{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 1.5665387326701654,
  "eval_steps": 100,
  "global_step": 5000,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.03,
      "grad_norm": 7.659268856048584,
      "learning_rate": 4.000000000000001e-06,
      "loss": 1.3919,
      "step": 100
    },
    {
      "epoch": 0.03,
      "eval_loss": 1.1893327236175537,
      "eval_runtime": 192.797,
      "eval_samples_per_second": 5.187,
      "eval_steps_per_second": 5.187,
      "step": 100
    },
    {
      "epoch": 0.06,
      "grad_norm": 3.593693971633911,
      "learning_rate": 8.000000000000001e-06,
      "loss": 0.7649,
      "step": 200
    },
    {
      "epoch": 0.06,
      "eval_loss": 0.37141919136047363,
      "eval_runtime": 193.0277,
      "eval_samples_per_second": 5.181,
      "eval_steps_per_second": 5.181,
      "step": 200
    },
    {
      "epoch": 0.09,
      "grad_norm": 0.2718026638031006,
      "learning_rate": 1.2e-05,
      "loss": 0.2492,
      "step": 300
    },
    {
      "epoch": 0.09,
      "eval_loss": 0.1235867515206337,
      "eval_runtime": 192.4574,
      "eval_samples_per_second": 5.196,
      "eval_steps_per_second": 5.196,
      "step": 300
    },
    {
      "epoch": 0.13,
      "grad_norm": 0.3505817651748657,
      "learning_rate": 1.6000000000000003e-05,
      "loss": 0.1111,
      "step": 400
    },
    {
      "epoch": 0.13,
      "eval_loss": 0.06019064038991928,
      "eval_runtime": 193.0151,
      "eval_samples_per_second": 5.181,
      "eval_steps_per_second": 5.181,
      "step": 400
    },
    {
      "epoch": 0.16,
      "grad_norm": 2.2099411487579346,
      "learning_rate": 2e-05,
      "loss": 0.1213,
      "step": 500
    },
    {
      "epoch": 0.16,
      "eval_loss": 0.05992453545331955,
      "eval_runtime": 193.5019,
      "eval_samples_per_second": 5.168,
      "eval_steps_per_second": 5.168,
      "step": 500
    },
    {
      "epoch": 0.19,
      "grad_norm": 0.006523499730974436,
      "learning_rate": 1.9555555555555557e-05,
      "loss": 0.1161,
      "step": 600
    },
    {
      "epoch": 0.19,
      "eval_loss": 0.03172233700752258,
      "eval_runtime": 192.8437,
      "eval_samples_per_second": 5.186,
      "eval_steps_per_second": 5.186,
      "step": 600
    },
    {
      "epoch": 0.22,
      "grad_norm": 0.0743604227900505,
      "learning_rate": 1.9111111111111113e-05,
      "loss": 0.0853,
      "step": 700
    },
    {
      "epoch": 0.22,
      "eval_loss": 0.03372843191027641,
      "eval_runtime": 193.2798,
      "eval_samples_per_second": 5.174,
      "eval_steps_per_second": 5.174,
      "step": 700
    },
    {
      "epoch": 0.25,
      "grad_norm": 2.5127480030059814,
      "learning_rate": 1.866666666666667e-05,
      "loss": 0.0801,
      "step": 800
    },
    {
      "epoch": 0.25,
      "eval_loss": 0.02089008316397667,
      "eval_runtime": 193.1941,
      "eval_samples_per_second": 5.176,
      "eval_steps_per_second": 5.176,
      "step": 800
    },
    {
      "epoch": 0.28,
      "grad_norm": 2.219334125518799,
      "learning_rate": 1.8222222222222224e-05,
      "loss": 0.0619,
      "step": 900
    },
    {
      "epoch": 0.28,
      "eval_loss": 0.03212675824761391,
      "eval_runtime": 192.8222,
      "eval_samples_per_second": 5.186,
      "eval_steps_per_second": 5.186,
      "step": 900
    },
    {
      "epoch": 0.31,
      "grad_norm": 0.01359077449887991,
      "learning_rate": 1.7777777777777777e-05,
      "loss": 0.0707,
      "step": 1000
    },
    {
      "epoch": 0.31,
      "eval_loss": 0.0244381632655859,
      "eval_runtime": 192.3716,
      "eval_samples_per_second": 5.198,
      "eval_steps_per_second": 5.198,
      "step": 1000
    },
    {
      "epoch": 0.34,
      "grad_norm": 0.007412207778543234,
      "learning_rate": 1.7333333333333336e-05,
      "loss": 0.0631,
      "step": 1100
    },
    {
      "epoch": 0.34,
      "eval_loss": 0.025782734155654907,
      "eval_runtime": 192.5677,
      "eval_samples_per_second": 5.193,
      "eval_steps_per_second": 5.193,
      "step": 1100
    },
    {
      "epoch": 0.38,
      "grad_norm": 0.004010587465018034,
      "learning_rate": 1.688888888888889e-05,
      "loss": 0.0645,
      "step": 1200
    },
    {
      "epoch": 0.38,
      "eval_loss": 0.026380345225334167,
      "eval_runtime": 192.1928,
      "eval_samples_per_second": 5.203,
      "eval_steps_per_second": 5.203,
      "step": 1200
    },
    {
      "epoch": 0.41,
      "grad_norm": 0.0805005207657814,
      "learning_rate": 1.6444444444444444e-05,
      "loss": 0.0419,
      "step": 1300
    },
    {
      "epoch": 0.41,
      "eval_loss": 0.027041926980018616,
      "eval_runtime": 192.535,
      "eval_samples_per_second": 5.194,
      "eval_steps_per_second": 5.194,
      "step": 1300
    },
    {
      "epoch": 0.44,
      "grad_norm": 0.010087325237691402,
      "learning_rate": 1.6000000000000003e-05,
      "loss": 0.0612,
      "step": 1400
    },
    {
      "epoch": 0.44,
      "eval_loss": 0.018431425094604492,
      "eval_runtime": 192.3449,
      "eval_samples_per_second": 5.199,
      "eval_steps_per_second": 5.199,
      "step": 1400
    },
    {
      "epoch": 0.47,
      "grad_norm": 0.6794092655181885,
      "learning_rate": 1.555555555555556e-05,
      "loss": 0.0569,
      "step": 1500
    },
    {
      "epoch": 0.47,
      "eval_loss": 0.017885498702526093,
      "eval_runtime": 192.6531,
      "eval_samples_per_second": 5.191,
      "eval_steps_per_second": 5.191,
      "step": 1500
    },
    {
      "epoch": 0.5,
      "grad_norm": 1.6970280408859253,
      "learning_rate": 1.5111111111111112e-05,
      "loss": 0.0546,
      "step": 1600
    },
    {
      "epoch": 0.5,
      "eval_loss": 0.022140460088849068,
      "eval_runtime": 192.8968,
      "eval_samples_per_second": 5.184,
      "eval_steps_per_second": 5.184,
      "step": 1600
    },
    {
      "epoch": 0.53,
      "grad_norm": 0.013943832367658615,
      "learning_rate": 1.4666666666666666e-05,
      "loss": 0.069,
      "step": 1700
    },
    {
      "epoch": 0.53,
      "eval_loss": 0.020227737724781036,
      "eval_runtime": 193.1829,
      "eval_samples_per_second": 5.176,
      "eval_steps_per_second": 5.176,
      "step": 1700
    },
    {
      "epoch": 0.56,
      "grad_norm": 2.067197322845459,
      "learning_rate": 1.4222222222222224e-05,
      "loss": 0.0509,
      "step": 1800
    },
    {
      "epoch": 0.56,
      "eval_loss": 0.01612325944006443,
      "eval_runtime": 192.6314,
      "eval_samples_per_second": 5.191,
      "eval_steps_per_second": 5.191,
      "step": 1800
    },
    {
      "epoch": 0.6,
      "grad_norm": 2.2480263710021973,
      "learning_rate": 1.377777777777778e-05,
      "loss": 0.0495,
      "step": 1900
    },
    {
      "epoch": 0.6,
      "eval_loss": 0.01796303130686283,
      "eval_runtime": 192.3154,
      "eval_samples_per_second": 5.2,
      "eval_steps_per_second": 5.2,
      "step": 1900
    },
    {
      "epoch": 0.63,
      "grad_norm": 0.0029044542461633682,
      "learning_rate": 1.3333333333333333e-05,
      "loss": 0.0444,
      "step": 2000
    },
    {
      "epoch": 0.63,
      "eval_loss": 0.02335376851260662,
      "eval_runtime": 192.3608,
      "eval_samples_per_second": 5.199,
      "eval_steps_per_second": 5.199,
      "step": 2000
    },
    {
      "epoch": 0.66,
      "grad_norm": 1.1315308809280396,
      "learning_rate": 1.288888888888889e-05,
      "loss": 0.0523,
      "step": 2100
    },
    {
      "epoch": 0.66,
      "eval_loss": 0.019370460882782936,
      "eval_runtime": 192.4625,
      "eval_samples_per_second": 5.196,
      "eval_steps_per_second": 5.196,
      "step": 2100
    },
    {
      "epoch": 0.69,
      "grad_norm": 0.0026583941653370857,
      "learning_rate": 1.2444444444444446e-05,
      "loss": 0.0553,
      "step": 2200
    },
    {
      "epoch": 0.69,
      "eval_loss": 0.01725778914988041,
      "eval_runtime": 192.3261,
      "eval_samples_per_second": 5.2,
      "eval_steps_per_second": 5.2,
      "step": 2200
    },
    {
      "epoch": 0.72,
      "grad_norm": 0.004530389327555895,
      "learning_rate": 1.2e-05,
      "loss": 0.0538,
      "step": 2300
    },
    {
      "epoch": 0.72,
      "eval_loss": 0.022032534703612328,
      "eval_runtime": 192.3926,
      "eval_samples_per_second": 5.198,
      "eval_steps_per_second": 5.198,
      "step": 2300
    },
    {
      "epoch": 0.75,
      "grad_norm": 0.0024464745074510574,
      "learning_rate": 1.1555555555555556e-05,
      "loss": 0.0456,
      "step": 2400
    },
    {
      "epoch": 0.75,
      "eval_loss": 0.030184298753738403,
      "eval_runtime": 192.2035,
      "eval_samples_per_second": 5.203,
      "eval_steps_per_second": 5.203,
      "step": 2400
    },
    {
      "epoch": 0.78,
      "grad_norm": 0.0019896693993359804,
      "learning_rate": 1.1111111111111113e-05,
      "loss": 0.0499,
      "step": 2500
    },
    {
      "epoch": 0.78,
      "eval_loss": 0.01628696359694004,
      "eval_runtime": 192.4383,
      "eval_samples_per_second": 5.196,
      "eval_steps_per_second": 5.196,
      "step": 2500
    },
    {
      "epoch": 0.81,
      "grad_norm": 0.001884658238850534,
      "learning_rate": 1.0666666666666667e-05,
      "loss": 0.0493,
      "step": 2600
    },
    {
      "epoch": 0.81,
      "eval_loss": 0.025292817503213882,
      "eval_runtime": 192.2059,
      "eval_samples_per_second": 5.203,
      "eval_steps_per_second": 5.203,
      "step": 2600
    },
    {
      "epoch": 0.85,
      "grad_norm": 0.04386008903384209,
      "learning_rate": 1.0222222222222223e-05,
      "loss": 0.0452,
      "step": 2700
    },
    {
      "epoch": 0.85,
      "eval_loss": 0.022591974586248398,
      "eval_runtime": 192.8163,
      "eval_samples_per_second": 5.186,
      "eval_steps_per_second": 5.186,
      "step": 2700
    },
    {
      "epoch": 0.88,
      "grad_norm": 0.0019288246985524893,
      "learning_rate": 9.777777777777779e-06,
      "loss": 0.0438,
      "step": 2800
    },
    {
      "epoch": 0.88,
      "eval_loss": 0.0226399265229702,
      "eval_runtime": 192.6394,
      "eval_samples_per_second": 5.191,
      "eval_steps_per_second": 5.191,
      "step": 2800
    },
    {
      "epoch": 0.91,
      "grad_norm": 0.056311335414648056,
      "learning_rate": 9.333333333333334e-06,
      "loss": 0.0579,
      "step": 2900
    },
    {
      "epoch": 0.91,
      "eval_loss": 0.02137078531086445,
      "eval_runtime": 192.3652,
      "eval_samples_per_second": 5.198,
      "eval_steps_per_second": 5.198,
      "step": 2900
    },
    {
      "epoch": 0.94,
      "grad_norm": 0.003287563333287835,
      "learning_rate": 8.888888888888888e-06,
      "loss": 0.036,
      "step": 3000
    },
    {
      "epoch": 0.94,
      "eval_loss": 0.015787875279784203,
      "eval_runtime": 192.7182,
      "eval_samples_per_second": 5.189,
      "eval_steps_per_second": 5.189,
      "step": 3000
    },
    {
      "epoch": 0.97,
      "grad_norm": 0.0015150770777836442,
      "learning_rate": 8.444444444444446e-06,
      "loss": 0.055,
      "step": 3100
    },
    {
      "epoch": 0.97,
      "eval_loss": 0.020349696278572083,
      "eval_runtime": 192.752,
      "eval_samples_per_second": 5.188,
      "eval_steps_per_second": 5.188,
      "step": 3100
    },
    {
      "epoch": 1.0,
      "grad_norm": 0.8284673690795898,
      "learning_rate": 8.000000000000001e-06,
      "loss": 0.0424,
      "step": 3200
    },
    {
      "epoch": 1.0,
      "eval_loss": 0.011587778106331825,
      "eval_runtime": 192.4082,
      "eval_samples_per_second": 5.197,
      "eval_steps_per_second": 5.197,
      "step": 3200
    },
    {
      "epoch": 1.03,
      "grad_norm": 0.00324226007796824,
      "learning_rate": 7.555555555555556e-06,
      "loss": 0.0232,
      "step": 3300
    },
    {
      "epoch": 1.03,
      "eval_loss": 0.02541309781372547,
      "eval_runtime": 192.4753,
      "eval_samples_per_second": 5.195,
      "eval_steps_per_second": 5.195,
      "step": 3300
    },
    {
      "epoch": 1.07,
      "grad_norm": 0.0018368299352005124,
      "learning_rate": 7.111111111111112e-06,
      "loss": 0.0391,
      "step": 3400
    },
    {
      "epoch": 1.07,
      "eval_loss": 0.019817600026726723,
      "eval_runtime": 192.2908,
      "eval_samples_per_second": 5.2,
      "eval_steps_per_second": 5.2,
      "step": 3400
    },
    {
      "epoch": 1.1,
      "grad_norm": 0.001375267980620265,
      "learning_rate": 6.666666666666667e-06,
      "loss": 0.029,
      "step": 3500
    },
    {
      "epoch": 1.1,
      "eval_loss": 0.014760646037757397,
      "eval_runtime": 192.1713,
      "eval_samples_per_second": 5.204,
      "eval_steps_per_second": 5.204,
      "step": 3500
    },
    {
      "epoch": 1.13,
      "grad_norm": 1.2048271894454956,
      "learning_rate": 6.222222222222223e-06,
      "loss": 0.028,
      "step": 3600
    },
    {
      "epoch": 1.13,
      "eval_loss": 0.024096647277474403,
      "eval_runtime": 191.7597,
      "eval_samples_per_second": 5.215,
      "eval_steps_per_second": 5.215,
      "step": 3600
    },
    {
      "epoch": 1.16,
      "grad_norm": 1.7010436058044434,
      "learning_rate": 5.777777777777778e-06,
      "loss": 0.033,
      "step": 3700
    },
    {
      "epoch": 1.16,
      "eval_loss": 0.024101875722408295,
      "eval_runtime": 191.6566,
      "eval_samples_per_second": 5.218,
      "eval_steps_per_second": 5.218,
      "step": 3700
    },
    {
      "epoch": 1.19,
      "grad_norm": 0.4044632613658905,
      "learning_rate": 5.333333333333334e-06,
      "loss": 0.0411,
      "step": 3800
    },
    {
      "epoch": 1.19,
      "eval_loss": 0.020846880972385406,
      "eval_runtime": 192.1837,
      "eval_samples_per_second": 5.203,
      "eval_steps_per_second": 5.203,
      "step": 3800
    },
    {
      "epoch": 1.22,
      "grad_norm": 0.47499576210975647,
      "learning_rate": 4.888888888888889e-06,
      "loss": 0.0233,
      "step": 3900
    },
    {
      "epoch": 1.22,
      "eval_loss": 0.010989435017108917,
      "eval_runtime": 192.4351,
      "eval_samples_per_second": 5.197,
      "eval_steps_per_second": 5.197,
      "step": 3900
    },
    {
      "epoch": 1.25,
      "grad_norm": 0.017001571133732796,
      "learning_rate": 4.444444444444444e-06,
      "loss": 0.0204,
      "step": 4000
    },
    {
      "epoch": 1.25,
      "eval_loss": 0.02116994932293892,
      "eval_runtime": 192.8643,
      "eval_samples_per_second": 5.185,
      "eval_steps_per_second": 5.185,
      "step": 4000
    },
    {
      "epoch": 1.28,
      "grad_norm": 0.0578785166144371,
      "learning_rate": 4.004444444444445e-06,
      "loss": 0.0378,
      "step": 4100
    },
    {
      "epoch": 1.28,
      "eval_loss": 0.018333839252591133,
      "eval_runtime": 192.9576,
      "eval_samples_per_second": 5.182,
      "eval_steps_per_second": 5.182,
      "step": 4100
    },
    {
      "epoch": 1.32,
      "grad_norm": 0.0014218598371371627,
      "learning_rate": 3.5600000000000002e-06,
      "loss": 0.0289,
      "step": 4200
    },
    {
      "epoch": 1.32,
      "eval_loss": 0.02419031597673893,
      "eval_runtime": 192.6293,
      "eval_samples_per_second": 5.191,
      "eval_steps_per_second": 5.191,
      "step": 4200
    },
    {
      "epoch": 1.35,
      "grad_norm": 0.0013137555215507746,
      "learning_rate": 3.1155555555555555e-06,
      "loss": 0.0298,
      "step": 4300
    },
    {
      "epoch": 1.35,
      "eval_loss": 0.02638879045844078,
      "eval_runtime": 192.8273,
      "eval_samples_per_second": 5.186,
      "eval_steps_per_second": 5.186,
      "step": 4300
    },
    {
      "epoch": 1.38,
      "grad_norm": 0.35259732604026794,
      "learning_rate": 2.6711111111111116e-06,
      "loss": 0.0382,
      "step": 4400
    },
    {
      "epoch": 1.38,
      "eval_loss": 0.01607164740562439,
      "eval_runtime": 192.7408,
      "eval_samples_per_second": 5.188,
      "eval_steps_per_second": 5.188,
      "step": 4400
    },
    {
      "epoch": 1.41,
      "grad_norm": 0.0020605421159416437,
      "learning_rate": 2.226666666666667e-06,
      "loss": 0.0339,
      "step": 4500
    },
    {
      "epoch": 1.41,
      "eval_loss": 0.014907135628163815,
      "eval_runtime": 192.8289,
      "eval_samples_per_second": 5.186,
      "eval_steps_per_second": 5.186,
      "step": 4500
    },
    {
      "epoch": 1.44,
      "grad_norm": 0.0019016048172488809,
      "learning_rate": 1.7822222222222225e-06,
      "loss": 0.0195,
      "step": 4600
    },
    {
      "epoch": 1.44,
      "eval_loss": 0.015925556421279907,
      "eval_runtime": 192.4215,
      "eval_samples_per_second": 5.197,
      "eval_steps_per_second": 5.197,
      "step": 4600
    },
    {
      "epoch": 1.47,
      "grad_norm": 0.734219491481781,
      "learning_rate": 1.337777777777778e-06,
      "loss": 0.0328,
      "step": 4700
    },
    {
      "epoch": 1.47,
      "eval_loss": 0.014890914782881737,
      "eval_runtime": 192.4767,
      "eval_samples_per_second": 5.195,
      "eval_steps_per_second": 5.195,
      "step": 4700
    },
    {
      "epoch": 1.5,
      "grad_norm": 2.157243251800537,
      "learning_rate": 8.933333333333334e-07,
      "loss": 0.0429,
      "step": 4800
    },
    {
      "epoch": 1.5,
      "eval_loss": 0.01486047450453043,
      "eval_runtime": 191.7644,
      "eval_samples_per_second": 5.215,
      "eval_steps_per_second": 5.215,
      "step": 4800
    },
    {
      "epoch": 1.54,
      "grad_norm": 0.00161929486785084,
      "learning_rate": 4.488888888888889e-07,
      "loss": 0.0312,
      "step": 4900
    },
    {
      "epoch": 1.54,
      "eval_loss": 0.02127786912024021,
      "eval_runtime": 191.9957,
      "eval_samples_per_second": 5.208,
      "eval_steps_per_second": 5.208,
      "step": 4900
    },
    {
      "epoch": 1.57,
      "grad_norm": 0.001571273198351264,
      "learning_rate": 4.444444444444445e-09,
      "loss": 0.0364,
      "step": 5000
    },
    {
      "epoch": 1.57,
      "eval_loss": 0.01901436597108841,
      "eval_runtime": 191.5985,
      "eval_samples_per_second": 5.219,
      "eval_steps_per_second": 5.219,
      "step": 5000
    }
  ],
  "logging_steps": 100,
  "max_steps": 5000,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 2,
  "save_steps": 1000,
  "total_flos": 8.051062996992e+16,
  "train_batch_size": 1,
  "trial_name": null,
  "trial_params": null
}