0001,
      "loss": 0.0008,
      "step": 4600
    },
    {
      "epoch": 62.21,
      "learning_rate": 0.0001,
      "loss": 0.0003,
      "step": 4604
    },
    {
      "epoch": 62.27,
      "learning_rate": 0.0001,
      "loss": 0.0007,
      "step": 4608
    },
    {
      "epoch": 62.27,
      "eval_exec": 0.723404255319149,
      "eval_loss": 0.3098083734512329,
      "eval_runtime": 291.4042,
      "eval_samples_per_second": 3.548,
      "step": 4608
    },
    {
      "epoch": 62.32,
      "learning_rate": 0.0001,
      "loss": 0.0038,
      "step": 4612
    },
    {
      "epoch": 62.37,
      "learning_rate": 0.0001,
      "loss": 0.0013,
      "step": 4616
    },
    {
      "epoch": 62.43,
      "learning_rate": 0.0001,
      "loss": 0.0008,
      "step": 4620
    },
    {
      "epoch": 62.48,
      "learning_rate": 0.0001,
      "loss": 0.0011,
      "step": 4624
    },
    {
      "epoch": 62.53,
      "learning_rate": 0.0001,
      "loss": 0.002,
      "step": 4628
    },
    {
      "epoch": 62.59,
      "learning_rate": 0.0001,
      "loss": 0.0018,
      "step": 4632
    },
    {
      "epoch": 62.64,
      "learning_rate": 0.0001,
      "loss": 0.0005,
      "step": 4636
    },
    {
      "epoch": 62.7,
      "learning_rate": 0.0001,
      "loss": 0.0008,
      "step": 4640
    },
    {
      "epoch": 62.75,
      "learning_rate": 0.0001,
      "loss": 0.0009,
      "step": 4644
    },
    {
      "epoch": 62.8,
      "learning_rate": 0.0001,
      "loss": 0.001,
      "step": 4648
    },
    {
      "epoch": 62.86,
      "learning_rate": 0.0001,
      "loss": 0.0004,
      "step": 4652
    },
    {
      "epoch": 62.91,
      "learning_rate": 0.0001,
      "loss": 0.0003,
      "step": 4656
    },
    {
      "epoch": 62.96,
      "learning_rate": 0.0001,
      "loss": 0.0002,
      "step": 4660
    },
    {
      "epoch": 63.03,
      "learning_rate": 0.0001,
      "loss": 0.0011,
      "step": 4664
    },
    {
      "epoch": 63.08,
      "learning_rate": 0.0001,
      "loss": 0.001,
      "step": 4668
    },
    {
      "epoch": 63.13,
      "learning_rate": 0.0001,
      "loss": 0.0003,
      "step": 4672
    },
    {
      "epoch": 63.13,
      "eval_exec": 0.7437137330754352,
      "eval_loss": 0.32493457198143005,
      "eval_runtime": 288.524,
      "eval_samples_per_second": 3.584,
      "step": 4672
    },
    {
      "epoch": 63.19,
      "learning_rate": 0.0001,
      "loss": 0.0005,
      "step": 4676
    },
    {
      "epoch": 63.24,
      "learning_rate": 0.0001,
      "loss": 0.0002,
      "step": 4680
    },
    {
      "epoch": 63.29,
      "learning_rate": 0.0001,
      "loss": 0.0012,
      "step": 4684
    },
    {
      "epoch": 63.35,
      "learning_rate": 0.0001,
      "loss": 0.0011,
      "step": 4688
    },
    {
      "epoch": 63.4,
      "learning_rate": 0.0001,
      "loss": 0.0007,
      "step": 4692
    },
    {
      "epoch": 63.45,
      "learning_rate": 0.0001,
      "loss": 0.0002,
      "step": 4696
    },
    {
      "epoch": 63.51,
      "learning_rate": 0.0001,
      "loss": 0.0005,
      "step": 4700
    },
    {
      "epoch": 63.56,
      "learning_rate": 0.0001,
      "loss": 0.0012,
      "step": 4704
    },
    {
      "epoch": 63.62,
      "learning_rate": 0.0001,
      "loss": 0.0003,
      "step": 4708
    },
    {
      "epoch": 63.67,
      "learning_rate": 0.0001,
      "loss": 0.0003,
      "step": 4712
    },
    {
      "epoch": 63.72,
      "learning_rate": 0.0001,
      "loss": 0.0009,
      "step": 4716
    },
    {
      "epoch": 63.78,
      "learning_rate": 0.0001,
      "loss": 0.0009,
      "step": 4720
    },
    {
      "epoch": 63.83,
      "learning_rate": 0.0001,
      "loss": 0.0006,
      "step": 4724
    },
    {
      "epoch": 63.88,
      "learning_rate": 0.0001,
      "loss": 0.0004,
      "step": 4728
    },
    {
      "epoch": 63.94,
      "learning_rate": 0.0001,
      "loss": 0.0004,
      "step": 4732
    },
    {
      "epoch": 63.99,
      "learning_rate": 0.0001,
      "loss": 0.0012,
      "step": 4736
    },
    {
      "epoch": 63.99,
      "eval_exec": 0.7427466150870407,
      "eval_loss": 0.3356438875198364,
      "eval_runtime": 295.1379,
      "eval_samples_per_second": 3.503,
      "step": 4736
    },
    {
      "epoch": 64.05,
      "learning_rate": 0.0001,
      "loss": 0.0009,
      "step": 4740
    },
    {
      "epoch": 64.11,
      "learning_rate": 0.0001,
      "loss": 0.0004,
      "step": 4744
    },
    {
      "epoch": 64.16,
      "learning_rate": 0.0001,
      "loss": 0.001,
      "step": 4748
    },
    {
      "epoch": 64.21,
      "learning_rate": 0.0001,
      "loss": 0.0024,
      "step": 4752
    },
    {
      "epoch": 64.27,
      "learning_rate": 0.0001,
      "loss": 0.0006,
      "step": 4756
    },
    {
      "epoch": 64.32,
      "learning_rate": 0.0001,
      "loss": 0.0008,
      "step": 4760
    },
    {
      "epoch": 64.37,
      "learning_rate": 0.0001,
      "loss": 0.0005,
      "step": 4764
    },
    {
      "epoch": 64.43,
      "learning_rate": 0.0001,
      "loss": 0.0002,
      "step": 4768
    },
    {
      "epoch": 64.48,
      "learning_rate": 0.0001,
      "loss": 0.0007,
      "step": 4772
    },
    {
      "epoch": 64.53,
      "learning_rate": 0.0001,
      "loss": 0.0006,
      "step": 4776
    },
    {
      "epoch": 64.59,
      "learning_rate": 0.0001,
      "loss": 0.0002,
      "step": 4780
    },
    {
      "epoch": 64.64,
      "learning_rate": 0.0001,
      "loss": 0.0002,
      "step": 4784
    },
    {
      "epoch": 64.7,
      "learning_rate": 0.0001,
      "loss": 0.0003,
      "step": 4788
    },
    {
      "epoch": 64.75,
      "learning_rate": 0.0001,
      "loss": 0.0009,
      "step": 4792
    },
    {
      "epoch": 64.8,
      "learning_rate": 0.0001,
      "loss": 0.0008,
      "step": 4796
    },
    {
      "epoch": 64.86,
      "learning_rate": 0.0001,
      "loss": 0.0002,
      "step": 4800
    },
    {
      "epoch": 64.86,
      "eval_exec": 0.7398452611218569,
      "eval_loss": 0.3491382598876953,
      "eval_runtime": 296.7752,
      "eval_samples_per_second": 3.484,
      "step": 4800
    },
    {
      "epoch": 64.91,
      "learning_rate": 0.0001,
      "loss": 0.001,
      "step": 4804
    },
    {
      "epoch": 64.96,
      "learning_rate": 0.0001,
      "loss": 0.0014,
      "step": 4808
    },
    {
      "epoch": 65.03,
      "learning_rate": 0.0001,
      "loss": 0.0013,
      "step": 4812
    },
    {
      "epoch": 65.08,
      "learning_rate": 0.0001,
      "loss": 0.0008,
      "step": 4816
    },
    {
      "epoch": 65.13,
      "learning_rate": 0.0001,
      "loss": 0.0005,
      "step": 4820
    },
    {
      "epoch": 65.19,
      "learning_rate": 0.0001,
      "loss": 0.0011,
      "step": 4824
    },
    {
      "epoch": 65.24,
      "learning_rate": 0.0001,
      "loss": 0.002,
      "step": 4828
    },
    {
      "epoch": 65.29,
      "learning_rate": 0.0001,
      "loss": 0.0007,
      "step": 4832
    },
    {
      "epoch": 65.35,
      "learning_rate": 0.0001,
      "loss": 0.0011,
      "step": 4836
    },
    {
      "epoch": 65.4,
      "learning_rate": 0.0001,
      "loss": 0.0008,
      "step": 4840
    },
    {
      "epoch": 65.45,
      "learning_rate": 0.0001,
      "loss": 0.0008,
      "step": 4844
    },
    {
      "epoch": 65.51,
      "learning_rate": 0.0001,
      "loss": 0.0004,
      "step": 4848
    },
    {
      "epoch": 65.56,
      "learning_rate": 0.0001,
      "loss": 0.0007,
      "step": 4852
    },
    {
      "epoch": 65.62,
      "learning_rate": 0.0001,
      "loss": 0.0003,
      "step": 4856
    },
    {
      "epoch": 65.67,
      "learning_rate": 0.0001,
      "loss": 0.0005,
      "step": 4860
    },
    {
      "epoch": 65.72,
      "learning_rate": 0.0001,
      "loss": 0.0006,
      "step": 4864
    },
    {
      "epoch": 65.72,
      "eval_exec": 0.7272727272727273,
      "eval_loss": 0.34306177496910095,
      "eval_runtime": 286.5304,
      "eval_samples_per_second": 3.609,
      "step": 4864
    },
    {
      "epoch": 65.78,
      "learning_rate": 0.0001,
      "loss": 0.0008,
      "step": 4868
    },
    {
      "epoch": 65.83,
      "learning_rate": 0.0001,
      "loss": 0.0006,
      "step": 4872
    },
    {
      "epoch": 65.88,
      "learning_rate": 0.0001,
      "loss": 0.0006,
      "step": 4876
    },
    {
      "epoch": 65.94,
      "learning_rate": 0.0001,
      "loss": 0.0004,
      "step": 4880
    },
    {
      "epoch": 65.99,
      "learning_rate": 0.0001,
      "loss": 0.0005,
      "step": 4884
    },
    {
      "epoch": 66.05,
      "learning_rate": 0.0001,
      "loss": 0.0005,
      "step": 4888
    },
    {
      "epoch": 66.11,
      "learning_rate": 0.0001,
      "loss": 0.0007,
      "step": 4892
    },
    {
      "epoch": 66.16,
      "learning_rate": 0.0001,
      "loss": 0.0007,
      "step": 4896
    },
    {
      "epoch": 66.21,
      "learning_rate": 0.0001,
      "loss": 0.0005,
      "step": 4900
    },
    {
      "epoch": 66.27,
      "learning_rate": 0.0001,
      "loss": 0.0013,
      "step": 4904
    },
    {
      "epoch": 66.32,
      "learning_rate": 0.0001,
      "loss": 0.001,
      "step": 4908
    },
    {
      "epoch": 66.37,
      "learning_rate": 0.0001,
      "loss": 0.0008,
      "step": 4912
    },
    {
      "epoch": 66.43,
      "learning_rate": 0.0001,
      "loss": 0.0004,
      "step": 4916
    },
    {
      "epoch": 66.48,
      "learning_rate": 0.0001,
      "loss": 0.0006,
      "step": 4920
    },
    {
      "epoch": 66.53,
      "learning_rate": 0.0001,
      "loss": 0.0019,
      "step": 4924
    },
    {
      "epoch": 66.59,
      "learning_rate": 0.0001,
      "loss": 0.0013,
      "step": 4928
    },
    {
      "epoch": 66.59,
      "eval_exec": 0.7050290135396519,
      "eval_loss": 0.3411843478679657,
      "eval_runtime": 292.9024,
      "eval_samples_per_second": 3.53,
      "step": 4928
    },
    {
      "epoch": 66.64,
      "learning_rate": 0.0001,
      "loss": 0.0005,
      "step": 4932
    },
    {
      "epoch": 66.7,
      "learning_rate": 0.0001,
      "loss": 0.0003,
      "step": 4936
    },
    {
      "epoch": 66.75,
      "learning_rate": 0.0001,
      "loss": 0.0011,
      "step": 4940
    },
    {
      "epoch": 66.8,
      "learning_rate": 0.0001,
      "loss": 0.0006,
      "step": 4944
    },
    {
      "epoch": 66.86,
      "learning_rate": 0.0001,
      "loss": 0.0006,
      "step": 4948
    },
    {
      "epoch": 66.91,
      "learning_rate": 0.0001,
      "loss": 0.001,
      "step": 4952
    },
    {
      "epoch": 66.96,
      "learning_rate": 0.0001,
      "loss": 0.0002,
      "step": 4956
    },
    {
      "epoch": 67.03,
      "learning_rate": 0.0001,
      "loss": 0.0009,
      "step": 4960
    },
    {
      "epoch": 67.08,
      "learning_rate": 0.0001,
      "loss": 0.0007,
      "step": 4964
    },
    {
      "epoch": 67.13,
      "learning_rate": 0.0001,
      "loss": 0.0002,
      "step": 4968
    },
    {
      "epoch": 67.19,
      "learning_rate": 0.0001,
      "loss": 0.0003,
      "step": 4972
    },
    {
      "epoch": 67.24,
      "learning_rate": 0.0001,
      "loss": 0.0002,
      "step": 4976
    },
    {
      "epoch": 67.29,
      "learning_rate": 0.0001,
      "loss": 0.001,
      "step": 4980
    },
    {
      "epoch": 67.35,
      "learning_rate": 0.0001,
      "loss": 0.0014,
      "step": 4984
    },
    {
      "epoch": 67.4,
      "learning_rate": 0.0001,
      "loss": 0.0003,
      "step": 4988
    },
    {
      "epoch": 67.45,
      "learning_rate": 0.0001,
      "loss": 0.0012,
      "step": 4992
    },
    {
      "epoch": 67.45,
      "eval_exec": 0.7224371373307543,
      "eval_loss": 0.33966416120529175,
      "eval_runtime": 289.6629,
      "eval_samples_per_second": 3.57,
      "step": 4992
    },
    {
      "epoch": 67.51,
      "learning_rate": 0.0001,
      "loss": 0.0002,
      "step": 4996
    },
    {
      "epoch": 67.56,
      "learning_rate": 0.0001,
      "loss": 0.0015,
      "step": 5000
    },
    {
      "epoch": 67.62,
      "learning_rate": 0.0001,
      "loss": 0.0004,
      "step": 5004
    },
    {
      "epoch": 67.67,
      "learning_rate": 0.0001,
      "loss": 0.0016,
      "step": 5008
    },
    {
      "epoch": 67.72,
      "learning_rate": 0.0001,
      "loss": 0.0013,
      "step": 5012
    },
    {
      "epoch": 67.78,
      "learning_rate": 0.0001,
      "loss": 0.0017,
      "step": 5016
    },
    {
      "epoch": 67.83,
      "learning_rate": 0.0001,
      "loss": 0.0003,
      "step": 5020
    },
    {
      "epoch": 67.88,
      "learning_rate": 0.0001,
      "loss": 0.0006,
      "step": 5024
    },
    {
      "epoch": 67.94,
      "learning_rate": 0.0001,
      "loss": 0.0004,
      "step": 5028
    },
    {
      "epoch": 67.99,
      "learning_rate": 0.0001,
      "loss": 0.0011,
      "step": 5032
    },
    {
      "epoch": 68.05,
      "learning_rate": 0.0001,
      "loss": 0.0009,
      "step": 5036
    },
    {
      "epoch": 68.11,
      "learning_rate": 0.0001,
      "loss": 0.0005,
      "step": 5040
    },
    {
      "epoch": 68.16,
      "learning_rate": 0.0001,
      "loss": 0.0006,
      "step": 5044
    },
    {
      "epoch": 68.21,
      "learning_rate": 0.0001,
      "loss": 0.0008,
      "step": 5048
    },
    {
      "epoch": 68.27,
      "learning_rate": 0.0001,
      "loss": 0.0006,
      "step": 5052
    },
    {
      "epoch": 68.32,
      "learning_rate": 0.0001,
      "loss": 0.0007,
      "step": 5056
    },
    {
      "epoch": 68.32,
      "eval_exec": 0.723404255319149,
      "eval_loss": 0.31566286087036133,
      "eval_runtime": 285.389,
      "eval_samples_per_second": 3.623,
      "step": 5056
    },
    {
      "epoch": 68.37,
      "learning_rate": 0.0001,
      "loss": 0.0008,
      "step": 5060
    },
    {
      "epoch": 68.43,
      "learning_rate": 0.0001,
      "loss": 0.0008,
      "step": 5064
    },
    {
      "epoch": 68.48,
      "learning_rate": 0.0001,
      "loss": 0.0009,
      "step": 5068
    },
    {
      "epoch": 68.53,
      "learning_rate": 0.0001,
      "loss": 0.0013,
      "step": 5072
    },
    {
      "epoch": 68.59,
      "learning_rate": 0.0001,
      "loss": 0.0004,
      "step": 5076
    },
    {
      "epoch": 68.64,
      "learning_rate": 0.0001,
      "loss": 0.0011,
      "step": 5080
    },
    {
      "epoch": 68.7,
      "learning_rate": 0.0001,
      "loss": 0.0004,
      "step": 5084
    },
    {
      "epoch": 68.75,
      "learning_rate": 0.0001,
      "loss": 0.0023,
      "step": 5088
    },
    {
      "epoch": 68.8,
      "learning_rate": 0.0001,
      "loss": 0.001,
      "step": 5092
    },
    {
      "epoch": 68.86,
      "learning_rate": 0.0001,
      "loss": 0.0024,
      "step": 5096
    },
    {
      "epoch": 68.91,
      "learning_rate": 0.0001,
      "loss": 0.0011,
      "step": 5100
    },
    {
      "epoch": 68.96,
      "learning_rate": 0.0001,
      "loss": 0.0012,
      "step": 5104
    },
    {
      "epoch": 69.03,
      "learning_rate": 0.0001,
      "loss": 0.001,
      "step": 5108
    },
    {
      "epoch": 69.08,
      "learning_rate": 0.0001,
      "loss": 0.0016,
      "step": 5112
    },
    {
      "epoch": 69.13,
      "learning_rate": 0.0001,
      "loss": 0.0007,
      "step": 5116
    },
    {
      "epoch": 69.19,
      "learning_rate": 0.0001,
      "loss": 0.0009,
      "step": 5120
    },
    {
      "epoch": 69.19,
      "eval_exec": 0.7108317214700194,
      "eval_loss": 0.3202356994152069,
      "eval_runtime": 289.377,
      "eval_samples_per_second": 3.573,
      "step": 5120
    },
    {
      "epoch": 69.24,
      "learning_rate": 0.0001,
      "loss": 0.0003,
      "step": 5124
    },
    {
      "epoch": 69.29,
      "learning_rate": 0.0001,
      "loss": 0.0004,
      "step": 5128
    },
    {
      "epoch": 69.35,
      "learning_rate": 0.0001,
      "loss": 0.0002,
      "step": 5132
    },
    {
      "epoch": 69.4,
      "learning_rate": 0.0001,
      "loss": 0.0002,
      "step": 5136
    },
    {
      "epoch": 69.45,
      "learning_rate": 0.0001,
      "loss": 0.0001,
      "step": 5140
    },
    {
      "epoch": 69.51,
      "learning_rate": 0.0001,
      "loss": 0.0015,
      "step": 5144
    },
    {
      "epoch": 69.56,
      "learning_rate": 0.0001,
      "loss": 0.001,
      "step": 5148
    },
    {
      "epoch": 69.62,
      "learning_rate": 0.0001,
      "loss": 0.0016,
      "step": 5152
    },
    {
      "epoch": 69.67,
      "learning_rate": 0.0001,
      "loss": 0.0006,
      "step": 5156
    },
    {
      "epoch": 69.72,
      "learning_rate": 0.0001,
      "loss": 0.0071,
      "step": 5160
    },
    {
      "epoch": 69.78,
      "learning_rate": 0.0001,
      "loss": 0.0006,
      "step": 5164
    },
    {
      "epoch": 69.83,
      "learning_rate": 0.0001,
      "loss": 0.0003,
      "step": 5168
    },
    {
      "epoch": 69.88,
      "learning_rate": 0.0001,
      "loss": 0.0006,
      "step": 5172
    },
    {
      "epoch": 69.94,
      "learning_rate": 0.0001,
      "loss": 0.0003,
      "step": 5176
    },
    {
      "epoch": 69.99,
      "learning_rate": 0.0001,
      "loss": 0.0032,
      "step": 5180
    },
    {
      "epoch": 70.05,
      "learning_rate": 0.0001,
      "loss": 0.0007,
      "step": 5184
    },
    {
      "epoch": 70.05,
      "eval_exec": 0.7040618955512572,
      "eval_loss": 0.31820809841156006,
      "eval_runtime": 291.8188,
      "eval_samples_per_second": 3.543,
      "step": 5184
    },
    {
      "epoch": 70.11,
      "learning_rate": 0.0001,
      "loss": 0.0002,
      "step": 5188
    },
    {
      "epoch": 70.16,
      "learning_rate": 0.0001,
      "loss": 0.0002,
      "step": 5192
    },
    {
      "epoch": 70.21,
      "learning_rate": 0.0001,
      "loss": 0.0001,
      "step": 5196
    },
    {
      "epoch": 70.27,
      "learning_rate": 0.0001,
      "loss": 0.0006,
      "step": 5200
    },
    {
      "epoch": 70.32,
      "learning_rate": 0.0001,
      "loss": 0.0013,
      "step": 5204
    },
    {
      "epoch": 70.37,
      "learning_rate": 0.0001,
      "loss": 0.0004,
      "step": 5208
    },
    {
      "epoch": 70.43,
      "learning_rate": 0.0001,
      "loss": 0.0002,
      "step": 5212
    },
    {
      "epoch": 70.48,
      "learning_rate": 0.0001,
      "loss": 0.0004,
      "step": 5216
    },
    {
      "epoch": 70.53,
      "learning_rate": 0.0001,
      "loss": 0.0012,
      "step": 5220
    },
    {
      "epoch": 70.59,
      "learning_rate": 0.0001,
      "loss": 0.0002,
      "step": 5224
    },
    {
      "epoch": 70.64,
      "learning_rate": 0.0001,
      "loss": 0.0004,
      "step": 5228
    },
    {
      "epoch": 70.7,
      "learning_rate": 0.0001,
      "loss": 0.0004,
      "step": 5232
    },
    {
      "epoch": 70.75,
      "learning_rate": 0.0001,
      "loss": 0.0017,
      "step": 5236
    },
    {
      "epoch": 70.8,
      "learning_rate": 0.0001,
      "loss": 0.0007,
      "step": 5240
    },
    {
      "epoch": 70.86,
      "learning_rate": 0.0001,
      "loss": 0.0001,
      "step": 5244
    },
    {
      "epoch": 70.91,
      "learning_rate": 0.0001,
      "loss": 0.0004,
      "step": 5248
    },
    {
      "epoch": 70.91,
      "eval_exec": 0.7176015473887815,
      "eval_loss": 0.35949328541755676,
      "eval_runtime": 287.4088,
      "eval_samples_per_second": 3.598,
      "step": 5248
    },
    {
      "epoch": 70.96,
      "learning_rate": 0.0001,
      "loss": 0.0005,
      "step": 5252
    },
    {
      "epoch": 71.03,
      "learning_rate": 0.0001,
      "loss": 0.0002,
      "step": 5256
    },
    {
      "epoch": 71.08,
      "learning_rate": 0.0001,
      "loss": 0.0009,
      "step": 5260
    },
    {
      "epoch": 71.13,
      "learning_rate": 0.0001,
      "loss": 0.0007,
      "step": 5264
    },
    {
      "epoch": 71.19,
      "learning_rate": 0.0001,
      "loss": 0.0002,
      "step": 5268
    },
    {
      "epoch": 71.24,
      "learning_rate": 0.0001,
      "loss": 0.0007,
      "step": 5272
    },
    {
      "epoch": 71.29,
      "learning_rate": 0.0001,
      "loss": 0.0006,
      "step": 5276
    },
    {
      "epoch": 71.35,
      "learning_rate": 0.0001,
      "loss": 0.0013,
      "step": 5280
    },
    {
      "epoch": 71.4,
      "learning_rate": 0.0001,
      "loss": 0.0002,
      "step": 5284
    },
    {
      "epoch": 71.45,
      "learning_rate": 0.0001,
      "loss": 0.0014,
      "step": 5288
    },
    {
      "epoch": 71.51,
      "learning_rate": 0.0001,
      "loss": 0.0007,
      "step": 5292
    },
    {
      "epoch": 71.56,
      "learning_rate": 0.0001,
      "loss": 0.0004,
      "step": 5296
    },
    {
      "epoch": 71.62,
      "learning_rate": 0.0001,
      "loss": 0.005,
      "step": 5300
    },
    {
      "epoch": 71.67,
      "learning_rate": 0.0001,
      "loss": 0.0011,
      "step": 5304
    },
    {
      "epoch": 71.72,
      "learning_rate": 0.0001,
      "loss": 0.0026,
      "step": 5308
    },
    {
      "epoch": 71.78,
      "learning_rate": 0.0001,
      "loss": 0.0009,
      "step": 5312
    },
    {
      "epoch": 71.78,
      "eval_exec": 0.7205029013539652,
      "eval_loss": 0.304627001285553,
      "eval_runtime": 285.1203,
      "eval_samples_per_second": 3.627,
      "step": 5312
    },
    {
      "epoch": 71.83,
      "learning_rate": 0.0001,
      "loss": 0.0005,
      "step": 5316
    },
    {
      "epoch": 71.88,
      "learning_rate": 0.0001,
      "loss": 0.0007,
      "step": 5320
    },
    {
      "epoch": 71.94,
      "learning_rate": 0.0001,
      "loss": 0.0016,
      "step": 5324
    },
    {
      "epoch": 71.99,
      "learning_rate": 0.0001,
      "loss": 0.003,
      "step": 5328
    },
    {
      "epoch": 72.05,
      "learning_rate": 0.0001,
      "loss": 0.0007,
      "step": 5332
    },
    {
      "epoch": 72.11,
      "learning_rate": 0.0001,
      "loss": 0.0011,
      "step": 5336
    },
    {
      "epoch": 72.16,
      "learning_rate": 0.0001,
      "loss": 0.0003,
      "step": 5340
    },
    {
      "epoch": 72.21,
      "learning_rate": 0.0001,
      "loss": 0.0002,
      "step": 5344
    },
    {
      "epoch": 72.27,
      "learning_rate": 0.0001,
      "loss": 0.0015,
      "step": 5348
    },
    {
      "epoch": 72.32,
      "learning_rate": 0.0001,
      "loss": 0.0016,
      "step": 5352
    },
    {
      "epoch": 72.37,
      "learning_rate": 0.0001,
      "loss": 0.0009,
      "step": 5356
    },
    {
      "epoch": 72.43,
      "learning_rate": 0.0001,
      "loss": 0.0004,
      "step": 5360
    },
    {
      "epoch": 72.48,
      "learning_rate": 0.0001,
      "loss": 0.0013,
      "step": 5364
    },
    {
      "epoch": 72.53,
      "learning_rate": 0.0001,
      "loss": 0.0005,
      "step": 5368
    },
    {
      "epoch": 72.59,
      "learning_rate": 0.0001,
      "loss": 0.0004,
      "step": 5372
    },
    {
      "epoch": 72.64,
      "learning_rate": 0.0001,
      "loss": 0.0007,
      "step": 5376
    },
    {
      "epoch": 72.64,
      "eval_exec": 0.7321083172147002,
      "eval_loss": 0.3293524980545044,
      "eval_runtime": 290.4949,
      "eval_samples_per_second": 3.559,
      "step": 5376
    },
    {
      "epoch": 72.7,
      "learning_rate": 0.0001,
      "loss": 0.0003,
      "step": 5380
    },
    {
      "epoch": 72.75,
      "learning_rate": 0.0001,
      "loss": 0.0004,
      "step": 5384
    },
    {
      "epoch": 72.8,
      "learning_rate": 0.0001,
      "loss": 0.0005,
      "step": 5388
    },
    {
      "epoch": 72.86,
      "learning_rate": 0.0001,
      "loss": 0.0012,
      "step": 5392
    },
    {
      "epoch": 72.91,
      "learning_rate": 0.0001,
      "loss": 0.0007,
      "step": 5396
    },
    {
      "epoch": 72.96,
      "learning_rate": 0.0001,
      "loss": 0.001,
      "step": 5400
    },
    {
      "epoch": 73.03,
      "learning_rate": 0.0001,
      "loss": 0.0015,
      "step": 5404
    },
    {
      "epoch": 73.08,
      "learning_rate": 0.0001,
      "loss": 0.0008,
      "step": 5408
    },
    {
      "epoch": 73.13,
      "learning_rate": 0.0001,
      "loss": 0.0008,
      "step": 5412
    },
    {
      "epoch": 73.19,
      "learning_rate": 0.0001,
      "loss": 0.0006,
      "step": 5416
    },
    {
      "epoch": 73.24,
      "learning_rate": 0.0001,
      "loss": 0.0056,
      "step": 5420
    },
    {
      "epoch": 73.29,
      "learning_rate": 0.0001,
      "loss": 0.0005,
      "step": 5424
    },
    {
      "epoch": 73.35,
      "learning_rate": 0.0001,
      "loss": 0.0022,
      "step": 5428
    },
    {
      "epoch": 73.4,
      "learning_rate": 0.0001,
      "loss": 0.0008,
      "step": 5432
    },
    {
      "epoch": 73.45,
      "learning_rate": 0.0001,
      "loss": 0.0003,
      "step": 5436
    },
    {
      "epoch": 73.51,
      "learning_rate": 0.0001,
      "loss": 0.0022,
      "step": 5440
    },
    {
      "epoch": 73.51,
      "eval_exec": 0.7350096711798839,
      "eval_loss": 0.31576061248779297,
      "eval_runtime": 291.4555,
      "eval_samples_per_second": 3.548,
      "step": 5440
    },
    {
      "epoch": 73.56,
      "learning_rate": 0.0001,
      "loss": 0.0005,
      "step": 5444
    },
    {
      "epoch": 73.62,
      "learning_rate": 0.0001,
      "loss": 0.0003,
      "step": 5448
    },
    {
      "epoch": 73.67,
      "learning_rate": 0.0001,
      "loss": 0.0002,
      "step": 5452
    },
    {
      "epoch": 73.72,
      "learning_rate": 0.0001,
      "loss": 0.001,
      "step": 5456
    },
    {
      "epoch": 73.78,
      "learning_rate": 0.0001,
      "loss": 0.0006,
      "step": 5460
    },
    {
      "epoch": 73.83,
      "learning_rate": 0.0001,
      "loss": 0.0009,
      "step": 5464
    },
    {
      "epoch": 73.88,
      "learning_rate": 0.0001,
      "loss": 0.0005,
      "step": 5468
    },
    {
      "epoch": 73.94,
      "learning_rate": 0.0001,
      "loss": 0.0003,
      "step": 5472
    },
    {
      "epoch": 73.99,
      "learning_rate": 0.0001,
      "loss": 0.0004,
      "step": 5476
    },
    {
      "epoch": 74.05,
      "learning_rate": 0.0001,
      "loss": 0.0008,
      "step": 5480
    },
    {
      "epoch": 74.11,
      "learning_rate": 0.0001,
      "loss": 0.0006,
      "step": 5484
    },
    {
      "epoch": 74.16,
      "learning_rate": 0.0001,
      "loss": 0.0024,
      "step": 5488
    },
    {
      "epoch": 74.21,
      "learning_rate": 0.0001,
      "loss": 0.0006,
      "step": 5492
    },
    {
      "epoch": 74.27,
      "learning_rate": 0.0001,
      "loss": 0.0004,
      "step": 5496
    },
    {
      "epoch": 74.32,
      "learning_rate": 0.0001,
      "loss": 0.0002,
      "step": 5500
    },
    {
      "epoch": 74.37,
      "learning_rate": 0.0001,
      "loss": 0.0009,
      "step": 5504
    },
    {
      "epoch": 74.37,
      "eval_exec": 0.7021276595744681,
      "eval_loss": 0.3458137512207031,
      "eval_runtime": 285.7195,
      "eval_samples_per_second": 3.619,
      "step": 5504
    },
    {
      "epoch": 74.43,
      "learning_rate": 0.0001,
      "loss": 0.0013,
      "step": 5508
    },
    {
      "epoch": 74.48,
      "learning_rate": 0.0001,
      "loss": 0.0022,
      "step": 5512
    },
    {
      "epoch": 74.53,
      "learning_rate": 0.0001,
      "loss": 0.0006,
      "step": 5516
    },
    {
      "epoch": 74.59,
      "learning_rate": 0.0001,
      "loss": 0.0013,
      "step": 5520
    },
    {
      "epoch": 74.64,
      "learning_rate": 0.0001,
      "loss": 0.0002,
      "step": 5524
    },
    {
      "epoch": 74.7,
      "learning_rate": 0.0001,
      "loss": 0.0005,
      "step": 5528
    },
    {
      "epoch": 74.75,
      "learning_rate": 0.0001,
      "loss": 0.004,
      "step": 5532
    },
    {
      "epoch": 74.8,
      "learning_rate": 0.0001,
      "loss": 0.0006,
      "step": 5536
    },
    {
      "epoch": 74.86,
      "learning_rate": 0.0001,
      "loss": 0.0017,
      "step": 5540
    },
    {
      "epoch": 74.91,
      "learning_rate": 0.0001,
      "loss": 0.0006,
      "step": 5544
    },
    {
      "epoch": 74.96,
      "learning_rate": 0.0001,
      "loss": 0.003,
      "step": 5548
    },
    {
      "epoch": 75.03,
      "learning_rate": 0.0001,
      "loss": 0.0015,
      "step": 5552
    },
    {
      "epoch": 75.08,
      "learning_rate": 0.0001,
      "loss": 0.0003,
      "step": 5556
    },
    {
      "epoch": 75.13,
      "learning_rate": 0.0001,
      "loss": 0.0005,
      "step": 5560
    },
    {
      "epoch": 75.19,
      "learning_rate": 0.0001,
      "loss": 0.0001,
      "step": 5564
    },
    {
      "epoch": 75.24,
      "learning_rate": 0.0001,
      "loss": 0.0006,
      "step": 5568
    },
    {
      "epoch": 75.24,
      "eval_exec": 0.7108317214700194,
      "eval_loss": 0.3360079526901245,
      "eval_runtime": 289.9247,
      "eval_samples_per_second": 3.566,
      "step": 5568
    },
    {
      "epoch": 75.29,
      "learning_rate": 0.0001,
      "loss": 0.0003,
      "step": 5572
    },
    {
      "epoch": 75.35,
      "learning_rate": 0.0001,
      "loss": 0.0003,
      "step": 5576
    },
    {
      "epoch": 75.4,
      "learning_rate": 0.0001,
      "loss": 0.0003,
      "step": 5580
    },
    {
      "epoch": 75.45,
      "learning_rate": 0.0001,
      "loss": 0.0016,
      "step": 5584
    },
    {
      "epoch": 75.51,
      "learning_rate": 0.0001,
      "loss": 0.0007,
      "step": 5588
    },
    {
      "epoch": 75.56,
      "learning_rate": 0.0001,
      "loss": 0.0014,
      "step": 5592
    },
    {
      "epoch": 75.62,
      "learning_rate": 0.0001,
      "loss": 0.0006,
      "step": 5596
    },
    {
      "epoch": 75.67,
      "learning_rate": 0.0001,
      "loss": 0.0009,
      "step": 5600
    },
    {
      "epoch": 75.72,
      "learning_rate": 0.0001,
      "loss": 0.0004,
      "step": 5604
    },
    {
      "epoch": 75.78,
      "learning_rate": 0.0001,
      "loss": 0.001,
      "step": 5608
    },
    {
      "epoch": 75.83,
      "learning_rate": 0.0001,
      "loss": 0.0015,
      "step": 5612
    },
    {
      "epoch": 75.88,
      "learning_rate": 0.0001,
      "loss": 0.0008,
      "step": 5616
    },
    {
      "epoch": 75.94,
      "learning_rate": 0.0001,
      "loss": 0.0025,
      "step": 5620
    },
    {
      "epoch": 75.99,
      "learning_rate": 0.0001,
      "loss": 0.0009,
      "step": 5624
    },
    {
      "epoch": 76.05,
      "learning_rate": 0.0001,
      "loss": 0.0009,
      "step": 5628
    },
    {
      "epoch": 76.11,
      "learning_rate": 0.0001,
      "loss": 0.001,
      "step": 5632
    },
    {
      "epoch": 76.11,
      "eval_exec": 0.706963249516441,
      "eval_loss": 0.3310911953449249,
      "eval_runtime": 288.5245,
      "eval_samples_per_second": 3.584,
      "step": 5632
    },
    {
      "epoch": 76.16,
      "learning_rate": 0.0001,
      "loss": 0.0007,
      "step": 5636
    },
    {
      "epoch": 76.21,
      "learning_rate": 0.0001,
      "loss": 0.0007,
      "step": 5640
    },
    {
      "epoch": 76.27,
      "learning_rate": 0.0001,
      "loss": 0.0009,
      "step": 5644
    },
    {
      "epoch": 76.32,
      "learning_rate": 0.0001,
      "loss": 0.0008,
      "step": 5648
    },
    {
      "epoch": 76.37,
      "learning_rate": 0.0001,
      "loss": 0.0036,
      "step": 5652
    },
    {
      "epoch": 76.43,
      "learning_rate": 0.0001,
      "loss": 0.0003,
      "step": 5656
    },
    {
      "epoch": 76.48,
      "learning_rate": 0.0001,
      "loss": 0.0017,
      "step": 5660
    },
    {
      "epoch": 76.53,
      "learning_rate": 0.0001,
      "loss": 0.0017,
      "step": 5664
    },
    {
      "epoch": 76.59,
      "learning_rate": 0.0001,
      "loss": 0.0006,
      "step": 5668
    },
    {
      "epoch": 76.64,
      "learning_rate": 0.0001,
      "loss": 0.0008,
      "step": 5672
    },
    {
      "epoch": 76.7,
      "learning_rate": 0.0001,
      "loss": 0.0014,
      "step": 5676
    },
    {
      "epoch": 76.75,
      "learning_rate": 0.0001,
      "loss": 0.0005,
      "step": 5680
    },
    {
      "epoch": 76.8,
      "learning_rate": 0.0001,
      "loss": 0.0005,
      "step": 5684
    },
    {
      "epoch": 76.86,
      "learning_rate": 0.0001,
      "loss": 0.0012,
      "step": 5688
    },
    {
      "epoch": 76.91,
      "learning_rate": 0.0001,
      "loss": 0.0008,
      "step": 5692
    },
    {
      "epoch": 76.96,
      "learning_rate": 0.0001,
      "loss": 0.0008,
      "step": 5696
    },
    {
      "epoch": 76.96,
      "eval_exec": 0.706963249516441,
      "eval_loss": 0.3491285741329193,
      "eval_runtime": 290.4295,
      "eval_samples_per_second": 3.56,
      "step": 5696
    },
    {
      "epoch": 77.03,
      "learning_rate": 0.0001,
      "loss": 0.0008,
      "step": 5700
    },
    {
      "epoch": 77.08,
      "learning_rate": 0.0001,
      "loss": 0.0004,
      "step": 5704
    },
    {
      "epoch": 77.13,
      "learning_rate": 0.0001,
      "loss": 0.0014,
      "step": 5708
    },
    {
      "epoch": 77.19,
      "learning_rate": 0.0001,
      "loss": 0.0009,
      "step": 5712
    },
    {
      "epoch": 77.24,
      "learning_rate": 0.0001,
      "loss": 0.0005,
      "step": 5716
    },
    {
      "epoch": 77.29,
      "learning_rate": 0.0001,
      "loss": 0.0009,
      "step": 5720
    },
    {
      "epoch": 77.35,
      "learning_rate": 0.0001,
      "loss": 0.0009,
      "step": 5724
    },
    {
      "epoch": 77.4,
      "learning_rate": 0.0001,
      "loss": 0.0018,
      "step": 5728
    },
    {
      "epoch": 77.45,
      "learning_rate": 0.0001,
      "loss": 0.0002,
      "step": 5732
    },
    {
      "epoch": 77.51,
      "learning_rate": 0.0001,
      "loss": 0.0004,
      "step": 5736
    },
    {
      "epoch": 77.56,
      "learning_rate": 0.0001,
      "loss": 0.0014,
      "step": 5740
    },
    {
      "epoch": 77.62,
      "learning_rate": 0.0001,
      "loss": 0.0004,
      "step": 5744
    },
    {
      "epoch": 77.67,
      "learning_rate": 0.0001,
      "loss": 0.0003,
      "step": 5748
    },
    {
      "epoch": 77.72,
      "learning_rate": 0.0001,
      "loss": 0.0008,
      "step": 5752
    },
    {
      "epoch": 77.78,
      "learning_rate": 0.0001,
      "loss": 0.0007,
      "step": 5756
    },
    {
      "epoch": 77.83,
      "learning_rate": 0.0001,
      "loss": 0.0005,
      "step": 5760
    },
    {
      "epoch": 77.83,
      "eval_exec": 0.7108317214700194,
      "eval_loss": 0.33614587783813477,
      "eval_runtime": 284.0797,
      "eval_samples_per_second": 3.64,
      "step": 5760
    },
    {
      "epoch": 77.88,
      "learning_rate": 0.0001,
      "loss": 0.0007,
      "step": 5764
    },
    {
      "epoch": 77.94,
      "learning_rate": 0.0001,
      "loss": 0.0011,
      "step": 5768
    },
    {
      "epoch": 77.99,
      "learning_rate": 0.0001,
      "loss": 0.0003,
      "step": 5772
    },
    {
      "epoch": 78.05,
      "learning_rate": 0.0001,
      "loss": 0.0004,
      "step": 5776
    },
    {
      "epoch": 78.11,
      "learning_rate": 0.0001,
      "loss": 0.0007,
      "step": 5780
    },
    {
      "epoch": 78.16,
      "learning_rate": 0.0001,
      "loss": 0.0005,
      "step": 5784
    },
    {
      "epoch": 78.21,
      "learning_rate": 0.0001,
      "loss": 0.0005,
      "step": 5788
    },
    {
      "epoch": 78.27,
      "learning_rate": 0.0001,
      "loss": 0.0013,
      "step": 5792
    },
    {
      "epoch": 78.32,
      "learning_rate": 0.0001,
      "loss": 0.0003,
      "step": 5796
    },
    {
      "epoch": 78.37,
      "learning_rate": 0.0001,
      "loss": 0.0008,
      "step": 5800
    },
    {
      "epoch": 78.43,
      "learning_rate": 0.0001,
      "loss": 0.0006,
      "step": 5804
    },
    {
      "epoch": 78.48,
      "learning_rate": 0.0001,
      "loss": 0.0003,
      "step": 5808
    },
    {
      "epoch": 78.53,
      "learning_rate": 0.0001,
      "loss": 0.0004,
      "step": 5812
    },
    {
      "epoch": 78.59,
      "learning_rate": 0.0001,
      "loss": 0.001,
      "step": 5816
    },
    {
      "epoch": 78.64,
      "learning_rate": 0.0001,
      "loss": 0.0016,
      "step": 5820
    },
    {
      "epoch": 78.7,
      "learning_rate": 0.0001,
      "loss": 0.0002,
      "step": 5824
    },
    {
      "epoch": 78.7,
      "eval_exec": 0.7117988394584139,
      "eval_loss": 0.3427754342556,
      "eval_runtime": 290.913,
      "eval_samples_per_second": 3.554,
      "step": 5824
    },
    {
      "epoch": 78.75,
      "learning_rate": 0.0001,
      "loss": 0.0002,
      "step": 5828
    },
    {
      "epoch": 78.8,
      "learning_rate": 0.0001,
      "loss": 0.0028,
      "step": 5832
    },
    {
      "epoch": 78.86,
      "learning_rate": 0.0001,
      "loss": 0.0006,
      "step": 5836
    },
    {
      "epoch": 78.91,
      "learning_rate": 0.0001,
      "loss": 0.0003,
      "step": 5840
    },
    {
      "epoch": 78.96,
      "learning_rate": 0.0001,
      "loss": 0.0008,
      "step": 5844
    },
    {
      "epoch": 79.03,
      "learning_rate": 0.0001,
      "loss": 0.0007,
      "step": 5848
    },
    {
      "epoch": 79.08,
      "learning_rate": 0.0001,
      "loss": 0.0005,
      "step": 5852
    },
    {
      "epoch": 79.13,
      "learning_rate": 0.0001,
      "loss": 0.0003,
      "step": 5856
    },
    {
      "epoch": 79.19,
      "learning_rate": 0.0001,
      "loss": 0.0003,
      "step": 5860
    },
    {
      "epoch": 79.24,
      "learning_rate": 0.0001,
      "loss": 0.0006,
      "step": 5864
    },
    {
      "epoch": 79.29,
      "learning_rate": 0.0001,
      "loss": 0.0005,
      "step": 5868
    },
    {
      "epoch": 79.35,
      "learning_rate": 0.0001,
      "loss": 0.0002,
      "step": 5872
    },
    {
      "epoch": 79.4,
      "learning_rate": 0.0001,
      "loss": 0.0006,
      "step": 5876
    },
    {
      "epoch": 79.45,
      "learning_rate": 0.0001,
      "loss": 0.0004,
      "step": 5880
    },
    {
      "epoch": 79.51,
      "learning_rate": 0.0001,
      "loss": 0.0003,
      "step": 5884
    },
    {
      "epoch": 79.56,
      "learning_rate": 0.0001,
      "loss": 0.0007,
      "step": 5888
    },
    {
      "epoch": 79.56,
      "eval_exec": 0.7011605415860735,
      "eval_loss": 0.34577926993370056,
      "eval_runtime": 287.6027,
      "eval_samples_per_second": 3.595,
      "step": 5888
    },
    {
      "epoch": 79.62,
      "learning_rate": 0.0001,
      "loss": 0.0004,
      "step": 5892
    },
    {
      "epoch": 79.67,
      "learning_rate": 0.0001,
      "loss": 0.0003,
      "step": 5896
    },
    {
      "epoch": 79.72,
      "learning_rate": 0.0001,
      "loss": 0.0004,
      "step": 5900
    },
    {
      "epoch": 79.78,
      "learning_rate": 0.0001,
      "loss": 0.0007,
      "step": 5904
    },
    {
      "epoch": 79.83,
      "learning_rate": 0.0001,
      "loss": 0.0008,
      "step": 5908
    },
    {
      "epoch": 79.88,
      "learning_rate": 0.0001,
      "loss": 0.0008,
      "step": 5912
    },
    {
      "epoch": 79.94,
      "learning_rate": 0.0001,
      "loss": 0.001,
      "step": 5916
    },
    {
      "epoch": 79.99,
      "learning_rate": 0.0001,
      "loss": 0.0008,
      "step": 5920
    },
    {
      "epoch": 80.05,
      "learning_rate": 0.0001,
      "loss": 0.0008,
      "step": 5924
    },
    {
      "epoch": 80.11,
      "learning_rate": 0.0001,
      "loss": 0.0006,
      "step": 5928
    },
    {
      "epoch": 80.16,
      "learning_rate": 0.0001,
      "loss": 0.0004,
      "step": 5932
    },
    {
      "epoch": 80.21,
      "learning_rate": 0.0001,
      "loss": 0.0007,
      "step": 5936
    },
    {
      "epoch": 80.27,
      "learning_rate": 0.0001,
      "loss": 0.0017,
      "step": 5940
    },
    {
      "epoch": 80.32,
      "learning_rate": 0.0001,
      "loss": 0.0006,
      "step": 5944
    },
    {
      "epoch": 80.37,
      "learning_rate": 0.0001,
      "loss": 0.0004,
      "step": 5948
    },
    {
      "epoch": 80.43,
      "learning_rate": 0.0001,
      "loss": 0.0011,
      "step": 5952
    },
    {
      "epoch": 80.43,
      "eval_exec": 0.7050290135396519,
      "eval_loss": 0.32949239015579224,
      "eval_runtime": 280.7542,
      "eval_samples_per_second": 3.683,
      "step": 5952
    },
    {
      "epoch": 80.48,
      "learning_rate": 0.0001,
      "loss": 0.0003,
      "step": 5956
    },
    {
      "epoch": 80.53,
      "learning_rate": 0.0001,
      "loss": 0.0007,
      "step": 5960
    },
    {
      "epoch": 80.59,
      "learning_rate": 0.0001,
      "loss": 0.0004,
      "step": 5964
    },
    {
      "epoch": 80.64,
      "learning_rate": 0.0001,
      "loss": 0.0011,
      "step": 5968
    },
    {
      "epoch": 80.7,
      "learning_rate": 0.0001,
      "loss": 0.0007,
      "step": 5972
    },
    {
      "epoch": 80.75,
      "learning_rate": 0.0001,
      "loss": 0.0003,
      "step": 5976
    },
    {
      "epoch": 80.8,
      "learning_rate": 0.0001,
      "loss": 0.0004,
      "step": 5980
    },
    {
      "epoch": 80.86,
      "learning_rate": 0.0001,
      "loss": 0.0003,
      "step": 5984
    },
    {
      "epoch": 80.91,
      "learning_rate": 0.0001,
      "loss": 0.0001,
      "step": 5988
    },
    {
      "epoch": 80.96,
      "learning_rate": 0.0001,
      "loss": 0.0008,
      "step": 5992
    },
    {
      "epoch": 81.03,
      "learning_rate": 0.0001,
      "loss": 0.0003,
      "step": 5996
    },
    {
      "epoch": 81.08,
      "learning_rate": 0.0001,
      "loss": 0.0004,
      "step": 6000
    },
    {
      "epoch": 81.13,
      "learning_rate": 0.0001,
      "loss": 0.0002,
      "step": 6004
    },
    {
      "epoch": 81.19,
      "learning_rate": 0.0001,
      "loss": 0.0001,
      "step": 6008
    },
    {
      "epoch": 81.24,
      "learning_rate": 0.0001,
      "loss": 0.0012,
      "step": 6012
    },
    {
      "epoch": 81.29,
      "learning_rate": 0.0001,
      "loss": 0.0004,
      "step": 6016
    },
    {
      "epoch": 81.29,
      "eval_exec": 0.7040618955512572,
      "eval_loss": 0.3443618714809418,
      "eval_runtime": 284.4305,
      "eval_samples_per_second": 3.635,
      "step": 6016
    },
    {
      "epoch": 81.35,
      "learning_rate": 0.0001,
      "loss": 0.0004,
      "step": 6020
    },
    {
      "epoch": 81.4,
      "learning_rate": 0.0001,
      "loss": 0.0002,
      "step": 6024
    },
    {
      "epoch": 81.45,
      "learning_rate": 0.0001,
      "loss": 0.0002,
      "step": 6028
    },
    {
      "epoch": 81.51,
      "learning_rate": 0.0001,
      "loss": 0.0001,
      "step": 6032
    },
    {
      "epoch": 81.56,
      "learning_rate": 0.0001,
      "loss": 0.0016,
      "step": 6036
    },
    {
      "epoch": 81.62,
      "learning_rate": 0.0001,
      "loss": 0.0001,
      "step": 6040
    },
    {
      "epoch": 81.67,
      "learning_rate": 0.0001,
      "loss": 0.0003,
      "step": 6044
    },
    {
      "epoch": 81.72,
      "learning_rate": 0.0001,
      "loss": 0.0001,
      "step": 6048
    },
    {
      "epoch": 81.78,
      "learning_rate": 0.0001,
      "loss": 0.0003,
      "step": 6052
    },
    {
      "epoch": 81.83,
      "learning_rate": 0.0001,
      "loss": 0.001,
      "step": 6056
    },
    {
      "epoch": 81.88,
      "learning_rate": 0.0001,
      "loss": 0.0008,
      "step": 6060
    },
    {
      "epoch": 81.94,
      "learning_rate": 0.0001,
      "loss": 0.0005,
      "step": 6064
    },
    {
      "epoch": 81.99,
      "learning_rate": 0.0001,
      "loss": 0.0005,
      "step": 6068
    },
    {
      "epoch": 82.05,
      "learning_rate": 0.0001,
      "loss": 0.0002,
      "step": 6072
    },
    {
      "epoch": 82.11,
      "learning_rate": 0.0001,
      "loss": 0.0004,
      "step": 6076
    },
    {
      "epoch": 82.16,
      "learning_rate": 0.0001,
      "loss": 0.0001,
      "step": 6080
    },
    {
      "epoch": 82.16,
      "eval_exec": 0.7253384912959381,
      "eval_loss": 0.35400378704071045,
      "eval_runtime": 287.759,
      "eval_samples_per_second": 3.593,
      "step": 6080
    },
    {
      "epoch": 82.21,
      "learning_rate": 0.0001,
      "loss": 0.0003,
      "step": 6084
    },
    {
      "epoch": 82.27,
      "learning_rate": 0.0001,
      "loss": 0.001,
      "step": 6088
    },
    {
      "epoch": 82.32,
      "learning_rate": 0.0001,
      "loss": 0.0011,
      "step": 6092
    },
    {
      "epoch": 82.37,
      "learning_rate": 0.0001,
      "loss": 0.0003,
      "step": 6096
    },
    {
      "epoch": 82.43,
      "learning_rate": 0.0001,
      "loss": 0.0002,
      "step": 6100
    },
    {
      "epoch": 82.48,
      "learning_rate": 0.0001,
      "loss": 0.0018,
      "step": 6104
    },
    {
      "epoch": 82.53,
      "learning_rate": 0.0001,
      "loss": 0.0004,
      "step": 6108
    },
    {
      "epoch": 82.59,
      "learning_rate": 0.0001,
      "loss": 0.0007,
      "step": 6112
    },
    {
      "epoch": 82.64,
      "learning_rate": 0.0001,
      "loss": 0.001,
      "step": 6116
    },
    {
      "epoch": 82.7,
      "learning_rate": 0.0001,
      "loss": 0.0002,
      "step": 6120
    },
    {
      "epoch": 82.75,
      "learning_rate": 0.0001,
      "loss": 0.0012,
      "step": 6124
    },
    {
      "epoch": 82.8,
      "learning_rate": 0.0001,
      "loss": 0.0005,
      "step": 6128
    },
    {
      "epoch": 82.86,
      "learning_rate": 0.0001,
      "loss": 0.0005,
      "step": 6132
    },
    {
      "epoch": 82.91,
      "learning_rate": 0.0001,
      "loss": 0.0002,
      "step": 6136
    },
    {
      "epoch": 82.96,
      "learning_rate": 0.0001,
      "loss": 0.0005,
      "step": 6140
    },
    {
      "epoch": 83.03,
      "learning_rate": 0.0001,
      "loss": 0.0001,
      "step": 6144
    },
    {
      "epoch": 83.03,
      "eval_exec": 0.690522243713733,
      "eval_loss": 0.3448183834552765,
      "eval_runtime": 284.9854,
      "eval_samples_per_second": 3.628,
      "step": 6144
    },
    {
      "epoch": 83.08,
      "learning_rate": 0.0001,
      "loss": 0.0004,
      "step": 6148
    },
    {
      "epoch": 83.13,
      "learning_rate": 0.0001,
      "loss": 0.0001,
      "step": 6152
    },
    {
      "epoch": 83.19,
      "learning_rate": 0.0001,
      "loss": 0.0005,
      "step": 6156
    },
    {
      "epoch": 83.24,
      "learning_rate": 0.0001,
      "loss": 0.0006,
      "step": 6160
    },
    {
      "epoch": 83.29,
      "learning_rate": 0.0001,
      "loss": 0.0001,
      "step": 6164
    },
    {
      "epoch": 83.35,
      "learning_rate": 0.0001,
      "loss": 0.0006,
      "step": 6168
    },
    {
      "epoch": 83.4,
      "learning_rate": 0.0001,
      "loss": 0.0002,
      "step": 6172
    },
    {
      "epoch": 83.45,
      "learning_rate": 0.0001,
      "loss": 0.0005,
      "step": 6176
    },
    {
      "epoch": 83.51,
      "learning_rate": 0.0001,
      "loss": 0.0006,
      "step": 6180
    },
    {
      "epoch": 83.56,
      "learning_rate": 0.0001,
      "loss": 0.0008,
      "step": 6184
    },
    {
      "epoch": 83.62,
      "learning_rate": 0.0001,
      "loss": 0.0005,
      "step": 6188
    },
    {
      "epoch": 83.67,
      "learning_rate": 0.0001,
      "loss": 0.0,
      "step": 6192
    },
    {
      "epoch": 83.72,
      "learning_rate": 0.0001,
      "loss": 0.0002,
      "step": 6196
    },
    {
      "epoch": 83.78,
      "learning_rate": 0.0001,
      "loss": 0.0006,
      "step": 6200
    },
    {
      "epoch": 83.83,
      "learning_rate": 0.0001,
      "loss": 0.0003,
      "step": 6204
    },
    {
      "epoch": 83.88,
      "learning_rate": 0.0001,
      "loss": 0.0002,
      "step": 6208
    },
    {
      "epoch": 83.88,
      "eval_exec": 0.7156673114119922,
      "eval_loss": 0.3485656976699829,
      "eval_runtime": 290.1762,
      "eval_samples_per_second": 3.563,
      "step": 6208
    },
    {
      "epoch": 83.94,
      "learning_rate": 0.0001,
      "loss": 0.0012,
      "step": 6212
    },
    {
      "epoch": 83.99,
      "learning_rate": 0.0001,
      "loss": 0.0005,
      "step": 6216
    },
    {
      "epoch": 84.05,
      "learning_rate": 0.0001,
      "loss": 0.0005,
      "step": 6220
    },
    {
      "epoch": 84.11,
      "learning_rate": 0.0001,
      "loss": 0.0022,
      "step": 6224
    },
    {
      "epoch": 84.16,
      "learning_rate": 0.0001,
      "loss": 0.0002,
      "step": 6228
    },
    {
      "epoch": 84.21,
      "learning_rate": 0.0001,
      "loss": 0.0011,
      "step": 6232
    },
    {
      "epoch": 84.27,
      "learning_rate": 0.0001,
      "loss": 0.0005,
      "step": 6236
    },
    {
      "epoch": 84.32,
      "learning_rate": 0.0001,
      "loss": 0.001,
      "step": 6240
    },
    {
      "epoch": 84.37,
      "learning_rate": 0.0001,
      "loss": 0.0011,
      "step": 6244
    },
    {
      "epoch": 84.43,
      "learning_rate": 0.0001,
      "loss": 0.0008,
      "step": 6248
    },
    {
      "epoch": 84.48,
      "learning_rate": 0.0001,
      "loss": 0.0029,
      "step": 6252
    },
    {
      "epoch": 84.53,
      "learning_rate": 0.0001,
      "loss": 0.0004,
      "step": 6256
    },
    {
      "epoch": 84.59,
      "learning_rate": 0.0001,
      "loss": 0.0007,
      "step": 6260
    },
    {
      "epoch": 84.64,
      "learning_rate": 0.0001,
      "loss": 0.0007,
      "step": 6264
    },
    {
      "epoch": 84.7,
      "learning_rate": 0.0001,
      "loss": 0.0003,
      "step": 6268
    },
    {
      "epoch": 84.75,
      "learning_rate": 0.0001,
      "loss": 0.0002,
      "step": 6272
    },
    {
      "epoch": 84.75,
      "eval_exec": 0.7243713733075435,
      "eval_loss": 0.33773335814476013,
      "eval_runtime": 286.9564,
      "eval_samples_per_second": 3.603,
      "step": 6272
    },
    {
      "epoch": 84.8,
      "learning_rate": 0.0001,
      "loss": 0.0004,
      "step": 6276
    },
    {
      "epoch": 84.86,
      "learning_rate": 0.0001,
      "loss": 0.0003,
      "step": 6280
    },
    {
      "epoch": 84.91,
      "learning_rate": 0.0001,
      "loss": 0.0007,
      "step": 6284
    },
    {
      "epoch": 84.96,
      "learning_rate": 0.0001,
      "loss": 0.001,
      "step": 6288
    },
    {
      "epoch": 85.03,
      "learning_rate": 0.0001,
      "loss": 0.0005,
      "step": 6292
    },
    {
      "epoch": 85.08,
      "learning_rate": 0.0001,
      "loss": 0.0015,
      "step": 6296
    },
    {
      "epoch": 85.13,
      "learning_rate": 0.0001,
      "loss": 0.0002,
      "step": 6300
    },
    {
      "epoch": 85.19,
      "learning_rate": 0.0001,
      "loss": 0.0006,
      "step": 6304
    },
    {
      "epoch": 85.24,
      "learning_rate": 0.0001,
      "loss": 0.0003,
      "step": 6308
    },
    {
      "epoch": 85.29,
      "learning_rate": 0.0001,
      "loss": 0.0001,
      "step": 6312
    },
    {
      "epoch": 85.35,
      "learning_rate": 0.0001,
      "loss": 0.0014,
      "step": 6316
    },
    {
      "epoch": 85.4,
      "learning_rate": 0.0001,
      "loss": 0.0003,
      "step": 6320
    },
    {
      "epoch": 85.45,
      "learning_rate": 0.0001,
      "loss": 0.0008,
      "step": 6324
    },
    {
      "epoch": 85.51,
      "learning_rate": 0.0001,
      "loss": 0.0003,
      "step": 6328
    },
    {
      "epoch": 85.56,
      "learning_rate": 0.0001,
      "loss": 0.0004,
      "step": 6332
    },
    {
      "epoch": 85.62,
      "learning_rate": 0.0001,
      "loss": 0.0009,
      "step": 6336
    },
    {
      "epoch": 85.62,
      "eval_exec": 0.7243713733075435,
      "eval_loss": 0.33152279257774353,
      "eval_runtime": 290.7991,
      "eval_samples_per_second": 3.556,
      "step": 6336
    },
    {
      "epoch": 85.67,
      "learning_rate": 0.0001,
      "loss": 0.0003,
      "step": 6340
    },
    {
      "epoch": 85.72,
      "learning_rate": 0.0001,
      "loss": 0.0016,
      "step": 6344
    },
    {
      "epoch": 85.78,
      "learning_rate": 0.0001,
      "loss": 0.0027,
      "step": 6348
    },
    {
      "epoch": 85.83,
      "learning_rate": 0.0001,
      "loss": 0.0013,
      "step": 6352
    },
    {
      "epoch": 85.88,
      "learning_rate": 0.0001,
      "loss": 0.0004,
      "step": 6356
    },
    {
      "epoch": 85.94,
      "learning_rate": 0.0001,
      "loss": 0.001,
      "step": 6360
    },
    {
      "epoch": 85.99,
      "learning_rate": 0.0001,
      "loss": 0.0005,
      "step": 6364
    },
    {
      "epoch": 86.05,
      "learning_rate": 0.0001,
      "loss": 0.0005,
      "step": 6368
    },
    {
      "epoch": 86.11,
      "learning_rate": 0.0001,
      "loss": 0.0004,
      "step": 6372
    },
    {
      "epoch": 86.16,
      "learning_rate": 0.0001,
      "loss": 0.0001,
      "step": 6376
    },
    {
      "epoch": 86.21,
      "learning_rate": 0.0001,
      "loss": 0.0005,
      "step": 6380
    },
    {
      "epoch": 86.27,
      "learning_rate": 0.0001,
      "loss": 0.0004,
      "step": 6384
    },
    {
      "epoch": 86.32,
      "learning_rate": 0.0001,
      "loss": 0.0009,
      "step": 6388
    },
    {
      "epoch": 86.37,
      "learning_rate": 0.0001,
      "loss": 0.0019,
      "step": 6392
    },
    {
      "epoch": 86.43,
      "learning_rate": 0.0001,
      "loss": 0.0004,
      "step": 6396
    },
    {
      "epoch": 86.48,
      "learning_rate": 0.0001,
      "loss": 0.0004,
      "step": 6400
    },
    {
      "epoch": 86.48,
      "eval_exec": 0.7147001934235977,
      "eval_loss": 0.33703288435935974,
      "eval_runtime": 284.2076,
      "eval_samples_per_second": 3.638,
      "step": 6400
    },
    {
      "epoch": 86.53,
      "learning_rate": 0.0001,
      "loss": 0.0002,
      "step": 6404
    },
    {
      "epoch": 86.59,
      "learning_rate": 0.0001,
      "loss": 0.0002,
      "step": 6408
    },
    {
      "epoch": 86.64,
      "learning_rate": 0.0001,
      "loss": 0.0007,
      "step": 6412
    },
    {
      "epoch": 86.7,
      "learning_rate": 0.0001,
      "loss": 0.0003,
      "step": 6416
    },
    {
      "epoch": 86.75,
      "learning_rate": 0.0001,
      "loss": 0.0005,
      "step": 6420
    },
    {
      "epoch": 86.8,
      "learning_rate": 0.0001,
      "loss": 0.0004,
      "step": 6424
    },
    {
      "epoch": 86.86,
      "learning_rate": 0.0001,
      "loss": 0.0004,
      "step": 6428
    },
    {
      "epoch": 86.91,
      "learning_rate": 0.0001,
      "loss": 0.0001,
      "step": 6432
    },
    {
      "epoch": 86.96,
      "learning_rate": 0.0001,
      "loss": 0.0012,
      "step": 6436
    },
    {
      "epoch": 87.03,
      "learning_rate": 0.0001,
      "loss": 0.0002,
      "step": 6440
    },
    {
      "epoch": 87.08,
      "learning_rate": 0.0001,
      "loss": 0.0007,
      "step": 6444
    },
    {
      "epoch": 87.13,
      "learning_rate": 0.0001,
      "loss": 0.0003,
      "step": 6448
    },
    {
      "epoch": 87.19,
      "learning_rate": 0.0001,
      "loss": 0.0003,
      "step": 6452
    },
    {
      "epoch": 87.24,
      "learning_rate": 0.0001,
      "loss": 0.0007,
      "step": 6456
    },
    {
      "epoch": 87.29,
      "learning_rate": 0.0001,
      "loss": 0.0014,
      "step": 6460
    },
    {
      "epoch": 87.35,
      "learning_rate": 0.0001,
      "loss": 0.0005,
      "step": 6464
    },
    {
      "epoch": 87.35,
      "eval_exec": 0.7253384912959381,
      "eval_loss": 0.3449646830558777,
      "eval_runtime": 294.4474,
      "eval_samples_per_second": 3.512,
      "step": 6464
    },
    {
      "epoch": 87.4,
      "learning_rate": 0.0001,
      "loss": 0.0005,
      "step": 6468
    },
    {
      "epoch": 87.45,
      "learning_rate": 0.0001,
      "loss": 0.0009,
      "step": 6472
    },
    {
      "epoch": 87.51,
      "learning_rate": 0.0001,
      "loss": 0.0011,
      "step": 6476
    },
    {
      "epoch": 87.56,
      "learning_rate": 0.0001,
      "loss": 0.0011,
      "step": 6480
    },
    {
      "epoch": 87.62,
      "learning_rate": 0.0001,
      "loss": 0.0004,
      "step": 6484
    },
    {
      "epoch": 87.67,
      "learning_rate": 0.0001,
      "loss": 0.0003,
      "step": 6488
    },
    {
      "epoch": 87.72,
      "learning_rate": 0.0001,
      "loss": 0.0003,
      "step": 6492
    },
    {
      "epoch": 87.78,
      "learning_rate": 0.0001,
      "loss": 0.001,
      "step": 6496
    },
    {
      "epoch": 87.83,
      "learning_rate": 0.0001,
      "loss": 0.0007,
      "step": 6500
    },
    {
      "epoch": 87.88,
      "learning_rate": 0.0001,
      "loss": 0.0006,
      "step": 6504
    },
    {
      "epoch": 87.94,
      "learning_rate": 0.0001,
      "loss": 0.0001,
      "step": 6508
    },
    {
      "epoch": 87.99,
      "learning_rate": 0.0001,
      "loss": 0.0017,
      "step": 6512
    },
    {
      "epoch": 88.05,
      "learning_rate": 0.0001,
      "loss": 0.0012,
      "step": 6516
    },
    {
      "epoch": 88.11,
      "learning_rate": 0.0001,
      "loss": 0.0004,
      "step": 6520
    },
    {
      "epoch": 88.16,
      "learning_rate": 0.0001,
      "loss": 0.0006,
      "step": 6524
    },
    {
      "epoch": 88.21,
      "learning_rate": 0.0001,
      "loss": 0.0006,
      "step": 6528
    },
    {
      "epoch": 88.21,
      "eval_exec": 0.7050290135396519,
      "eval_loss": 0.3305092751979828,
      "eval_runtime": 286.599,
      "eval_samples_per_second": 3.608,
      "step": 6528
    },
    {
      "epoch": 88.27,
      "learning_rate": 0.0001,
      "loss": 0.0033,
      "step": 6532
    },
    {
      "epoch": 88.32,
      "learning_rate": 0.0001,
      "loss": 0.0004,
      "step": 6536
    },
    {
      "epoch": 88.37,
      "learning_rate": 0.0001,
      "loss": 0.0009,
      "step": 6540
    },
    {
      "epoch": 88.43,
      "learning_rate": 0.0001,
      "loss": 0.0013,
      "step": 6544
    },
    {
      "epoch": 88.48,
      "learning_rate": 0.0001,
      "loss": 0.0014,
      "step": 6548
    },
    {
      "epoch": 88.53,
      "learning_rate": 0.0001,
      "loss": 0.0011,
      "step": 6552
    },
    {
      "epoch": 88.59,
      "learning_rate": 0.0001,
      "loss": 0.0005,
      "step": 6556
    },
    {
      "epoch": 88.64,
      "learning_rate": 0.0001,
      "loss": 0.0006,
      "step": 6560
    },
    {
      "epoch": 88.7,
      "learning_rate": 0.0001,
      "loss": 0.0003,
      "step": 6564
    },
    {
      "epoch": 88.75,
      "learning_rate": 0.0001,
      "loss": 0.0008,
      "step": 6568
    },
    {
      "epoch": 88.8,
      "learning_rate": 0.0001,
      "loss": 0.0002,
      "step": 6572
    },
    {
      "epoch": 88.86,
      "learning_rate": 0.0001,
      "loss": 0.0012,
      "step": 6576
    },
    {
      "epoch": 88.91,
      "learning_rate": 0.0001,
      "loss": 0.0001,
      "step": 6580
    },
    {
      "epoch": 88.96,
      "learning_rate": 0.0001,
      "loss": 0.0012,
      "step": 6584
    },
    {
      "epoch": 89.03,
      "learning_rate": 0.0001,
      "loss": 0.0008,
      "step": 6588
    },
    {
      "epoch": 89.08,
      "learning_rate": 0.0001,
      "loss": 0.0007,
      "step": 6592
    },
    {
      "epoch": 89.08,
      "eval_exec": 0.7166344294003868,
      "eval_loss": 0.3231656849384308,
      "eval_runtime": 284.3494,
      "eval_samples_per_second": 3.636,
      "step": 6592
    },
    {
      "epoch": 89.13,
      "learning_rate": 0.0001,
      "loss": 0.0002,
      "step": 6596
    },
    {
      "epoch": 89.19,
      "learning_rate": 0.0001,
      "loss": 0.0001,
      "step": 6600
    },
    {
      "epoch": 89.24,
      "learning_rate": 0.0001,
      "loss": 0.0001,
      "step": 6604
    },
    {
      "epoch": 89.29,
      "learning_rate": 0.0001,
      "loss": 0.0008,
      "step": 6608
    },
    {
      "epoch": 89.35,
      "learning_rate": 0.0001,
      "loss": 0.0006,
      "step": 6612
    },
    {
      "epoch": 89.4,
      "learning_rate": 0.0001,
      "loss": 0.0001,
      "step": 6616
    },
    {
      "epoch": 89.45,
      "learning_rate": 0.0001,
      "loss": 0.0002,
      "step": 6620
    },
    {
      "epoch": 89.51,
      "learning_rate": 0.0001,
      "loss": 0.001,
      "step": 6624
    },
    {
      "epoch": 89.56,
      "learning_rate": 0.0001,
      "loss": 0.0007,
      "step": 6628
    },
    {
      "epoch": 89.62,
      "learning_rate": 0.0001,
      "loss": 0.0005,
      "step": 6632
    },
    {
      "epoch": 89.67,
      "learning_rate": 0.0001,
      "loss": 0.0001,
      "step": 6636
    },
    {
      "epoch": 89.72,
      "learning_rate": 0.0001,
      "loss": 0.0016,
      "step": 6640
    },
    {
      "epoch": 89.78,
      "learning_rate": 0.0001,
      "loss": 0.0008,
      "step": 6644
    },
    {
      "epoch": 89.83,
      "learning_rate": 0.0001,
      "loss": 0.0009,
      "step": 6648
    },
    {
      "epoch": 89.88,
      "learning_rate": 0.0001,
      "loss": 0.0007,
      "step": 6652
    },
    {
      "epoch": 89.94,
      "learning_rate": 0.0001,
      "loss": 0.0003,
      "step": 6656
    },
    {
      "epoch": 89.94,
      "eval_exec": 0.7127659574468085,
      "eval_loss": 0.3296290338039398,
      "eval_runtime": 289.4346,
      "eval_samples_per_second": 3.572,
      "step": 6656
    },
    {
      "epoch": 89.99,
      "learning_rate": 0.0001,
      "loss": 0.0008,
      "step": 6660
    },
    {
      "epoch": 90.05,
      "learning_rate": 0.0001,
      "loss": 0.0002,
      "step": 6664
    },
    {
      "epoch": 90.11,
      "learning_rate": 0.0001,
      "loss": 0.0013,
      "step": 6668
    },
    {
      "epoch": 90.16,
      "learning_rate": 0.0001,
      "loss": 0.0002,
      "step": 6672
    },
    {
      "epoch": 90.21,
      "learning_rate": 0.0001,
      "loss": 0.0003,
      "step": 6676
    },
    {
      "epoch": 90.27,
      "learning_rate": 0.0001,
      "loss": 0.0007,
      "step": 6680
    },
    {
      "epoch": 90.32,
      "learning_rate": 0.0001,
      "loss": 0.0003,
      "step": 6684
    },
    {
      "epoch": 90.37,
      "learning_rate": 0.0001,
      "loss": 0.0028,
      "step": 6688
    },
    {
      "epoch": 90.43,
      "learning_rate": 0.0001,
      "loss": 0.0005,
      "step": 6692
    },
    {
      "epoch": 90.48,
      "learning_rate": 0.0001,
      "loss": 0.0014,
      "step": 6696
    },
    {
      "epoch": 90.53,
      "learning_rate": 0.0001,
      "loss": 0.0005,
      "step": 6700
    },
    {
      "epoch": 90.59,
      "learning_rate": 0.0001,
      "loss": 0.0004,
      "step": 6704
    },
    {
      "epoch": 90.64,
      "learning_rate": 0.0001,
      "loss": 0.0003,
      "step": 6708
    },
    {
      "epoch": 90.7,
      "learning_rate": 0.0001,
      "loss": 0.0001,
      "step": 6712
    },
    {
      "epoch": 90.75,
      "learning_rate": 0.0001,
      "loss": 0.0005,
      "step": 6716
    },
    {
      "epoch": 90.8,
      "learning_rate": 0.0001,
      "loss": 0.0009,
      "step": 6720
    },
    {
      "epoch": 90.8,
      "eval_exec": 0.7050290135396519,
      "eval_loss": 0.33742156624794006,
      "eval_runtime": 286.2212,
      "eval_samples_per_second": 3.613,
      "step": 6720
    },
    {
      "epoch": 90.86,
      "learning_rate": 0.0001,
      "loss": 0.0012,
      "step": 6724
    },
    {
      "epoch": 90.91,
      "learning_rate": 0.0001,
      "loss": 0.0001,
      "step": 6728
    },
    {
      "epoch": 90.96,
      "learning_rate": 0.0001,
      "loss": 0.0024,
      "step": 6732
    },
    {
      "epoch": 91.03,
      "learning_rate": 0.0001,
      "loss": 0.0014,
      "step": 6736
    },
    {
      "epoch": 91.08,
      "learning_rate": 0.0001,
      "loss": 0.0002,
      "step": 6740
    },
    {
      "epoch": 91.13,
      "learning_rate": 0.0001,
      "loss": 0.0014,
      "step": 6744
    },
    {
      "epoch": 91.19,
      "learning_rate": 0.0001,
      "loss": 0.0004,
      "step": 6748
    },
    {
      "epoch": 91.24,
      "learning_rate": 0.0001,
      "loss": 0.0005,
      "step": 6752
    },
    {
      "epoch": 91.29,
      "learning_rate": 0.0001,
      "loss": 0.0002,
      "step": 6756
    },
    {
      "epoch": 91.35,
      "learning_rate": 0.0001,
      "loss": 0.0002,
      "step": 6760
    },
    {
      "epoch": 91.4,
      "learning_rate": 0.0001,
      "loss": 0.0014,
      "step": 6764
    },
    {
      "epoch": 91.45,
      "learning_rate": 0.0001,
      "loss": 0.0001,
      "step": 6768
    },
    {
      "epoch": 91.51,
      "learning_rate": 0.0001,
      "loss": 0.0002,
      "step": 6772
    },
    {
      "epoch": 91.56,
      "learning_rate": 0.0001,
      "loss": 0.0002,
      "step": 6776
    },
    {
      "epoch": 91.62,
      "learning_rate": 0.0001,
      "loss": 0.0006,
      "step": 6780
    },
    {
      "epoch": 91.67,
      "learning_rate": 0.0001,
      "loss": 0.0011,
      "step": 6784
    },
    {
      "epoch": 91.67,
      "eval_exec": 0.7030947775628626,
      "eval_loss": 0.3439423739910126,
      "eval_runtime": 283.1873,
      "eval_samples_per_second": 3.651,
      "step": 6784
    },
    {
      "epoch": 91.72,
      "learning_rate": 0.0001,
      "loss": 0.0002,
      "step": 6788
    },
    {
      "epoch": 91.78,
      "learning_rate": 0.0001,
      "loss": 0.0013,
      "step": 6792
    },
    {
      "epoch": 91.83,
      "learning_rate": 0.0001,
      "loss": 0.0005,
      "step": 6796
    },
    {
      "epoch": 91.88,
      "learning_rate": 0.0001,
      "loss": 0.0014,
      "step": 6800
    },
    {
      "epoch": 91.94,
      "learning_rate": 0.0001,
      "loss": 0.0002,
      "step": 6804
    },
    {
      "epoch": 91.99,
      "learning_rate": 0.0001,
      "loss": 0.0008,
      "step": 6808
    },
    {
      "epoch": 92.05,
      "learning_rate": 0.0001,
      "loss": 0.0005,
      "step": 6812
    },
    {
      "epoch": 92.11,
      "learning_rate": 0.0001,
      "loss": 0.0007,
      "step": 6816
    },
    {
      "epoch": 92.16,
      "learning_rate": 0.0001,
      "loss": 0.0014,
      "step": 6820
    },
    {
      "epoch": 92.21,
      "learning_rate": 0.0001,
      "loss": 0.0002,
      "step": 6824
    },
    {
      "epoch": 92.27,
      "learning_rate": 0.0001,
      "loss": 0.0005,
      "step": 6828
    },
    {
      "epoch": 92.32,
      "learning_rate": 0.0001,
      "loss": 0.0006,
      "step": 6832
    },
    {
      "epoch": 92.37,
      "learning_rate": 0.0001,
      "loss": 0.0017,
      "step": 6836
    },
    {
      "epoch": 92.43,
      "learning_rate": 0.0001,
      "loss": 0.0003,
      "step": 6840
    },
    {
      "epoch": 92.48,
      "learning_rate": 0.0001,
      "loss": 0.0004,
      "step": 6844
    },
    {
      "epoch": 92.53,
      "learning_rate": 0.0001,
      "loss": 0.0011,
      "step": 6848
    },
    {
      "epoch": 92.53,
      "eval_exec": 0.7050290135396519,
      "eval_loss": 0.3420410752296448,
      "eval_runtime": 289.1079,
      "eval_samples_per_second": 3.577,
      "step": 6848
    },
    {
      "epoch": 92.59,
      "learning_rate": 0.0001,
      "loss": 0.0006,
      "step": 6852
    },
    {
      "epoch": 92.64,
      "learning_rate": 0.0001,
      "loss": 0.0005,
      "step": 6856
    },
    {
      "epoch": 92.7,
      "learning_rate": 0.0001,
      "loss": 0.0003,
      "step": 6860
    },
    {
      "epoch": 92.75,
      "learning_rate": 0.0001,
      "loss": 0.0005,
      "step": 6864
    },
    {
      "epoch": 92.8,
      "learning_rate": 0.0001,
      "loss": 0.0003,
      "step": 6868
    },
    {
      "epoch": 92.86,
      "learning_rate": 0.0001,
      "loss": 0.001,
      "step": 6872
    },
    {
      "epoch": 92.91,
      "learning_rate": 0.0001,
      "loss": 0.0002,
      "step": 6876
    },
    {
      "epoch": 92.96,
      "learning_rate": 0.0001,
      "loss": 0.0011,
      "step": 6880
    },
    {
      "epoch": 93.03,
      "learning_rate": 0.0001,
      "loss": 0.0007,
      "step": 6884
    },
    {
      "epoch": 93.08,
      "learning_rate": 0.0001,
      "loss": 0.0002,
      "step": 6888
    },
    {
      "epoch": 93.13,
      "learning_rate": 0.0001,
      "loss": 0.0007,
      "step": 6892
    },
    {
      "epoch": 93.19,
      "learning_rate": 0.0001,
      "loss": 0.0006,
      "step": 6896
    },
    {
      "epoch": 93.24,
      "learning_rate": 0.0001,
      "loss": 0.0009,
      "step": 6900
    },
    {
      "epoch": 93.29,
      "learning_rate": 0.0001,
      "loss": 0.0002,
      "step": 6904
    },
    {
      "epoch": 93.35,
      "learning_rate": 0.0001,
      "loss": 0.0014,
      "step": 6908
    },
    {
      "epoch": 93.4,
      "learning_rate": 0.0001,
      "loss": 0.0004,
      "step": 6912
    },
    {
      "epoch": 93.4,
      "eval_exec": 0.7263056092843327,
      "eval_loss": 0.337966650724411,
      "eval_runtime": 283.4007,
      "eval_samples_per_second": 3.649,
      "step": 6912
    },
    {
      "epoch": 93.45,
      "learning_rate": 0.0001,
      "loss": 0.0011,
      "step": 6916
    },
    {
      "epoch": 93.51,
      "learning_rate": 0.0001,
      "loss": 0.0006,
      "step": 6920
    },
    {
      "epoch": 93.56,
      "learning_rate": 0.0001,
      "loss": 0.0004,
      "step": 6924
    },
    {
      "epoch": 93.62,
      "learning_rate": 0.0001,
      "loss": 0.001,
      "step": 6928
    },
    {
      "epoch": 93.67,
      "learning_rate": 0.0001,
      "loss": 0.0007,
      "step": 6932
    },
    {
      "epoch": 93.72,
      "learning_rate": 0.0001,
      "loss": 0.0003,
      "step": 6936
    },
    {
      "epoch": 93.78,
      "learning_rate": 0.0001,
      "loss": 0.0006,
      "step": 6940
    },
    {
      "epoch": 93.83,
      "learning_rate": 0.0001,
      "loss": 0.0006,
      "step": 6944
    },
    {
      "epoch": 93.88,
      "learning_rate": 0.0001,
      "loss": 0.0003,
      "step": 6948
    },
    {
      "epoch": 93.94,
      "learning_rate": 0.0001,
      "loss": 0.0002,
      "step": 6952
    },
    {
      "epoch": 93.99,
      "learning_rate": 0.0001,
      "loss": 0.001,
      "step": 6956
    },
    {
      "epoch": 94.05,
      "learning_rate": 0.0001,
      "loss": 0.0004,
      "step": 6960
    },
    {
      "epoch": 94.11,
      "learning_rate": 0.0001,
      "loss": 0.0008,
      "step": 6964
    },
    {
      "epoch": 94.16,
      "learning_rate": 0.0001,
      "loss": 0.0002,
      "step": 6968
    },
    {
      "epoch": 94.21,
      "learning_rate": 0.0001,
      "loss": 0.0001,
      "step": 6972
    },
    {
      "epoch": 94.27,
      "learning_rate": 0.0001,
      "loss": 0.0001,
      "step": 6976
    },
    {
      "epoch": 94.27,
      "eval_exec": 0.7176015473887815,
      "eval_loss": 0.364266037940979,
      "eval_runtime": 284.3795,
      "eval_samples_per_second": 3.636,
      "step": 6976
    },
    {
      "epoch": 94.32,
      "learning_rate": 0.0001,
      "loss": 0.0003,
      "step": 6980
    },
    {
      "epoch": 94.37,
      "learning_rate": 0.0001,
      "loss": 0.0002,
      "step": 6984
    },
    {
      "epoch": 94.43,
      "learning_rate": 0.0001,
      "loss": 0.0003,
      "step": 6988
    },
    {
      "epoch": 94.48,
      "learning_rate": 0.0001,
      "loss": 0.0004,
      "step": 6992
    },
    {
      "epoch": 94.53,
      "learning_rate": 0.0001,
      "loss": 0.0003,
      "step": 6996
    },
    {
      "epoch": 94.59,
      "learning_rate": 0.0001,
      "loss": 0.0005,
      "step": 7000
    },
    {
      "epoch": 94.64,
      "learning_rate": 0.0001,
      "loss": 0.0017,
      "step": 7004
    },
    {
      "epoch": 94.7,
      "learning_rate": 0.0001,
      "loss": 0.0006,
      "step": 7008
    },
    {
      "epoch": 94.75,
      "learning_rate": 0.0001,
      "loss": 0.0006,
      "step": 7012
    },
    {
      "epoch": 94.8,
      "learning_rate": 0.0001,
      "loss": 0.001,
      "step": 7016
    },
    {
      "epoch": 94.86,
      "learning_rate": 0.0001,
      "loss": 0.0001,
      "step": 7020
    },
    {
      "epoch": 94.91,
      "learning_rate": 0.0001,
      "loss": 0.0004,
      "step": 7024
    },
    {
      "epoch": 94.96,
      "learning_rate": 0.0001,
      "loss": 0.0004,
      "step": 7028
    },
    {
      "epoch": 95.03,
      "learning_rate": 0.0001,
      "loss": 0.0003,
      "step": 7032
    },
    {
      "epoch": 95.08,
      "learning_rate": 0.0001,
      "loss": 0.0006,
      "step": 7036
    },
    {
      "epoch": 95.13,
      "learning_rate": 0.0001,
      "loss": 0.0005,
      "step": 7040
    },
    {
      "epoch": 95.13,
      "eval_exec": 0.730174081237911,
      "eval_loss": 0.3686355650424957,
      "eval_runtime": 284.9995,
      "eval_samples_per_second": 3.628,
      "step": 7040
    },
    {
      "epoch": 95.19,
      "learning_rate": 0.0001,
      "loss": 0.0001,
      "step": 7044
    },
    {
      "epoch": 95.24,
      "learning_rate": 0.0001,
      "loss": 0.0002,
      "step": 7048
    },
    {
      "epoch": 95.29,
      "learning_rate": 0.0001,
      "loss": 0.0028,
      "step": 7052
    },
    {
      "epoch": 95.35,
      "learning_rate": 0.0001,
      "loss": 0.0006,
      "step": 7056
    },
    {
      "epoch": 95.4,
      "learning_rate": 0.0001,
      "loss": 0.0001,
      "step": 7060
    },
    {
      "epoch": 95.45,
      "learning_rate": 0.0001,
      "loss": 0.0004,
      "step": 7064
    },
    {
      "epoch": 95.51,
      "learning_rate": 0.0001,
      "loss": 0.0005,
      "step": 7068
    },
    {
      "epoch": 95.56,
      "learning_rate": 0.0001,
      "loss": 0.0012,
      "step": 7072
    },
    {
      "epoch": 95.62,
      "learning_rate": 0.0001,
      "loss": 0.0004,
      "step": 7076
    },
    {
      "epoch": 95.67,
      "learning_rate": 0.0001,
      "loss": 0.0003,
      "step": 7080
    },
    {
      "epoch": 95.72,
      "learning_rate": 0.0001,
      "loss": 0.0002,
      "step": 7084
    },
    {
      "epoch": 95.78,
      "learning_rate": 0.0001,
      "loss": 0.0002,
      "step": 7088
    },
    {
      "epoch": 95.83,
      "learning_rate": 0.0001,
      "loss": 0.001,
      "step": 7092
    },
    {
      "epoch": 95.88,
      "learning_rate": 0.0001,
      "loss": 0.0002,
      "step": 7096
    },
    {
      "epoch": 95.94,
      "learning_rate": 0.0001,
      "loss": 0.0001,
      "step": 7100
    },
    {
      "epoch": 95.99,
      "learning_rate": 0.0001,
      "loss": 0.0001,
      "step": 7104
    },
    {
      "epoch": 95.99,
      "eval_exec": 0.7176015473887815,
      "eval_loss": 0.3639739453792572,
      "eval_runtime": 281.8061,
      "eval_samples_per_second": 3.669,
      "step": 7104
    },
    {
      "epoch": 96.05,
      "learning_rate": 0.0001,
      "loss": 0.0008,
      "step": 7108
    },
    {
      "epoch": 96.11,
      "learning_rate": 0.0001,
      "loss": 0.0007,
      "step": 7112
    },
    {
      "epoch": 96.16,
      "learning_rate": 0.0001,
      "loss": 0.0006,
      "step": 7116
    },
    {
      "epoch": 96.21,
      "learning_rate": 0.0001,
      "loss": 0.0004,
      "step": 7120
    },
    {
      "epoch": 96.27,
      "learning_rate": 0.0001,
      "loss": 0.0002,
      "step": 7124
    },
    {
      "epoch": 96.32,
      "learning_rate": 0.0001,
      "loss": 0.001,
      "step": 7128
    },
    {
      "epoch": 96.37,
      "learning_rate": 0.0001,
      "loss": 0.0004,
      "step": 7132
    },
    {
      "epoch": 96.43,
      "learning_rate": 0.0001,
      "loss": 0.0001,
      "step": 7136
    },
    {
      "epoch": 96.48,
      "learning_rate": 0.0001,
      "loss": 0.0004,
      "step": 7140
    },
    {
      "epoch": 96.53,
      "learning_rate": 0.0001,
      "loss": 0.001,
      "step": 7144
    },
    {
      "epoch": 96.59,
      "learning_rate": 0.0001,
      "loss": 0.0009,
      "step": 7148
    },
    {
      "epoch": 96.64,
      "learning_rate": 0.0001,
      "loss": 0.0002,
      "step": 7152
    },
    {
      "epoch": 96.7,
      "learning_rate": 0.0001,
      "loss": 0.0005,
      "step": 7156
    },
    {
      "epoch": 96.75,
      "learning_rate": 0.0001,
      "loss": 0.0005,
      "step": 7160
    },
    {
      "epoch": 96.8,
      "learning_rate": 0.0001,
      "loss": 0.0011,
      "step": 7164
    },
    {
      "epoch": 96.86,
      "learning_rate": 0.0001,
      "loss": 0.0008,
      "step": 7168
    },
    {
      "epoch": 96.86,
      "eval_exec": 0.7224371373307543,
      "eval_loss": 0.35076427459716797,
      "eval_runtime": 285.1481,
      "eval_samples_per_second": 3.626,
      "step": 7168
    },
    {
      "epoch": 96.91,
      "learning_rate": 0.0001,
      "loss": 0.0005,
      "step": 7172
    },
    {
      "epoch": 96.96,
      "learning_rate": 0.0001,
      "loss": 0.0023,
      "step": 7176
    },
    {
      "epoch": 97.03,
      "learning_rate": 0.0001,
      "loss": 0.0007,
      "step": 7180
    },
    {
      "epoch": 97.08,
      "learning_rate": 0.0001,
      "loss": 0.0004,
      "step": 7184
    },
    {
      "epoch": 97.13,
      "learning_rate": 0.0001,
      "loss": 0.0009,
      "step": 7188
    },
    {
      "epoch": 97.19,
      "learning_rate": 0.0001,
      "loss": 0.0009,
      "step": 7192
    },
    {
      "epoch": 97.24,
      "learning_rate": 0.0001,
      "loss": 0.0002,
      "step": 7196
    },
    {
      "epoch": 97.29,
      "learning_rate": 0.0001,
      "loss": 0.0002,
      "step": 7200
    },
    {
      "epoch": 97.35,
      "learning_rate": 0.0001,
      "loss": 0.0007,
      "step": 7204
    },
    {
      "epoch": 97.4,
      "learning_rate": 0.0001,
      "loss": 0.0007,
      "step": 7208
    },
    {
      "epoch": 97.45,
      "learning_rate": 0.0001,
      "loss": 0.0007,
      "step": 7212
    },
    {
      "epoch": 97.51,
      "learning_rate": 0.0001,
      "loss": 0.0006,
      "step": 7216
    },
    {
      "epoch": 97.56,
      "learning_rate": 0.0001,
      "loss": 0.0004,
      "step": 7220
    },
    {
      "epoch": 97.62,
      "learning_rate": 0.0001,
      "loss": 0.0003,
      "step": 7224
    },
    {
      "epoch": 97.67,
      "learning_rate": 0.0001,
      "loss": 0.0009,
      "step": 7228
    },
    {
      "epoch": 97.72,
      "learning_rate": 0.0001,
      "loss": 0.0002,
      "step": 7232
    },
    {
      "epoch": 97.72,
      "eval_exec": 0.723404255319149,
      "eval_loss": 0.34438732266426086,
      "eval_runtime": 287.451,
      "eval_samples_per_second": 3.597,
      "step": 7232
    },
    {
      "epoch": 97.78,
      "learning_rate": 0.0001,
      "loss": 0.0003,
      "step": 7236
    },
    {
      "epoch": 97.83,
      "learning_rate": 0.0001,
      "loss": 0.0005,
      "step": 7240
    },
    {
      "epoch": 97.88,
      "learning_rate": 0.0001,
      "loss": 0.0003,
      "step": 7244
    },
    {
      "epoch": 97.94,
      "learning_rate": 0.0001,
      "loss": 0.0003,
      "step": 7248
    },
    {
      "epoch": 97.99,
      "learning_rate": 0.0001,
      "loss": 0.0004,
      "step": 7252
    },
    {
      "epoch": 98.05,
      "learning_rate": 0.0001,
      "loss": 0.0008,
      "step": 7256
    },
    {
      "epoch": 98.11,
      "learning_rate": 0.0001,
      "loss": 0.0007,
      "step": 7260
    },
    {
      "epoch": 98.16,
      "learning_rate": 0.0001,
      "loss": 0.0004,
      "step": 7264
    },
    {
      "epoch": 98.21,
      "learning_rate": 0.0001,
      "loss": 0.0004,
      "step": 7268
    },
    {
      "epoch": 98.27,
      "learning_rate": 0.0001,
      "loss": 0.0017,
      "step": 7272
    },
    {
      "epoch": 98.32,
      "learning_rate": 0.0001,
      "loss": 0.0006,
      "step": 7276
    },
    {
      "epoch": 98.37,
      "learning_rate": 0.0001,
      "loss": 0.0004,
      "step": 7280
    },
    {
      "epoch": 98.43,
      "learning_rate": 0.0001,
      "loss": 0.0005,
      "step": 7284
    },
    {
      "epoch": 98.48,
      "learning_rate": 0.0001,
      "loss": 0.0001,
      "step": 7288
    },
    {
      "epoch": 98.53,
      "learning_rate": 0.0001,
      "loss": 0.0003,
      "step": 7292
    },
    {
      "epoch": 98.59,
      "learning_rate": 0.0001,
      "loss": 0.0002,
      "step": 7296
    },
    {
      "epoch": 98.59,
      "eval_exec": 0.718568665377176,
      "eval_loss": 0.3525933623313904,
      "eval_runtime": 283.2747,
      "eval_samples_per_second": 3.65,
      "step": 7296
    },
    {
      "epoch": 98.64,
      "learning_rate": 0.0001,
      "loss": 0.0002,
      "step": 7300
    },
    {
      "epoch": 98.7,
      "learning_rate": 0.0001,
      "loss": 0.0005,
      "step": 7304
    },
    {
      "epoch": 98.75,
      "learning_rate": 0.0001,
      "loss": 0.0002,
      "step": 7308
    },
    {
      "epoch": 98.8,
      "learning_rate": 0.0001,
      "loss": 0.0004,
      "step": 7312
    },
    {
      "epoch": 98.86,
      "learning_rate": 0.0001,
      "loss": 0.0002,
      "step": 7316
    },
    {
      "epoch": 98.91,
      "learning_rate": 0.0001,
      "loss": 0.0008,
      "step": 7320
    },
    {
      "epoch": 98.96,
      "learning_rate": 0.0001,
      "loss": 0.0004,
      "step": 7324
    },
    {
      "epoch": 99.03,
      "learning_rate": 0.0001,
      "loss": 0.0003,
      "step": 7328
    },
    {
      "epoch": 99.08,
      "learning_rate": 0.0001,
      "loss": 0.0002,
      "step": 7332
    },
    {
      "epoch": 99.13,
      "learning_rate": 0.0001,
      "loss": 0.0002,
      "step": 7336
    },
    {
      "epoch": 99.19,
      "learning_rate": 0.0001,
      "loss": 0.001,
      "step": 7340
    },
    {
      "epoch": 99.24,
      "learning_rate": 0.0001,
      "loss": 0.0001,
      "step": 7344
    },
    {
      "epoch": 99.29,
      "learning_rate": 0.0001,
      "loss": 0.0003,
      "step": 7348
    },
    {
      "epoch": 99.35,
      "learning_rate": 0.0001,
      "loss": 0.0002,
      "step": 7352
    },
    {
      "epoch": 99.4,
      "learning_rate": 0.0001,
      "loss": 0.0002,
      "step": 7356
    },
    {
      "epoch": 99.45,
      "learning_rate": 0.0001,
      "loss": 0.0002,
      "step": 7360
    },
    {
      "epoch": 99.45,
      "eval_exec": 0.7311411992263056,
      "eval_loss": 0.3651147186756134,
      "eval_runtime": 287.5898,
      "eval_samples_per_second": 3.595,
      "step": 7360
    },
    {
      "epoch": 99.51,
      "learning_rate": 0.0001,
      "loss": 0.0007,
      "step": 7364
    },
    {
      "epoch": 99.56,
      "learning_rate": 0.0001,
      "loss": 0.0002,
      "step": 7368
    },
    {
      "epoch": 99.62,
      "learning_rate": 0.0001,
      "loss": 0.0004,
      "step": 7372
    },
    {
      "epoch": 99.67,
      "learning_rate": 0.0001,
      "loss": 0.0002,
      "step": 7376
    },
    {
      "epoch": 99.72,
      "learning_rate": 0.0001,
      "loss": 0.0008,
      "step": 7380
    },
    {
      "epoch": 99.78,
      "learning_rate": 0.0001,
      "loss": 0.0009,
      "step": 7384
    },
    {
      "epoch": 99.83,
      "learning_rate": 0.0001,
      "loss": 0.0001,
      "step": 7388
    },
    {
      "epoch": 99.88,
      "learning_rate": 0.0001,
      "loss": 0.001,
      "step": 7392
    },
    {
      "epoch": 99.94,
      "learning_rate": 0.0001,
      "loss": 0.0002,
      "step": 7396
    },
    {
      "epoch": 99.99,
      "learning_rate": 0.0001,
      "loss": 0.0006,
      "step": 7400
    },
    {
      "epoch": 99.99,
      "step": 7400,
      "total_flos": 5.036125754774262e+18,
      "train_loss": 0.00013556757536748224,
      "train_runtime": 24498.5901,
      "train_samples_per_second": 61.04,
      "train_steps_per_second": 0.302
    }
  ],
  "max_steps": 7400,
  "num_train_epochs": 100,
  "total_flos": 5.036125754774262e+18,
  "trial_name": null,
  "trial_params": null
}