{
  "best_metric": 0.08136745542287827,
  "best_model_checkpoint": "model/checkpoint-8500",
  "epoch": 2.903997266826102,
  "eval_steps": 500,
  "global_step": 8500,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.17,
      "grad_norm": 1.9025481939315796,
      "learning_rate": 9.430588771210569e-06,
      "loss": 8.0821,
      "step": 500
    },
    {
      "epoch": 0.17,
      "eval_loss": 0.15024334192276,
      "eval_runtime": 39.7831,
      "eval_samples_per_second": 61.961,
      "eval_steps_per_second": 12.392,
      "step": 500
    },
    {
      "epoch": 0.34,
      "grad_norm": 0.37238627672195435,
      "learning_rate": 8.861177542421137e-06,
      "loss": 0.1353,
      "step": 1000
    },
    {
      "epoch": 0.34,
      "eval_loss": 0.09424333274364471,
      "eval_runtime": 39.7674,
      "eval_samples_per_second": 61.986,
      "eval_steps_per_second": 12.397,
      "step": 1000
    },
    {
      "epoch": 0.51,
      "grad_norm": 0.28026077151298523,
      "learning_rate": 8.291766313631705e-06,
      "loss": 0.1056,
      "step": 1500
    },
    {
      "epoch": 0.51,
      "eval_loss": 0.08929836004972458,
      "eval_runtime": 39.8141,
      "eval_samples_per_second": 61.913,
      "eval_steps_per_second": 12.383,
      "step": 1500
    },
    {
      "epoch": 0.68,
      "grad_norm": 0.242831751704216,
      "learning_rate": 7.722355084842274e-06,
      "loss": 0.0994,
      "step": 2000
    },
    {
      "epoch": 0.68,
      "eval_loss": 0.08687467128038406,
      "eval_runtime": 39.8819,
      "eval_samples_per_second": 61.808,
      "eval_steps_per_second": 12.362,
      "step": 2000
    },
    {
      "epoch": 0.85,
      "grad_norm": 0.31068557500839233,
      "learning_rate": 7.152943856052842e-06,
      "loss": 0.0957,
      "step": 2500
    },
    {
      "epoch": 0.85,
      "eval_loss": 0.08578581362962723,
      "eval_runtime": 39.8992,
      "eval_samples_per_second": 61.781,
      "eval_steps_per_second": 12.356,
      "step": 2500
    },
    {
      "epoch": 1.02,
      "grad_norm": 0.1952580064535141,
      "learning_rate": 6.583532627263411e-06,
      "loss": 0.0951,
      "step": 3000
    },
    {
      "epoch": 1.02,
      "eval_loss": 0.0845487117767334,
      "eval_runtime": 39.7723,
      "eval_samples_per_second": 61.978,
      "eval_steps_per_second": 12.396,
      "step": 3000
    },
    {
      "epoch": 1.2,
      "grad_norm": 0.1910097897052765,
      "learning_rate": 6.014121398473979e-06,
      "loss": 0.0923,
      "step": 3500
    },
    {
      "epoch": 1.2,
      "eval_loss": 0.08382081985473633,
      "eval_runtime": 39.606,
      "eval_samples_per_second": 62.238,
      "eval_steps_per_second": 12.448,
      "step": 3500
    },
    {
      "epoch": 1.37,
      "grad_norm": 0.18350045382976532,
      "learning_rate": 5.444710169684546e-06,
      "loss": 0.0899,
      "step": 4000
    },
    {
      "epoch": 1.37,
      "eval_loss": 0.08325745165348053,
      "eval_runtime": 39.6287,
      "eval_samples_per_second": 62.202,
      "eval_steps_per_second": 12.44,
      "step": 4000
    },
    {
      "epoch": 1.54,
      "grad_norm": 0.2158266305923462,
      "learning_rate": 4.8752989408951145e-06,
      "loss": 0.0899,
      "step": 4500
    },
    {
      "epoch": 1.54,
      "eval_loss": 0.08282188326120377,
      "eval_runtime": 39.6573,
      "eval_samples_per_second": 62.158,
      "eval_steps_per_second": 12.432,
      "step": 4500
    },
    {
      "epoch": 1.71,
      "grad_norm": 0.2279905378818512,
      "learning_rate": 4.305887712105683e-06,
      "loss": 0.0894,
      "step": 5000
    },
    {
      "epoch": 1.71,
      "eval_loss": 0.08228254318237305,
      "eval_runtime": 39.8586,
      "eval_samples_per_second": 61.844,
      "eval_steps_per_second": 12.369,
      "step": 5000
    },
    {
      "epoch": 1.88,
      "grad_norm": 0.18786436319351196,
      "learning_rate": 3.7364764833162513e-06,
      "loss": 0.089,
      "step": 5500
    },
    {
      "epoch": 1.88,
      "eval_loss": 0.08236683160066605,
      "eval_runtime": 39.8662,
      "eval_samples_per_second": 61.832,
      "eval_steps_per_second": 12.366,
      "step": 5500
    },
    {
      "epoch": 2.05,
      "grad_norm": 0.20143625140190125,
      "learning_rate": 3.1670652545268194e-06,
      "loss": 0.0882,
      "step": 6000
    },
    {
      "epoch": 2.05,
      "eval_loss": 0.08215593546628952,
      "eval_runtime": 39.8891,
      "eval_samples_per_second": 61.796,
      "eval_steps_per_second": 12.359,
      "step": 6000
    },
    {
      "epoch": 2.22,
      "grad_norm": 0.3449649512767792,
      "learning_rate": 2.5976540257373876e-06,
      "loss": 0.0876,
      "step": 6500
    },
    {
      "epoch": 2.22,
      "eval_loss": 0.08190125972032547,
      "eval_runtime": 39.9241,
      "eval_samples_per_second": 61.742,
      "eval_steps_per_second": 12.348,
      "step": 6500
    },
    {
      "epoch": 2.39,
      "grad_norm": 0.17690658569335938,
      "learning_rate": 2.028242796947956e-06,
      "loss": 0.0876,
      "step": 7000
    },
    {
      "epoch": 2.39,
      "eval_loss": 0.08178862929344177,
      "eval_runtime": 39.8989,
      "eval_samples_per_second": 61.781,
      "eval_steps_per_second": 12.356,
      "step": 7000
    },
    {
      "epoch": 2.56,
      "grad_norm": 0.16777564585208893,
      "learning_rate": 1.4588315681585242e-06,
      "loss": 0.0874,
      "step": 7500
    },
    {
      "epoch": 2.56,
      "eval_loss": 0.08147595077753067,
      "eval_runtime": 39.629,
      "eval_samples_per_second": 62.202,
      "eval_steps_per_second": 12.44,
      "step": 7500
    },
    {
      "epoch": 2.73,
      "grad_norm": 0.2061876803636551,
      "learning_rate": 8.894203393690924e-07,
      "loss": 0.088,
      "step": 8000
    },
    {
      "epoch": 2.73,
      "eval_loss": 0.08150825649499893,
      "eval_runtime": 39.6669,
      "eval_samples_per_second": 62.142,
      "eval_steps_per_second": 12.428,
      "step": 8000
    },
    {
      "epoch": 2.9,
      "grad_norm": 0.21220079064369202,
      "learning_rate": 3.2000911057966066e-07,
      "loss": 0.0871,
      "step": 8500
    },
    {
      "epoch": 2.9,
      "eval_loss": 0.08136745542287827,
      "eval_runtime": 39.796,
      "eval_samples_per_second": 61.941,
      "eval_steps_per_second": 12.388,
      "step": 8500
    }
  ],
  "logging_steps": 500,
  "max_steps": 8781,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 3,
  "save_steps": 500,
  "total_flos": 2.7281294636544e+16,
  "train_batch_size": 16,
  "trial_name": null,
  "trial_params": null
}