{
  "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": 7.473833084106445,
      "learning_rate": 0.0,
      "loss": 4.2699,
      "step": 1
    },
    {
      "epoch": 0.3586800573888092,
      "grad_norm": 6.972509860992432,
      "learning_rate": 1.428981348637016e-05,
      "loss": 3.0799,
      "step": 1000
    },
    {
      "epoch": 0.3586800573888092,
      "eval_cosine_accuracy": 0.9463666081428528,
      "eval_loss": 0.5312139987945557,
      "eval_runtime": 56.9482,
      "eval_samples_per_second": 166.976,
      "eval_steps_per_second": 0.667,
      "step": 1000
    },
    {
      "epoch": 0.7173601147776184,
      "grad_norm": 7.692445278167725,
      "learning_rate": 1.9040331579786388e-05,
      "loss": 1.6967,
      "step": 2000
    },
    {
      "epoch": 0.7173601147776184,
      "eval_cosine_accuracy": 0.9519402384757996,
      "eval_loss": 0.4940159022808075,
      "eval_runtime": 56.3183,
      "eval_samples_per_second": 168.844,
      "eval_steps_per_second": 0.675,
      "step": 2000
    },
    {
      "epoch": 1.0759856630824374,
      "grad_norm": 6.095524311065674,
      "learning_rate": 1.744619799139168e-05,
      "loss": 1.2766,
      "step": 3000
    },
    {
      "epoch": 1.0759856630824374,
      "eval_cosine_accuracy": 0.9552003145217896,
      "eval_loss": 0.4655364751815796,
      "eval_runtime": 56.2099,
      "eval_samples_per_second": 169.17,
      "eval_steps_per_second": 0.676,
      "step": 3000
    },
    {
      "epoch": 1.4344086021505376,
      "grad_norm": 7.97337007522583,
      "learning_rate": 1.585206440299697e-05,
      "loss": 1.3959,
      "step": 4000
    },
    {
      "epoch": 1.4344086021505376,
      "eval_cosine_accuracy": 0.9605636596679688,
      "eval_loss": 0.46521449089050293,
      "eval_runtime": 55.8952,
      "eval_samples_per_second": 170.122,
      "eval_steps_per_second": 0.68,
      "step": 4000
    },
    {
      "epoch": 1.792831541218638,
      "grad_norm": 6.2700090408325195,
      "learning_rate": 1.4261119081779054e-05,
      "loss": 1.2793,
      "step": 5000
    },
    {
      "epoch": 1.792831541218638,
      "eval_cosine_accuracy": 0.9587758779525757,
      "eval_loss": 0.46579667925834656,
      "eval_runtime": 56.6266,
      "eval_samples_per_second": 167.924,
      "eval_steps_per_second": 0.671,
      "step": 5000
    },
    {
      "epoch": 2.151254480286738,
      "grad_norm": 5.520406246185303,
      "learning_rate": 1.2666985493384347e-05,
      "loss": 1.1814,
      "step": 6000
    },
    {
      "epoch": 2.151254480286738,
      "eval_cosine_accuracy": 0.9608791470527649,
      "eval_loss": 0.46335798501968384,
      "eval_runtime": 56.6782,
      "eval_samples_per_second": 167.772,
      "eval_steps_per_second": 0.67,
      "step": 6000
    },
    {
      "epoch": 2.509677419354839,
      "grad_norm": 6.737025260925293,
      "learning_rate": 1.1072851904989638e-05,
      "loss": 1.1045,
      "step": 7000
    },
    {
      "epoch": 2.509677419354839,
      "eval_cosine_accuracy": 0.9618256092071533,
      "eval_loss": 0.45776286721229553,
      "eval_runtime": 56.1532,
      "eval_samples_per_second": 169.34,
      "eval_steps_per_second": 0.677,
      "step": 7000
    },
    {
      "epoch": 2.868100358422939,
      "grad_norm": 7.002987861633301,
      "learning_rate": 9.480312450183326e-06,
      "loss": 1.0585,
      "step": 8000
    },
    {
      "epoch": 2.868100358422939,
      "eval_cosine_accuracy": 0.9616153240203857,
      "eval_loss": 0.4550265073776245,
      "eval_runtime": 56.0314,
      "eval_samples_per_second": 169.708,
      "eval_steps_per_second": 0.678,
      "step": 8000
    },
    {
      "epoch": 3.2265232974910396,
      "grad_norm": 5.350640773773193,
      "learning_rate": 7.886178861788618e-06,
      "loss": 0.9791,
      "step": 9000
    },
    {
      "epoch": 3.2265232974910396,
      "eval_cosine_accuracy": 0.9624566435813904,
      "eval_loss": 0.4573860764503479,
      "eval_runtime": 56.0494,
      "eval_samples_per_second": 169.654,
      "eval_steps_per_second": 0.678,
      "step": 9000
    },
    {
      "epoch": 3.58494623655914,
      "grad_norm": 6.804051876068115,
      "learning_rate": 6.29204527339391e-06,
      "loss": 0.9813,
      "step": 10000
    },
    {
      "epoch": 3.58494623655914,
      "eval_cosine_accuracy": 0.962246298789978,
      "eval_loss": 0.45354387164115906,
      "eval_runtime": 56.8485,
      "eval_samples_per_second": 167.269,
      "eval_steps_per_second": 0.668,
      "step": 10000
    },
    {
      "epoch": 3.94336917562724,
      "grad_norm": 5.751834869384766,
      "learning_rate": 4.697911684999203e-06,
      "loss": 0.9417,
      "step": 11000
    },
    {
      "epoch": 3.94336917562724,
      "eval_cosine_accuracy": 0.9605636596679688,
      "eval_loss": 0.4583044946193695,
      "eval_runtime": 57.091,
      "eval_samples_per_second": 166.559,
      "eval_steps_per_second": 0.666,
      "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
}