{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 8.605851979345955,
  "eval_steps": 500,
  "global_step": 5000,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.17,
      "grad_norm": 1.3079030513763428,
      "learning_rate": 4.000000000000001e-06,
      "loss": 0.4896,
      "step": 100
    },
    {
      "epoch": 0.34,
      "grad_norm": 1.4786765575408936,
      "learning_rate": 8.000000000000001e-06,
      "loss": 0.4113,
      "step": 200
    },
    {
      "epoch": 0.52,
      "grad_norm": 1.2072057723999023,
      "learning_rate": 1.2e-05,
      "loss": 0.3924,
      "step": 300
    },
    {
      "epoch": 0.69,
      "grad_norm": 1.5262311697006226,
      "learning_rate": 1.6000000000000003e-05,
      "loss": 0.3768,
      "step": 400
    },
    {
      "epoch": 0.86,
      "grad_norm": 1.114892840385437,
      "learning_rate": 2e-05,
      "loss": 0.353,
      "step": 500
    },
    {
      "epoch": 0.86,
      "eval_loss": 0.2806718051433563,
      "eval_runtime": 64.9966,
      "eval_samples_per_second": 15.385,
      "eval_steps_per_second": 3.846,
      "step": 500
    },
    {
      "epoch": 1.03,
      "grad_norm": 1.1068037748336792,
      "learning_rate": 1.9555555555555557e-05,
      "loss": 0.3373,
      "step": 600
    },
    {
      "epoch": 1.2,
      "grad_norm": 1.0337824821472168,
      "learning_rate": 1.9111111111111113e-05,
      "loss": 0.2773,
      "step": 700
    },
    {
      "epoch": 1.38,
      "grad_norm": 1.1779000759124756,
      "learning_rate": 1.866666666666667e-05,
      "loss": 0.2757,
      "step": 800
    },
    {
      "epoch": 1.55,
      "grad_norm": 1.2178924083709717,
      "learning_rate": 1.8222222222222224e-05,
      "loss": 0.2891,
      "step": 900
    },
    {
      "epoch": 1.72,
      "grad_norm": 1.1651198863983154,
      "learning_rate": 1.7777777777777777e-05,
      "loss": 0.2648,
      "step": 1000
    },
    {
      "epoch": 1.72,
      "eval_loss": 0.23123475909233093,
      "eval_runtime": 64.9606,
      "eval_samples_per_second": 15.394,
      "eval_steps_per_second": 3.848,
      "step": 1000
    },
    {
      "epoch": 1.89,
      "grad_norm": 0.941031813621521,
      "learning_rate": 1.7333333333333336e-05,
      "loss": 0.2556,
      "step": 1100
    },
    {
      "epoch": 2.07,
      "grad_norm": 1.5538479089736938,
      "learning_rate": 1.688888888888889e-05,
      "loss": 0.2188,
      "step": 1200
    },
    {
      "epoch": 2.24,
      "grad_norm": 1.7369763851165771,
      "learning_rate": 1.6444444444444444e-05,
      "loss": 0.1589,
      "step": 1300
    },
    {
      "epoch": 2.41,
      "grad_norm": 1.8593116998672485,
      "learning_rate": 1.6000000000000003e-05,
      "loss": 0.1526,
      "step": 1400
    },
    {
      "epoch": 2.58,
      "grad_norm": 0.9025155305862427,
      "learning_rate": 1.555555555555556e-05,
      "loss": 0.1512,
      "step": 1500
    },
    {
      "epoch": 2.58,
      "eval_loss": 0.23454324901103973,
      "eval_runtime": 64.9882,
      "eval_samples_per_second": 15.387,
      "eval_steps_per_second": 3.847,
      "step": 1500
    },
    {
      "epoch": 2.75,
      "grad_norm": 1.2440764904022217,
      "learning_rate": 1.5111111111111112e-05,
      "loss": 0.1596,
      "step": 1600
    },
    {
      "epoch": 2.93,
      "grad_norm": 0.9365782141685486,
      "learning_rate": 1.4666666666666666e-05,
      "loss": 0.1603,
      "step": 1700
    },
    {
      "epoch": 3.1,
      "grad_norm": 0.977448582649231,
      "learning_rate": 1.4222222222222224e-05,
      "loss": 0.1086,
      "step": 1800
    },
    {
      "epoch": 3.27,
      "grad_norm": 0.9493745565414429,
      "learning_rate": 1.377777777777778e-05,
      "loss": 0.0822,
      "step": 1900
    },
    {
      "epoch": 3.44,
      "grad_norm": 1.437554955482483,
      "learning_rate": 1.3333333333333333e-05,
      "loss": 0.0838,
      "step": 2000
    },
    {
      "epoch": 3.44,
      "eval_loss": 0.258994460105896,
      "eval_runtime": 65.0096,
      "eval_samples_per_second": 15.382,
      "eval_steps_per_second": 3.846,
      "step": 2000
    },
    {
      "epoch": 3.61,
      "grad_norm": 1.1978514194488525,
      "learning_rate": 1.288888888888889e-05,
      "loss": 0.0849,
      "step": 2100
    },
    {
      "epoch": 3.79,
      "grad_norm": 1.1100072860717773,
      "learning_rate": 1.2444444444444446e-05,
      "loss": 0.0822,
      "step": 2200
    },
    {
      "epoch": 3.96,
      "grad_norm": 1.0157575607299805,
      "learning_rate": 1.2e-05,
      "loss": 0.0799,
      "step": 2300
    },
    {
      "epoch": 4.13,
      "grad_norm": 1.1923385858535767,
      "learning_rate": 1.1555555555555556e-05,
      "loss": 0.05,
      "step": 2400
    },
    {
      "epoch": 4.3,
      "grad_norm": 1.1894739866256714,
      "learning_rate": 1.1111111111111113e-05,
      "loss": 0.0401,
      "step": 2500
    },
    {
      "epoch": 4.3,
      "eval_loss": 0.3230312466621399,
      "eval_runtime": 65.0356,
      "eval_samples_per_second": 15.376,
      "eval_steps_per_second": 3.844,
      "step": 2500
    },
    {
      "epoch": 4.48,
      "grad_norm": 1.5902525186538696,
      "learning_rate": 1.0666666666666667e-05,
      "loss": 0.0404,
      "step": 2600
    },
    {
      "epoch": 4.65,
      "grad_norm": 1.3404467105865479,
      "learning_rate": 1.0222222222222223e-05,
      "loss": 0.0418,
      "step": 2700
    },
    {
      "epoch": 4.82,
      "grad_norm": 1.4325445890426636,
      "learning_rate": 9.777777777777779e-06,
      "loss": 0.0441,
      "step": 2800
    },
    {
      "epoch": 4.99,
      "grad_norm": 1.3849271535873413,
      "learning_rate": 9.333333333333334e-06,
      "loss": 0.0429,
      "step": 2900
    },
    {
      "epoch": 5.16,
      "grad_norm": 1.0262218713760376,
      "learning_rate": 8.888888888888888e-06,
      "loss": 0.0196,
      "step": 3000
    },
    {
      "epoch": 5.16,
      "eval_loss": 0.38104376196861267,
      "eval_runtime": 64.9932,
      "eval_samples_per_second": 15.386,
      "eval_steps_per_second": 3.847,
      "step": 3000
    },
    {
      "epoch": 5.34,
      "grad_norm": 1.4700788259506226,
      "learning_rate": 8.444444444444446e-06,
      "loss": 0.0205,
      "step": 3100
    },
    {
      "epoch": 5.51,
      "grad_norm": 0.9892378449440002,
      "learning_rate": 8.000000000000001e-06,
      "loss": 0.0207,
      "step": 3200
    },
    {
      "epoch": 5.68,
      "grad_norm": 0.4820214509963989,
      "learning_rate": 7.555555555555556e-06,
      "loss": 0.0204,
      "step": 3300
    },
    {
      "epoch": 5.85,
      "grad_norm": 1.2511119842529297,
      "learning_rate": 7.111111111111112e-06,
      "loss": 0.0228,
      "step": 3400
    },
    {
      "epoch": 6.02,
      "grad_norm": 0.32089540362358093,
      "learning_rate": 6.666666666666667e-06,
      "loss": 0.0197,
      "step": 3500
    },
    {
      "epoch": 6.02,
      "eval_loss": 0.3612504303455353,
      "eval_runtime": 65.1059,
      "eval_samples_per_second": 15.36,
      "eval_steps_per_second": 3.84,
      "step": 3500
    },
    {
      "epoch": 6.2,
      "grad_norm": 0.4687187671661377,
      "learning_rate": 6.222222222222223e-06,
      "loss": 0.0085,
      "step": 3600
    },
    {
      "epoch": 6.37,
      "grad_norm": 0.696907103061676,
      "learning_rate": 5.777777777777778e-06,
      "loss": 0.0096,
      "step": 3700
    },
    {
      "epoch": 6.54,
      "grad_norm": 0.5151516795158386,
      "learning_rate": 5.333333333333334e-06,
      "loss": 0.009,
      "step": 3800
    },
    {
      "epoch": 6.71,
      "grad_norm": 0.8852140307426453,
      "learning_rate": 4.888888888888889e-06,
      "loss": 0.012,
      "step": 3900
    },
    {
      "epoch": 6.88,
      "grad_norm": 0.8389159440994263,
      "learning_rate": 4.444444444444444e-06,
      "loss": 0.0093,
      "step": 4000
    },
    {
      "epoch": 6.88,
      "eval_loss": 0.3735957741737366,
      "eval_runtime": 65.1742,
      "eval_samples_per_second": 15.343,
      "eval_steps_per_second": 3.836,
      "step": 4000
    },
    {
      "epoch": 7.06,
      "grad_norm": 0.158839613199234,
      "learning_rate": 4.000000000000001e-06,
      "loss": 0.0071,
      "step": 4100
    },
    {
      "epoch": 7.23,
      "grad_norm": 0.4010579586029053,
      "learning_rate": 3.555555555555556e-06,
      "loss": 0.0047,
      "step": 4200
    },
    {
      "epoch": 7.4,
      "grad_norm": 0.16058029234409332,
      "learning_rate": 3.1111111111111116e-06,
      "loss": 0.0037,
      "step": 4300
    },
    {
      "epoch": 7.57,
      "grad_norm": 0.1508949100971222,
      "learning_rate": 2.666666666666667e-06,
      "loss": 0.0039,
      "step": 4400
    },
    {
      "epoch": 7.75,
      "grad_norm": 0.3169790804386139,
      "learning_rate": 2.222222222222222e-06,
      "loss": 0.0037,
      "step": 4500
    },
    {
      "epoch": 7.75,
      "eval_loss": 0.38906845450401306,
      "eval_runtime": 65.0854,
      "eval_samples_per_second": 15.364,
      "eval_steps_per_second": 3.841,
      "step": 4500
    },
    {
      "epoch": 7.92,
      "grad_norm": 0.46604883670806885,
      "learning_rate": 1.777777777777778e-06,
      "loss": 0.0043,
      "step": 4600
    },
    {
      "epoch": 8.09,
      "grad_norm": 0.08523246645927429,
      "learning_rate": 1.3333333333333334e-06,
      "loss": 0.0042,
      "step": 4700
    },
    {
      "epoch": 8.26,
      "grad_norm": 0.2368686944246292,
      "learning_rate": 8.88888888888889e-07,
      "loss": 0.0032,
      "step": 4800
    },
    {
      "epoch": 8.43,
      "grad_norm": 0.11481987684965134,
      "learning_rate": 4.444444444444445e-07,
      "loss": 0.0019,
      "step": 4900
    },
    {
      "epoch": 8.61,
      "grad_norm": 0.13208821415901184,
      "learning_rate": 0.0,
      "loss": 0.002,
      "step": 5000
    },
    {
      "epoch": 8.61,
      "eval_loss": 0.4305289387702942,
      "eval_runtime": 65.0773,
      "eval_samples_per_second": 15.366,
      "eval_steps_per_second": 3.842,
      "step": 5000
    }
  ],
  "logging_steps": 100,
  "max_steps": 5000,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 9,
  "save_steps": 500,
  "total_flos": 1.258569996863275e+18,
  "train_batch_size": 4,
  "trial_name": null,
  "trial_params": null
}