{
  "best_global_step": null,
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 2.2727272727272725,
  "eval_steps": 500,
  "global_step": 50,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.045454545454545456,
      "grad_norm": 1.234143614768982,
      "learning_rate": 0.0,
      "loss": 0.5856,
      "memory/device_reserved (GiB)": 32.79,
      "memory/max_active (GiB)": 29.82,
      "memory/max_allocated (GiB)": 29.82,
      "step": 1,
      "tokens_per_second_per_gpu": 1177.42
    },
    {
      "epoch": 0.09090909090909091,
      "grad_norm": 1.2433714866638184,
      "learning_rate": 0.0001,
      "loss": 0.5849,
      "memory/device_reserved (GiB)": 33.2,
      "memory/max_active (GiB)": 30.44,
      "memory/max_allocated (GiB)": 30.44,
      "step": 2,
      "tokens_per_second_per_gpu": 1374.23
    },
    {
      "epoch": 0.13636363636363635,
      "grad_norm": 0.7277888655662537,
      "learning_rate": 0.0002,
      "loss": 0.3524,
      "memory/device_reserved (GiB)": 33.2,
      "memory/max_active (GiB)": 30.44,
      "memory/max_allocated (GiB)": 30.44,
      "step": 3,
      "tokens_per_second_per_gpu": 1492.46
    },
    {
      "epoch": 0.18181818181818182,
      "grad_norm": 0.7649520039558411,
      "learning_rate": 0.00019987954562051725,
      "loss": 0.1434,
      "memory/device_reserved (GiB)": 33.2,
      "memory/max_active (GiB)": 30.44,
      "memory/max_allocated (GiB)": 30.44,
      "step": 4,
      "tokens_per_second_per_gpu": 1487.31
    },
    {
      "epoch": 0.22727272727272727,
      "grad_norm": 0.6205183863639832,
      "learning_rate": 0.0001995184726672197,
      "loss": 0.1129,
      "memory/device_reserved (GiB)": 33.2,
      "memory/max_active (GiB)": 30.44,
      "memory/max_allocated (GiB)": 30.44,
      "step": 5,
      "tokens_per_second_per_gpu": 1230.96
    },
    {
      "epoch": 0.2727272727272727,
      "grad_norm": 0.42128750681877136,
      "learning_rate": 0.0001989176509964781,
      "loss": 0.0528,
      "memory/device_reserved (GiB)": 33.2,
      "memory/max_active (GiB)": 30.44,
      "memory/max_allocated (GiB)": 30.44,
      "step": 6,
      "tokens_per_second_per_gpu": 1193.16
    },
    {
      "epoch": 0.3181818181818182,
      "grad_norm": 0.2397449016571045,
      "learning_rate": 0.00019807852804032305,
      "loss": 0.0444,
      "memory/device_reserved (GiB)": 33.2,
      "memory/max_active (GiB)": 30.44,
      "memory/max_allocated (GiB)": 30.44,
      "step": 7,
      "tokens_per_second_per_gpu": 1290.71
    },
    {
      "epoch": 0.36363636363636365,
      "grad_norm": 0.2838428020477295,
      "learning_rate": 0.00019700312531945442,
      "loss": 0.0637,
      "memory/device_reserved (GiB)": 33.2,
      "memory/max_active (GiB)": 30.44,
      "memory/max_allocated (GiB)": 30.44,
      "step": 8,
      "tokens_per_second_per_gpu": 1274.6
    },
    {
      "epoch": 0.4090909090909091,
      "grad_norm": 0.21046645939350128,
      "learning_rate": 0.0001956940335732209,
      "loss": 0.0418,
      "memory/device_reserved (GiB)": 33.2,
      "memory/max_active (GiB)": 30.44,
      "memory/max_allocated (GiB)": 30.44,
      "step": 9,
      "tokens_per_second_per_gpu": 1336.17
    },
    {
      "epoch": 0.45454545454545453,
      "grad_norm": 0.21617285907268524,
      "learning_rate": 0.00019415440651830208,
      "loss": 0.0544,
      "memory/device_reserved (GiB)": 33.2,
      "memory/max_active (GiB)": 30.44,
      "memory/max_allocated (GiB)": 30.44,
      "step": 10,
      "tokens_per_second_per_gpu": 1197.26
    },
    {
      "epoch": 0.5,
      "grad_norm": 0.17148727178573608,
      "learning_rate": 0.0001923879532511287,
      "loss": 0.0339,
      "memory/device_reserved (GiB)": 33.2,
      "memory/max_active (GiB)": 30.44,
      "memory/max_allocated (GiB)": 30.44,
      "step": 11,
      "tokens_per_second_per_gpu": 1285.65
    },
    {
      "epoch": 0.5454545454545454,
      "grad_norm": 0.16145595908164978,
      "learning_rate": 0.00019039892931234435,
      "loss": 0.0377,
      "memory/device_reserved (GiB)": 33.2,
      "memory/max_active (GiB)": 30.44,
      "memory/max_allocated (GiB)": 30.44,
      "step": 12,
      "tokens_per_second_per_gpu": 1244.59
    },
    {
      "epoch": 0.5909090909090909,
      "grad_norm": 0.13145197927951813,
      "learning_rate": 0.0001881921264348355,
      "loss": 0.0256,
      "memory/device_reserved (GiB)": 33.2,
      "memory/max_active (GiB)": 30.44,
      "memory/max_allocated (GiB)": 30.44,
      "step": 13,
      "tokens_per_second_per_gpu": 1372.2
    },
    {
      "epoch": 0.6363636363636364,
      "grad_norm": 0.13290230929851532,
      "learning_rate": 0.00018577286100002723,
      "loss": 0.0356,
      "memory/device_reserved (GiB)": 33.2,
      "memory/max_active (GiB)": 30.44,
      "memory/max_allocated (GiB)": 30.44,
      "step": 14,
      "tokens_per_second_per_gpu": 1302.02
    },
    {
      "epoch": 0.6818181818181818,
      "grad_norm": 0.16713210940361023,
      "learning_rate": 0.00018314696123025454,
      "loss": 0.0344,
      "memory/device_reserved (GiB)": 33.2,
      "memory/max_active (GiB)": 30.44,
      "memory/max_allocated (GiB)": 30.44,
      "step": 15,
      "tokens_per_second_per_gpu": 1461.31
    },
    {
      "epoch": 0.7272727272727273,
      "grad_norm": 0.13399545848369598,
      "learning_rate": 0.00018032075314806448,
      "loss": 0.0276,
      "memory/device_reserved (GiB)": 33.2,
      "memory/max_active (GiB)": 30.44,
      "memory/max_allocated (GiB)": 30.44,
      "step": 16,
      "tokens_per_second_per_gpu": 1290.94
    },
    {
      "epoch": 0.7727272727272727,
      "grad_norm": 0.13537150621414185,
      "learning_rate": 0.0001773010453362737,
      "loss": 0.0294,
      "memory/device_reserved (GiB)": 33.2,
      "memory/max_active (GiB)": 30.44,
      "memory/max_allocated (GiB)": 30.44,
      "step": 17,
      "tokens_per_second_per_gpu": 1297.86
    },
    {
      "epoch": 0.8181818181818182,
      "grad_norm": 0.17200890183448792,
      "learning_rate": 0.00017409511253549593,
      "loss": 0.0325,
      "memory/device_reserved (GiB)": 33.2,
      "memory/max_active (GiB)": 30.44,
      "memory/max_allocated (GiB)": 30.44,
      "step": 18,
      "tokens_per_second_per_gpu": 1397.83
    },
    {
      "epoch": 0.8636363636363636,
      "grad_norm": 0.15987926721572876,
      "learning_rate": 0.00017071067811865476,
      "loss": 0.0257,
      "memory/device_reserved (GiB)": 33.2,
      "memory/max_active (GiB)": 30.44,
      "memory/max_allocated (GiB)": 30.44,
      "step": 19,
      "tokens_per_second_per_gpu": 1383.02
    },
    {
      "epoch": 0.9090909090909091,
      "grad_norm": 0.13878504931926727,
      "learning_rate": 0.00016715589548470185,
      "loss": 0.0366,
      "memory/device_reserved (GiB)": 33.2,
      "memory/max_active (GiB)": 30.44,
      "memory/max_allocated (GiB)": 30.44,
      "step": 20,
      "tokens_per_second_per_gpu": 1265.2
    },
    {
      "epoch": 0.9545454545454546,
      "grad_norm": 0.11577529460191727,
      "learning_rate": 0.00016343932841636456,
      "loss": 0.0359,
      "memory/device_reserved (GiB)": 33.2,
      "memory/max_active (GiB)": 30.44,
      "memory/max_allocated (GiB)": 30.44,
      "step": 21,
      "tokens_per_second_per_gpu": 1304.73
    },
    {
      "epoch": 1.0,
      "grad_norm": 0.11737319827079773,
      "learning_rate": 0.00015956993044924334,
      "loss": 0.0239,
      "memory/device_reserved (GiB)": 33.2,
      "memory/max_active (GiB)": 30.44,
      "memory/max_allocated (GiB)": 30.44,
      "step": 22,
      "tokens_per_second_per_gpu": 1178.56
    },
    {
      "epoch": 1.0454545454545454,
      "grad_norm": 0.07948925346136093,
      "learning_rate": 0.00015555702330196023,
      "loss": 0.0254,
      "memory/device_reserved (GiB)": 33.2,
      "memory/max_active (GiB)": 30.44,
      "memory/max_allocated (GiB)": 30.44,
      "step": 23,
      "tokens_per_second_per_gpu": 1413.56
    },
    {
      "epoch": 1.0909090909090908,
      "grad_norm": 0.08376588672399521,
      "learning_rate": 0.00015141027441932216,
      "loss": 0.0157,
      "memory/device_reserved (GiB)": 33.2,
      "memory/max_active (GiB)": 30.44,
      "memory/max_allocated (GiB)": 30.44,
      "step": 24,
      "tokens_per_second_per_gpu": 1346.91
    },
    {
      "epoch": 1.1363636363636362,
      "grad_norm": 0.10371027141809464,
      "learning_rate": 0.0001471396736825998,
      "loss": 0.0215,
      "memory/device_reserved (GiB)": 33.2,
      "memory/max_active (GiB)": 30.44,
      "memory/max_allocated (GiB)": 30.44,
      "step": 25,
      "tokens_per_second_per_gpu": 1324.34
    },
    {
      "epoch": 1.1818181818181819,
      "grad_norm": 0.09667334705591202,
      "learning_rate": 0.00014275550934302823,
      "loss": 0.0166,
      "memory/device_reserved (GiB)": 33.2,
      "memory/max_active (GiB)": 30.44,
      "memory/max_allocated (GiB)": 30.44,
      "step": 26,
      "tokens_per_second_per_gpu": 1206.6
    },
    {
      "epoch": 1.2272727272727273,
      "grad_norm": 0.1148822233080864,
      "learning_rate": 0.000138268343236509,
      "loss": 0.0184,
      "memory/device_reserved (GiB)": 33.2,
      "memory/max_active (GiB)": 30.44,
      "memory/max_allocated (GiB)": 30.44,
      "step": 27,
      "tokens_per_second_per_gpu": 1302.38
    },
    {
      "epoch": 1.2727272727272727,
      "grad_norm": 0.10378631949424744,
      "learning_rate": 0.000133688985339222,
      "loss": 0.0236,
      "memory/device_reserved (GiB)": 33.2,
      "memory/max_active (GiB)": 30.44,
      "memory/max_allocated (GiB)": 30.44,
      "step": 28,
      "tokens_per_second_per_gpu": 1368.17
    },
    {
      "epoch": 1.3181818181818181,
      "grad_norm": 0.0753244087100029,
      "learning_rate": 0.00012902846772544624,
      "loss": 0.0097,
      "memory/device_reserved (GiB)": 33.2,
      "memory/max_active (GiB)": 30.44,
      "memory/max_allocated (GiB)": 30.44,
      "step": 29,
      "tokens_per_second_per_gpu": 1367.39
    },
    {
      "epoch": 1.3636363636363638,
      "grad_norm": 0.06268369406461716,
      "learning_rate": 0.0001242980179903264,
      "loss": 0.0108,
      "memory/device_reserved (GiB)": 33.2,
      "memory/max_active (GiB)": 30.44,
      "memory/max_allocated (GiB)": 30.44,
      "step": 30,
      "tokens_per_second_per_gpu": 1301.1
    },
    {
      "epoch": 1.4090909090909092,
      "grad_norm": 0.09844817221164703,
      "learning_rate": 0.00011950903220161285,
      "loss": 0.0175,
      "memory/device_reserved (GiB)": 33.2,
      "memory/max_active (GiB)": 30.44,
      "memory/max_allocated (GiB)": 30.44,
      "step": 31,
      "tokens_per_second_per_gpu": 1302.37
    },
    {
      "epoch": 1.4545454545454546,
      "grad_norm": 0.12313925474882126,
      "learning_rate": 0.00011467304744553618,
      "loss": 0.0202,
      "memory/device_reserved (GiB)": 33.2,
      "memory/max_active (GiB)": 30.44,
      "memory/max_allocated (GiB)": 30.44,
      "step": 32,
      "tokens_per_second_per_gpu": 1467.43
    },
    {
      "epoch": 1.5,
      "grad_norm": 0.0872587263584137,
      "learning_rate": 0.0001098017140329561,
      "loss": 0.012,
      "memory/device_reserved (GiB)": 33.2,
      "memory/max_active (GiB)": 30.44,
      "memory/max_allocated (GiB)": 30.44,
      "step": 33,
      "tokens_per_second_per_gpu": 1429.68
    },
    {
      "epoch": 1.5454545454545454,
      "grad_norm": 0.1298132985830307,
      "learning_rate": 0.00010490676743274181,
      "loss": 0.0238,
      "memory/device_reserved (GiB)": 33.2,
      "memory/max_active (GiB)": 30.44,
      "memory/max_allocated (GiB)": 30.44,
      "step": 34,
      "tokens_per_second_per_gpu": 1150.27
    },
    {
      "epoch": 1.5909090909090908,
      "grad_norm": 0.13134033977985382,
      "learning_rate": 0.0001,
      "loss": 0.0235,
      "memory/device_reserved (GiB)": 33.2,
      "memory/max_active (GiB)": 30.44,
      "memory/max_allocated (GiB)": 30.44,
      "step": 35,
      "tokens_per_second_per_gpu": 1191.31
    },
    {
      "epoch": 1.6363636363636362,
      "grad_norm": 0.11833763122558594,
      "learning_rate": 9.509323256725821e-05,
      "loss": 0.0221,
      "memory/device_reserved (GiB)": 33.2,
      "memory/max_active (GiB)": 30.44,
      "memory/max_allocated (GiB)": 30.44,
      "step": 36,
      "tokens_per_second_per_gpu": 1315.84
    },
    {
      "epoch": 1.6818181818181817,
      "grad_norm": 0.1161683201789856,
      "learning_rate": 9.019828596704394e-05,
      "loss": 0.021,
      "memory/device_reserved (GiB)": 33.2,
      "memory/max_active (GiB)": 30.44,
      "memory/max_allocated (GiB)": 30.44,
      "step": 37,
      "tokens_per_second_per_gpu": 1331.9
    },
    {
      "epoch": 1.7272727272727273,
      "grad_norm": 0.13690337538719177,
      "learning_rate": 8.532695255446383e-05,
      "loss": 0.024,
      "memory/device_reserved (GiB)": 33.2,
      "memory/max_active (GiB)": 30.44,
      "memory/max_allocated (GiB)": 30.44,
      "step": 38,
      "tokens_per_second_per_gpu": 1355.23
    },
    {
      "epoch": 1.7727272727272727,
      "grad_norm": 0.11100757122039795,
      "learning_rate": 8.049096779838719e-05,
      "loss": 0.0184,
      "memory/device_reserved (GiB)": 33.2,
      "memory/max_active (GiB)": 30.44,
      "memory/max_allocated (GiB)": 30.44,
      "step": 39,
      "tokens_per_second_per_gpu": 1252.87
    },
    {
      "epoch": 1.8181818181818183,
      "grad_norm": 0.11401506513357162,
      "learning_rate": 7.570198200967362e-05,
      "loss": 0.0211,
      "memory/device_reserved (GiB)": 33.2,
      "memory/max_active (GiB)": 30.44,
      "memory/max_allocated (GiB)": 30.44,
      "step": 40,
      "tokens_per_second_per_gpu": 1204.12
    },
    {
      "epoch": 1.8636363636363638,
      "grad_norm": 0.0962221622467041,
      "learning_rate": 7.097153227455379e-05,
      "loss": 0.0136,
      "memory/device_reserved (GiB)": 33.2,
      "memory/max_active (GiB)": 30.44,
      "memory/max_allocated (GiB)": 30.44,
      "step": 41,
      "tokens_per_second_per_gpu": 1390.74
    },
    {
      "epoch": 1.9090909090909092,
      "grad_norm": 0.0952286571264267,
      "learning_rate": 6.6311014660778e-05,
      "loss": 0.0166,
      "memory/device_reserved (GiB)": 33.2,
      "memory/max_active (GiB)": 30.44,
      "memory/max_allocated (GiB)": 30.44,
      "step": 42,
      "tokens_per_second_per_gpu": 1350.28
    },
    {
      "epoch": 1.9545454545454546,
      "grad_norm": 0.09817808121442795,
      "learning_rate": 6.173165676349103e-05,
      "loss": 0.016,
      "memory/device_reserved (GiB)": 33.2,
      "memory/max_active (GiB)": 30.44,
      "memory/max_allocated (GiB)": 30.44,
      "step": 43,
      "tokens_per_second_per_gpu": 1316.92
    },
    {
      "epoch": 2.0,
      "grad_norm": 0.1389278769493103,
      "learning_rate": 5.7244490656971815e-05,
      "loss": 0.0274,
      "memory/device_reserved (GiB)": 33.2,
      "memory/max_active (GiB)": 30.44,
      "memory/max_allocated (GiB)": 30.44,
      "step": 44,
      "tokens_per_second_per_gpu": 1031.7
    },
    {
      "epoch": 2.0454545454545454,
      "grad_norm": 0.07030683755874634,
      "learning_rate": 5.286032631740023e-05,
      "loss": 0.0126,
      "memory/device_reserved (GiB)": 33.2,
      "memory/max_active (GiB)": 30.44,
      "memory/max_allocated (GiB)": 30.44,
      "step": 45,
      "tokens_per_second_per_gpu": 1292.29
    },
    {
      "epoch": 2.090909090909091,
      "grad_norm": 0.09299763292074203,
      "learning_rate": 4.8589725580677835e-05,
      "loss": 0.0105,
      "memory/device_reserved (GiB)": 33.2,
      "memory/max_active (GiB)": 30.44,
      "memory/max_allocated (GiB)": 30.44,
      "step": 46,
      "tokens_per_second_per_gpu": 1355.66
    },
    {
      "epoch": 2.1363636363636362,
      "grad_norm": 0.186285600066185,
      "learning_rate": 4.444297669803981e-05,
      "loss": 0.0134,
      "memory/device_reserved (GiB)": 33.2,
      "memory/max_active (GiB)": 30.44,
      "memory/max_allocated (GiB)": 30.44,
      "step": 47,
      "tokens_per_second_per_gpu": 1198.91
    },
    {
      "epoch": 2.1818181818181817,
      "grad_norm": 0.06503765285015106,
      "learning_rate": 4.0430069550756665e-05,
      "loss": 0.0107,
      "memory/device_reserved (GiB)": 33.2,
      "memory/max_active (GiB)": 30.44,
      "memory/max_allocated (GiB)": 30.44,
      "step": 48,
      "tokens_per_second_per_gpu": 1293.9
    },
    {
      "epoch": 2.227272727272727,
      "grad_norm": 0.0671035498380661,
      "learning_rate": 3.6560671583635467e-05,
      "loss": 0.0082,
      "memory/device_reserved (GiB)": 33.2,
      "memory/max_active (GiB)": 30.44,
      "memory/max_allocated (GiB)": 30.44,
      "step": 49,
      "tokens_per_second_per_gpu": 1389.49
    },
    {
      "epoch": 2.2727272727272725,
      "grad_norm": 0.07254406809806824,
      "learning_rate": 3.2844104515298155e-05,
      "loss": 0.0106,
      "memory/device_reserved (GiB)": 33.2,
      "memory/max_active (GiB)": 30.44,
      "memory/max_allocated (GiB)": 30.44,
      "step": 50,
      "tokens_per_second_per_gpu": 1490.8
    }
  ],
  "logging_steps": 1,
  "max_steps": 66,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 3,
  "save_steps": 50,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": false
      },
      "attributes": {}
    }
  },
  "total_flos": 7.441454260341965e+16,
  "train_batch_size": 2,
  "trial_name": null,
  "trial_params": null
}