{
  "best_metric": 0.02997952327132225,
  "best_model_checkpoint": "runs/deepseek_lora_20240424-122712/checkpoint-1000",
  "epoch": 0.31330774653403304,
  "eval_steps": 500,
  "global_step": 1000,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.0,
      "grad_norm": 2.6060585975646973,
      "learning_rate": 4.0000000000000003e-07,
      "loss": 1.6667,
      "step": 10
    },
    {
      "epoch": 0.01,
      "grad_norm": 5.726912021636963,
      "learning_rate": 8.000000000000001e-07,
      "loss": 1.6606,
      "step": 20
    },
    {
      "epoch": 0.01,
      "grad_norm": 6.3224616050720215,
      "learning_rate": 1.2000000000000002e-06,
      "loss": 1.6875,
      "step": 30
    },
    {
      "epoch": 0.01,
      "grad_norm": 4.4585185050964355,
      "learning_rate": 1.6000000000000001e-06,
      "loss": 1.5512,
      "step": 40
    },
    {
      "epoch": 0.02,
      "grad_norm": 2.849210023880005,
      "learning_rate": 2.0000000000000003e-06,
      "loss": 1.598,
      "step": 50
    },
    {
      "epoch": 0.02,
      "grad_norm": 1.7917169332504272,
      "learning_rate": 2.4000000000000003e-06,
      "loss": 1.3203,
      "step": 60
    },
    {
      "epoch": 0.02,
      "grad_norm": 10.056238174438477,
      "learning_rate": 2.8000000000000003e-06,
      "loss": 1.4914,
      "step": 70
    },
    {
      "epoch": 0.03,
      "grad_norm": 5.542996406555176,
      "learning_rate": 3.2000000000000003e-06,
      "loss": 1.3369,
      "step": 80
    },
    {
      "epoch": 0.03,
      "grad_norm": 1.5463522672653198,
      "learning_rate": 3.6000000000000003e-06,
      "loss": 1.3185,
      "step": 90
    },
    {
      "epoch": 0.03,
      "grad_norm": 9.401649475097656,
      "learning_rate": 4.000000000000001e-06,
      "loss": 1.2827,
      "step": 100
    },
    {
      "epoch": 0.03,
      "grad_norm": 1.238974928855896,
      "learning_rate": 4.4e-06,
      "loss": 1.0366,
      "step": 110
    },
    {
      "epoch": 0.04,
      "grad_norm": 1.2282520532608032,
      "learning_rate": 4.800000000000001e-06,
      "loss": 1.0715,
      "step": 120
    },
    {
      "epoch": 0.04,
      "grad_norm": 1.4583073854446411,
      "learning_rate": 5.2e-06,
      "loss": 0.868,
      "step": 130
    },
    {
      "epoch": 0.04,
      "grad_norm": 4.3194146156311035,
      "learning_rate": 5.600000000000001e-06,
      "loss": 0.8281,
      "step": 140
    },
    {
      "epoch": 0.05,
      "grad_norm": 4.280028820037842,
      "learning_rate": 6e-06,
      "loss": 0.8666,
      "step": 150
    },
    {
      "epoch": 0.05,
      "grad_norm": 0.9424476027488708,
      "learning_rate": 6.4000000000000006e-06,
      "loss": 0.6691,
      "step": 160
    },
    {
      "epoch": 0.05,
      "grad_norm": 0.9174453616142273,
      "learning_rate": 6.800000000000001e-06,
      "loss": 0.5405,
      "step": 170
    },
    {
      "epoch": 0.06,
      "grad_norm": 1.0179359912872314,
      "learning_rate": 7.2000000000000005e-06,
      "loss": 0.5103,
      "step": 180
    },
    {
      "epoch": 0.06,
      "grad_norm": 0.4516351521015167,
      "learning_rate": 7.600000000000001e-06,
      "loss": 0.2679,
      "step": 190
    },
    {
      "epoch": 0.06,
      "grad_norm": 6.054721355438232,
      "learning_rate": 8.000000000000001e-06,
      "loss": 0.2634,
      "step": 200
    },
    {
      "epoch": 0.07,
      "grad_norm": 0.40070173144340515,
      "learning_rate": 8.400000000000001e-06,
      "loss": 0.2254,
      "step": 210
    },
    {
      "epoch": 0.07,
      "grad_norm": 0.6031996011734009,
      "learning_rate": 8.8e-06,
      "loss": 0.1281,
      "step": 220
    },
    {
      "epoch": 0.07,
      "grad_norm": 3.0129740238189697,
      "learning_rate": 9.200000000000002e-06,
      "loss": 0.2365,
      "step": 230
    },
    {
      "epoch": 0.08,
      "grad_norm": 0.6611989140510559,
      "learning_rate": 9.600000000000001e-06,
      "loss": 0.3255,
      "step": 240
    },
    {
      "epoch": 0.08,
      "grad_norm": 0.3195549547672272,
      "learning_rate": 1e-05,
      "loss": 0.3287,
      "step": 250
    },
    {
      "epoch": 0.08,
      "grad_norm": 0.27022993564605713,
      "learning_rate": 1.04e-05,
      "loss": 0.1439,
      "step": 260
    },
    {
      "epoch": 0.08,
      "grad_norm": 4.671290874481201,
      "learning_rate": 1.0800000000000002e-05,
      "loss": 0.0676,
      "step": 270
    },
    {
      "epoch": 0.09,
      "grad_norm": 0.2435157746076584,
      "learning_rate": 1.1200000000000001e-05,
      "loss": 0.2338,
      "step": 280
    },
    {
      "epoch": 0.09,
      "grad_norm": 0.40632137656211853,
      "learning_rate": 1.16e-05,
      "loss": 0.1661,
      "step": 290
    },
    {
      "epoch": 0.09,
      "grad_norm": 0.2910804748535156,
      "learning_rate": 1.2e-05,
      "loss": 0.1781,
      "step": 300
    },
    {
      "epoch": 0.1,
      "grad_norm": 7.506131649017334,
      "learning_rate": 1.2400000000000002e-05,
      "loss": 0.1283,
      "step": 310
    },
    {
      "epoch": 0.1,
      "grad_norm": 0.04508267343044281,
      "learning_rate": 1.2800000000000001e-05,
      "loss": 0.1022,
      "step": 320
    },
    {
      "epoch": 0.1,
      "grad_norm": 0.23127306997776031,
      "learning_rate": 1.3200000000000002e-05,
      "loss": 0.1243,
      "step": 330
    },
    {
      "epoch": 0.11,
      "grad_norm": 6.1613383293151855,
      "learning_rate": 1.3600000000000002e-05,
      "loss": 0.1485,
      "step": 340
    },
    {
      "epoch": 0.11,
      "grad_norm": 0.023777758702635765,
      "learning_rate": 1.4e-05,
      "loss": 0.0538,
      "step": 350
    },
    {
      "epoch": 0.11,
      "grad_norm": 0.014769719913601875,
      "learning_rate": 1.4400000000000001e-05,
      "loss": 0.0285,
      "step": 360
    },
    {
      "epoch": 0.12,
      "grad_norm": 5.690178871154785,
      "learning_rate": 1.48e-05,
      "loss": 0.1325,
      "step": 370
    },
    {
      "epoch": 0.12,
      "grad_norm": 0.02117346040904522,
      "learning_rate": 1.5200000000000002e-05,
      "loss": 0.0507,
      "step": 380
    },
    {
      "epoch": 0.12,
      "grad_norm": 0.06630611419677734,
      "learning_rate": 1.5600000000000003e-05,
      "loss": 0.199,
      "step": 390
    },
    {
      "epoch": 0.13,
      "grad_norm": 0.16532257199287415,
      "learning_rate": 1.6000000000000003e-05,
      "loss": 0.0652,
      "step": 400
    },
    {
      "epoch": 0.13,
      "grad_norm": 0.009799620136618614,
      "learning_rate": 1.64e-05,
      "loss": 0.1011,
      "step": 410
    },
    {
      "epoch": 0.13,
      "grad_norm": 0.12324853241443634,
      "learning_rate": 1.6800000000000002e-05,
      "loss": 0.0364,
      "step": 420
    },
    {
      "epoch": 0.13,
      "grad_norm": 0.010026533156633377,
      "learning_rate": 1.72e-05,
      "loss": 0.1973,
      "step": 430
    },
    {
      "epoch": 0.14,
      "grad_norm": 0.0071570691652596,
      "learning_rate": 1.76e-05,
      "loss": 0.0819,
      "step": 440
    },
    {
      "epoch": 0.14,
      "grad_norm": 0.04149964451789856,
      "learning_rate": 1.8e-05,
      "loss": 0.1046,
      "step": 450
    },
    {
      "epoch": 0.14,
      "grad_norm": 3.45990252494812,
      "learning_rate": 1.8400000000000003e-05,
      "loss": 0.1011,
      "step": 460
    },
    {
      "epoch": 0.15,
      "grad_norm": 0.00734300771728158,
      "learning_rate": 1.88e-05,
      "loss": 0.223,
      "step": 470
    },
    {
      "epoch": 0.15,
      "grad_norm": 0.01419814396649599,
      "learning_rate": 1.9200000000000003e-05,
      "loss": 0.1178,
      "step": 480
    },
    {
      "epoch": 0.15,
      "grad_norm": 0.10111651569604874,
      "learning_rate": 1.9600000000000002e-05,
      "loss": 0.1416,
      "step": 490
    },
    {
      "epoch": 0.16,
      "grad_norm": 3.1214537620544434,
      "learning_rate": 2e-05,
      "loss": 0.1258,
      "step": 500
    },
    {
      "epoch": 0.16,
      "eval_loss": 0.05367860943078995,
      "eval_runtime": 62.0108,
      "eval_samples_per_second": 16.126,
      "eval_steps_per_second": 16.126,
      "step": 500
    },
    {
      "epoch": 0.16,
      "grad_norm": 0.013164438307285309,
      "learning_rate": 1.9955555555555557e-05,
      "loss": 0.0268,
      "step": 510
    },
    {
      "epoch": 0.16,
      "grad_norm": 0.15392057597637177,
      "learning_rate": 1.9911111111111112e-05,
      "loss": 0.0843,
      "step": 520
    },
    {
      "epoch": 0.17,
      "grad_norm": 0.006180985830724239,
      "learning_rate": 1.9866666666666667e-05,
      "loss": 0.1384,
      "step": 530
    },
    {
      "epoch": 0.17,
      "grad_norm": 0.2702454924583435,
      "learning_rate": 1.9822222222222226e-05,
      "loss": 0.0956,
      "step": 540
    },
    {
      "epoch": 0.17,
      "grad_norm": 1.5135071277618408,
      "learning_rate": 1.977777777777778e-05,
      "loss": 0.1615,
      "step": 550
    },
    {
      "epoch": 0.18,
      "grad_norm": 0.005121675785630941,
      "learning_rate": 1.9733333333333336e-05,
      "loss": 0.1201,
      "step": 560
    },
    {
      "epoch": 0.18,
      "grad_norm": 0.09728775173425674,
      "learning_rate": 1.968888888888889e-05,
      "loss": 0.1112,
      "step": 570
    },
    {
      "epoch": 0.18,
      "grad_norm": 4.736963748931885,
      "learning_rate": 1.9644444444444447e-05,
      "loss": 0.1546,
      "step": 580
    },
    {
      "epoch": 0.18,
      "grad_norm": 0.006898015271872282,
      "learning_rate": 1.9600000000000002e-05,
      "loss": 0.1116,
      "step": 590
    },
    {
      "epoch": 0.19,
      "grad_norm": 0.005388608202338219,
      "learning_rate": 1.9555555555555557e-05,
      "loss": 0.1597,
      "step": 600
    },
    {
      "epoch": 0.19,
      "grad_norm": 0.0133949751034379,
      "learning_rate": 1.9511111111111113e-05,
      "loss": 0.0799,
      "step": 610
    },
    {
      "epoch": 0.19,
      "grad_norm": 0.11226101964712143,
      "learning_rate": 1.9466666666666668e-05,
      "loss": 0.0448,
      "step": 620
    },
    {
      "epoch": 0.2,
      "grad_norm": 0.011205198243260384,
      "learning_rate": 1.9422222222222223e-05,
      "loss": 0.0739,
      "step": 630
    },
    {
      "epoch": 0.2,
      "grad_norm": 0.007312687113881111,
      "learning_rate": 1.9377777777777778e-05,
      "loss": 0.0779,
      "step": 640
    },
    {
      "epoch": 0.2,
      "grad_norm": 0.006550287362188101,
      "learning_rate": 1.9333333333333333e-05,
      "loss": 0.2039,
      "step": 650
    },
    {
      "epoch": 0.21,
      "grad_norm": 1.961844563484192,
      "learning_rate": 1.928888888888889e-05,
      "loss": 0.0618,
      "step": 660
    },
    {
      "epoch": 0.21,
      "grad_norm": 0.012318034656345844,
      "learning_rate": 1.9244444444444444e-05,
      "loss": 0.1166,
      "step": 670
    },
    {
      "epoch": 0.21,
      "grad_norm": 0.08778905868530273,
      "learning_rate": 1.9200000000000003e-05,
      "loss": 0.0822,
      "step": 680
    },
    {
      "epoch": 0.22,
      "grad_norm": 0.010623461566865444,
      "learning_rate": 1.9155555555555558e-05,
      "loss": 0.095,
      "step": 690
    },
    {
      "epoch": 0.22,
      "grad_norm": 0.0858495905995369,
      "learning_rate": 1.9111111111111113e-05,
      "loss": 0.0468,
      "step": 700
    },
    {
      "epoch": 0.22,
      "grad_norm": 1.7685880661010742,
      "learning_rate": 1.9066666666666668e-05,
      "loss": 0.0288,
      "step": 710
    },
    {
      "epoch": 0.23,
      "grad_norm": 0.032682083547115326,
      "learning_rate": 1.9022222222222223e-05,
      "loss": 0.1024,
      "step": 720
    },
    {
      "epoch": 0.23,
      "grad_norm": 0.010010900907218456,
      "learning_rate": 1.897777777777778e-05,
      "loss": 0.0129,
      "step": 730
    },
    {
      "epoch": 0.23,
      "grad_norm": 0.037693481892347336,
      "learning_rate": 1.8933333333333334e-05,
      "loss": 0.1217,
      "step": 740
    },
    {
      "epoch": 0.23,
      "grad_norm": 2.2111008167266846,
      "learning_rate": 1.888888888888889e-05,
      "loss": 0.1206,
      "step": 750
    },
    {
      "epoch": 0.24,
      "grad_norm": 1.4000599384307861,
      "learning_rate": 1.8844444444444444e-05,
      "loss": 0.1214,
      "step": 760
    },
    {
      "epoch": 0.24,
      "grad_norm": 4.3877034187316895,
      "learning_rate": 1.88e-05,
      "loss": 0.0408,
      "step": 770
    },
    {
      "epoch": 0.24,
      "grad_norm": 4.164296627044678,
      "learning_rate": 1.8755555555555558e-05,
      "loss": 0.1315,
      "step": 780
    },
    {
      "epoch": 0.25,
      "grad_norm": 0.07969211786985397,
      "learning_rate": 1.8711111111111113e-05,
      "loss": 0.0149,
      "step": 790
    },
    {
      "epoch": 0.25,
      "grad_norm": 3.9340951442718506,
      "learning_rate": 1.866666666666667e-05,
      "loss": 0.1169,
      "step": 800
    },
    {
      "epoch": 0.25,
      "grad_norm": 0.006517359986901283,
      "learning_rate": 1.8622222222222224e-05,
      "loss": 0.0517,
      "step": 810
    },
    {
      "epoch": 0.26,
      "grad_norm": 0.008951540105044842,
      "learning_rate": 1.857777777777778e-05,
      "loss": 0.0935,
      "step": 820
    },
    {
      "epoch": 0.26,
      "grad_norm": 0.03655437007546425,
      "learning_rate": 1.8533333333333334e-05,
      "loss": 0.0883,
      "step": 830
    },
    {
      "epoch": 0.26,
      "grad_norm": 0.0346522182226181,
      "learning_rate": 1.848888888888889e-05,
      "loss": 0.0479,
      "step": 840
    },
    {
      "epoch": 0.27,
      "grad_norm": 0.006223162170499563,
      "learning_rate": 1.8444444444444448e-05,
      "loss": 0.0075,
      "step": 850
    },
    {
      "epoch": 0.27,
      "grad_norm": 0.004037676844745874,
      "learning_rate": 1.8400000000000003e-05,
      "loss": 0.0333,
      "step": 860
    },
    {
      "epoch": 0.27,
      "grad_norm": 0.004569903016090393,
      "learning_rate": 1.835555555555556e-05,
      "loss": 0.0536,
      "step": 870
    },
    {
      "epoch": 0.28,
      "grad_norm": 0.005509174894541502,
      "learning_rate": 1.8311111111111114e-05,
      "loss": 0.1528,
      "step": 880
    },
    {
      "epoch": 0.28,
      "grad_norm": 0.007597978692501783,
      "learning_rate": 1.826666666666667e-05,
      "loss": 0.0052,
      "step": 890
    },
    {
      "epoch": 0.28,
      "grad_norm": 4.2217936515808105,
      "learning_rate": 1.8222222222222224e-05,
      "loss": 0.0776,
      "step": 900
    },
    {
      "epoch": 0.29,
      "grad_norm": 0.004342419560998678,
      "learning_rate": 1.817777777777778e-05,
      "loss": 0.0552,
      "step": 910
    },
    {
      "epoch": 0.29,
      "grad_norm": 0.0696156695485115,
      "learning_rate": 1.8133333333333335e-05,
      "loss": 0.0412,
      "step": 920
    },
    {
      "epoch": 0.29,
      "grad_norm": 0.022448547184467316,
      "learning_rate": 1.808888888888889e-05,
      "loss": 0.0109,
      "step": 930
    },
    {
      "epoch": 0.29,
      "grad_norm": 4.305741310119629,
      "learning_rate": 1.8044444444444445e-05,
      "loss": 0.0428,
      "step": 940
    },
    {
      "epoch": 0.3,
      "grad_norm": 3.344078779220581,
      "learning_rate": 1.8e-05,
      "loss": 0.0651,
      "step": 950
    },
    {
      "epoch": 0.3,
      "grad_norm": 2.585327386856079,
      "learning_rate": 1.7955555555555556e-05,
      "loss": 0.1474,
      "step": 960
    },
    {
      "epoch": 0.3,
      "grad_norm": 2.2476861476898193,
      "learning_rate": 1.791111111111111e-05,
      "loss": 0.0892,
      "step": 970
    },
    {
      "epoch": 0.31,
      "grad_norm": 0.06212488189339638,
      "learning_rate": 1.7866666666666666e-05,
      "loss": 0.0659,
      "step": 980
    },
    {
      "epoch": 0.31,
      "grad_norm": 0.012895594350993633,
      "learning_rate": 1.782222222222222e-05,
      "loss": 0.0915,
      "step": 990
    },
    {
      "epoch": 0.31,
      "grad_norm": 0.1314150094985962,
      "learning_rate": 1.7777777777777777e-05,
      "loss": 0.0456,
      "step": 1000
    },
    {
      "epoch": 0.31,
      "eval_loss": 0.02997952327132225,
      "eval_runtime": 62.0027,
      "eval_samples_per_second": 16.128,
      "eval_steps_per_second": 16.128,
      "step": 1000
    }
  ],
  "logging_steps": 10,
  "max_steps": 5000,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 2,
  "save_steps": 1000,
  "total_flos": 1.6102125993984e+16,
  "train_batch_size": 1,
  "trial_name": null,
  "trial_params": null
}