{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 16.46840148698885,
  "eval_steps": 500,
  "global_step": 8860,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.93,
      "grad_norm": 1.3550214767456055,
      "learning_rate": 2.8605947955390337e-05,
      "loss": 0.6834,
      "step": 500
    },
    {
      "epoch": 0.93,
      "eval_loss": 0.5703962445259094,
      "eval_runtime": 6.1216,
      "eval_samples_per_second": 248.138,
      "eval_steps_per_second": 15.519,
      "step": 500
    },
    {
      "epoch": 1.86,
      "grad_norm": 0.7966482043266296,
      "learning_rate": 2.721189591078067e-05,
      "loss": 0.5019,
      "step": 1000
    },
    {
      "epoch": 1.86,
      "eval_loss": 0.5145730972290039,
      "eval_runtime": 6.1234,
      "eval_samples_per_second": 248.066,
      "eval_steps_per_second": 15.514,
      "step": 1000
    },
    {
      "epoch": 2.79,
      "grad_norm": 0.9648380875587463,
      "learning_rate": 2.5817843866171006e-05,
      "loss": 0.4522,
      "step": 1500
    },
    {
      "epoch": 2.79,
      "eval_loss": 0.48453015089035034,
      "eval_runtime": 6.0895,
      "eval_samples_per_second": 249.446,
      "eval_steps_per_second": 15.601,
      "step": 1500
    },
    {
      "epoch": 3.72,
      "grad_norm": 0.7465741038322449,
      "learning_rate": 2.442379182156134e-05,
      "loss": 0.4105,
      "step": 2000
    },
    {
      "epoch": 3.72,
      "eval_loss": 0.4673249125480652,
      "eval_runtime": 6.11,
      "eval_samples_per_second": 248.609,
      "eval_steps_per_second": 15.548,
      "step": 2000
    },
    {
      "epoch": 4.65,
      "grad_norm": 0.8333841562271118,
      "learning_rate": 2.302973977695167e-05,
      "loss": 0.3705,
      "step": 2500
    },
    {
      "epoch": 4.65,
      "eval_loss": 0.4610813856124878,
      "eval_runtime": 6.0863,
      "eval_samples_per_second": 249.578,
      "eval_steps_per_second": 15.609,
      "step": 2500
    },
    {
      "epoch": 5.58,
      "grad_norm": 0.8087735176086426,
      "learning_rate": 2.1635687732342008e-05,
      "loss": 0.3484,
      "step": 3000
    },
    {
      "epoch": 5.58,
      "eval_loss": 0.44604435563087463,
      "eval_runtime": 6.1053,
      "eval_samples_per_second": 248.8,
      "eval_steps_per_second": 15.56,
      "step": 3000
    },
    {
      "epoch": 6.51,
      "grad_norm": 0.8467231392860413,
      "learning_rate": 2.0241635687732344e-05,
      "loss": 0.33,
      "step": 3500
    },
    {
      "epoch": 6.51,
      "eval_loss": 0.4419492185115814,
      "eval_runtime": 6.1029,
      "eval_samples_per_second": 248.896,
      "eval_steps_per_second": 15.566,
      "step": 3500
    },
    {
      "epoch": 7.43,
      "grad_norm": 0.9491092562675476,
      "learning_rate": 1.8847583643122676e-05,
      "loss": 0.3068,
      "step": 4000
    },
    {
      "epoch": 7.43,
      "eval_loss": 0.4374473989009857,
      "eval_runtime": 6.0863,
      "eval_samples_per_second": 249.578,
      "eval_steps_per_second": 15.609,
      "step": 4000
    },
    {
      "epoch": 8.36,
      "grad_norm": 0.5728708505630493,
      "learning_rate": 1.7453531598513013e-05,
      "loss": 0.2877,
      "step": 4500
    },
    {
      "epoch": 8.36,
      "eval_loss": 0.4360838234424591,
      "eval_runtime": 6.109,
      "eval_samples_per_second": 248.648,
      "eval_steps_per_second": 15.551,
      "step": 4500
    },
    {
      "epoch": 9.29,
      "grad_norm": 0.9121451377868652,
      "learning_rate": 1.6059479553903345e-05,
      "loss": 0.2774,
      "step": 5000
    },
    {
      "epoch": 9.29,
      "eval_loss": 0.4340021312236786,
      "eval_runtime": 6.1145,
      "eval_samples_per_second": 248.424,
      "eval_steps_per_second": 15.537,
      "step": 5000
    },
    {
      "epoch": 10.22,
      "grad_norm": 0.6917915344238281,
      "learning_rate": 1.4665427509293682e-05,
      "loss": 0.2643,
      "step": 5500
    },
    {
      "epoch": 10.22,
      "eval_loss": 0.4294339716434479,
      "eval_runtime": 6.1237,
      "eval_samples_per_second": 248.054,
      "eval_steps_per_second": 15.514,
      "step": 5500
    },
    {
      "epoch": 11.15,
      "grad_norm": 0.5932989120483398,
      "learning_rate": 1.3271375464684014e-05,
      "loss": 0.2521,
      "step": 6000
    },
    {
      "epoch": 11.15,
      "eval_loss": 0.429922491312027,
      "eval_runtime": 6.1392,
      "eval_samples_per_second": 247.428,
      "eval_steps_per_second": 15.474,
      "step": 6000
    },
    {
      "epoch": 12.08,
      "grad_norm": 0.6674472689628601,
      "learning_rate": 1.187732342007435e-05,
      "loss": 0.2445,
      "step": 6500
    },
    {
      "epoch": 12.08,
      "eval_loss": 0.43110528588294983,
      "eval_runtime": 6.2182,
      "eval_samples_per_second": 244.285,
      "eval_steps_per_second": 15.278,
      "step": 6500
    },
    {
      "epoch": 13.01,
      "grad_norm": 0.8294160962104797,
      "learning_rate": 1.0483271375464685e-05,
      "loss": 0.2343,
      "step": 7000
    },
    {
      "epoch": 13.01,
      "eval_loss": 0.42911913990974426,
      "eval_runtime": 6.1036,
      "eval_samples_per_second": 248.87,
      "eval_steps_per_second": 15.565,
      "step": 7000
    },
    {
      "epoch": 13.94,
      "grad_norm": 0.9937490224838257,
      "learning_rate": 9.089219330855018e-06,
      "loss": 0.227,
      "step": 7500
    },
    {
      "epoch": 13.94,
      "eval_loss": 0.4279896914958954,
      "eval_runtime": 6.1085,
      "eval_samples_per_second": 248.671,
      "eval_steps_per_second": 15.552,
      "step": 7500
    },
    {
      "epoch": 14.87,
      "grad_norm": 0.7265322804450989,
      "learning_rate": 7.695167286245354e-06,
      "loss": 0.2231,
      "step": 8000
    },
    {
      "epoch": 14.87,
      "eval_loss": 0.4289594292640686,
      "eval_runtime": 6.1309,
      "eval_samples_per_second": 247.763,
      "eval_steps_per_second": 15.495,
      "step": 8000
    },
    {
      "epoch": 15.8,
      "grad_norm": 2.6667771339416504,
      "learning_rate": 6.301115241635688e-06,
      "loss": 0.2154,
      "step": 8500
    },
    {
      "epoch": 15.8,
      "eval_loss": 0.4316449761390686,
      "eval_runtime": 6.1029,
      "eval_samples_per_second": 248.899,
      "eval_steps_per_second": 15.566,
      "step": 8500
    }
  ],
  "logging_steps": 500,
  "max_steps": 10760,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 20,
  "save_steps": 4430,
  "total_flos": 1.0656359168606208e+16,
  "train_batch_size": 16,
  "trial_name": null,
  "trial_params": null
}