{
  "best_global_step": null,
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 4.0,
  "eval_steps": 1000,
  "global_step": 11158,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.0003586800573888092,
      "grad_norm": 5.781733512878418,
      "learning_rate": 0.0,
      "loss": 4.1091,
      "step": 1
    },
    {
      "epoch": 0.3586800573888092,
      "grad_norm": 5.474567413330078,
      "learning_rate": 1.4332855093256816e-05,
      "loss": 3.2109,
      "step": 1000
    },
    {
      "epoch": 0.3586800573888092,
      "eval_cosine_accuracy": 0.9464717507362366,
      "eval_loss": 0.5677164196968079,
      "eval_runtime": 34.9651,
      "eval_samples_per_second": 271.957,
      "eval_steps_per_second": 1.087,
      "step": 1000
    },
    {
      "epoch": 0.7173601147776184,
      "grad_norm": 6.178121566772461,
      "learning_rate": 1.9035549179021205e-05,
      "loss": 1.8968,
      "step": 2000
    },
    {
      "epoch": 0.7173601147776184,
      "eval_cosine_accuracy": 0.9513092637062073,
      "eval_loss": 0.5238887667655945,
      "eval_runtime": 33.5859,
      "eval_samples_per_second": 283.125,
      "eval_steps_per_second": 1.131,
      "step": 2000
    },
    {
      "epoch": 1.0759856630824374,
      "grad_norm": 4.709190368652344,
      "learning_rate": 1.7441415590626496e-05,
      "loss": 1.4326,
      "step": 3000
    },
    {
      "epoch": 1.0759856630824374,
      "eval_cosine_accuracy": 0.9547796845436096,
      "eval_loss": 0.49907511472702026,
      "eval_runtime": 33.4837,
      "eval_samples_per_second": 283.989,
      "eval_steps_per_second": 1.135,
      "step": 3000
    },
    {
      "epoch": 1.4344086021505376,
      "grad_norm": 6.273707866668701,
      "learning_rate": 1.5847282002231787e-05,
      "loss": 1.5752,
      "step": 4000
    },
    {
      "epoch": 1.4344086021505376,
      "eval_cosine_accuracy": 0.9569880962371826,
      "eval_loss": 0.48718422651290894,
      "eval_runtime": 33.7017,
      "eval_samples_per_second": 282.152,
      "eval_steps_per_second": 1.128,
      "step": 4000
    },
    {
      "epoch": 1.792831541218638,
      "grad_norm": 5.117976188659668,
      "learning_rate": 1.4254742547425475e-05,
      "loss": 1.4586,
      "step": 5000
    },
    {
      "epoch": 1.792831541218638,
      "eval_cosine_accuracy": 0.9577242732048035,
      "eval_loss": 0.48695915937423706,
      "eval_runtime": 33.7015,
      "eval_samples_per_second": 282.154,
      "eval_steps_per_second": 1.128,
      "step": 5000
    },
    {
      "epoch": 2.151254480286738,
      "grad_norm": 4.084292411804199,
      "learning_rate": 1.2660608959030766e-05,
      "loss": 1.361,
      "step": 6000
    },
    {
      "epoch": 2.151254480286738,
      "eval_cosine_accuracy": 0.9586707353591919,
      "eval_loss": 0.48321250081062317,
      "eval_runtime": 35.2989,
      "eval_samples_per_second": 269.385,
      "eval_steps_per_second": 1.077,
      "step": 6000
    },
    {
      "epoch": 2.509677419354839,
      "grad_norm": 5.087512016296387,
      "learning_rate": 1.1068069504224455e-05,
      "loss": 1.2856,
      "step": 7000
    },
    {
      "epoch": 2.509677419354839,
      "eval_cosine_accuracy": 0.9583552479743958,
      "eval_loss": 0.4754229187965393,
      "eval_runtime": 33.8388,
      "eval_samples_per_second": 281.009,
      "eval_steps_per_second": 1.123,
      "step": 7000
    },
    {
      "epoch": 2.868100358422939,
      "grad_norm": 5.019012928009033,
      "learning_rate": 9.473935915829748e-06,
      "loss": 1.2338,
      "step": 8000
    },
    {
      "epoch": 2.868100358422939,
      "eval_cosine_accuracy": 0.958250105381012,
      "eval_loss": 0.4757404327392578,
      "eval_runtime": 35.2258,
      "eval_samples_per_second": 269.944,
      "eval_steps_per_second": 1.079,
      "step": 8000
    },
    {
      "epoch": 3.2265232974910396,
      "grad_norm": 4.120110511779785,
      "learning_rate": 7.88299059461183e-06,
      "loss": 1.1657,
      "step": 9000
    },
    {
      "epoch": 3.2265232974910396,
      "eval_cosine_accuracy": 0.960037887096405,
      "eval_loss": 0.47064778208732605,
      "eval_runtime": 34.2193,
      "eval_samples_per_second": 277.884,
      "eval_steps_per_second": 1.11,
      "step": 9000
    },
    {
      "epoch": 3.58494623655914,
      "grad_norm": 5.3556227684021,
      "learning_rate": 6.288857006217122e-06,
      "loss": 1.1664,
      "step": 10000
    },
    {
      "epoch": 3.58494623655914,
      "eval_cosine_accuracy": 0.9606688618659973,
      "eval_loss": 0.4687054455280304,
      "eval_runtime": 34.013,
      "eval_samples_per_second": 279.57,
      "eval_steps_per_second": 1.117,
      "step": 10000
    },
    {
      "epoch": 3.94336917562724,
      "grad_norm": 4.998316764831543,
      "learning_rate": 4.694723417822414e-06,
      "loss": 1.1221,
      "step": 11000
    },
    {
      "epoch": 3.94336917562724,
      "eval_cosine_accuracy": 0.9598275423049927,
      "eval_loss": 0.46798574924468994,
      "eval_runtime": 34.1933,
      "eval_samples_per_second": 278.095,
      "eval_steps_per_second": 1.111,
      "step": 11000
    }
  ],
  "logging_steps": 1000,
  "max_steps": 13940,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 5,
  "save_steps": 500,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": false
      },
      "attributes": {}
    }
  },
  "total_flos": 0.0,
  "train_batch_size": 256,
  "trial_name": null,
  "trial_params": null
}