{
  "best_global_step": null,
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 3.0,
  "eval_steps": 1000,
  "global_step": 9274,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.00032362459546925567,
      "grad_norm": 6.455742835998535,
      "learning_rate": 0.0,
      "loss": 3.5189,
      "step": 1
    },
    {
      "epoch": 0.32362459546925565,
      "grad_norm": 4.869843482971191,
      "learning_rate": 2.020631067961165e-05,
      "loss": 2.5841,
      "step": 1000
    },
    {
      "epoch": 0.32362459546925565,
      "eval_cosine_accuracy": 0.9525712728500366,
      "eval_loss": 0.5323272943496704,
      "eval_runtime": 35.0009,
      "eval_samples_per_second": 271.679,
      "eval_steps_per_second": 1.086,
      "step": 1000
    },
    {
      "epoch": 0.6472491909385113,
      "grad_norm": 6.113117694854736,
      "learning_rate": 4.04126213592233e-05,
      "loss": 1.5199,
      "step": 2000
    },
    {
      "epoch": 0.6472491909385113,
      "eval_cosine_accuracy": 0.9623514413833618,
      "eval_loss": 0.47361451387405396,
      "eval_runtime": 34.0145,
      "eval_samples_per_second": 279.557,
      "eval_steps_per_second": 1.117,
      "step": 2000
    },
    {
      "epoch": 0.970873786407767,
      "grad_norm": 6.593084812164307,
      "learning_rate": 4.7340210355987055e-05,
      "loss": 1.0615,
      "step": 3000
    },
    {
      "epoch": 0.970873786407767,
      "eval_cosine_accuracy": 0.9607740044593811,
      "eval_loss": 0.45996761322021484,
      "eval_runtime": 33.7873,
      "eval_samples_per_second": 281.437,
      "eval_steps_per_second": 1.125,
      "step": 3000
    },
    {
      "epoch": 1.2943078913324708,
      "grad_norm": 3.9310717582702637,
      "learning_rate": 4.228357605177994e-05,
      "loss": 1.1183,
      "step": 4000
    },
    {
      "epoch": 1.2943078913324708,
      "eval_cosine_accuracy": 0.9660322070121765,
      "eval_loss": 0.43745797872543335,
      "eval_runtime": 33.8541,
      "eval_samples_per_second": 280.882,
      "eval_steps_per_second": 1.122,
      "step": 4000
    },
    {
      "epoch": 1.6177231565329884,
      "grad_norm": 4.16636848449707,
      "learning_rate": 3.722694174757282e-05,
      "loss": 1.0372,
      "step": 5000
    },
    {
      "epoch": 1.6177231565329884,
      "eval_cosine_accuracy": 0.9657166600227356,
      "eval_loss": 0.4422674775123596,
      "eval_runtime": 33.8581,
      "eval_samples_per_second": 280.849,
      "eval_steps_per_second": 1.122,
      "step": 5000
    },
    {
      "epoch": 1.941138421733506,
      "grad_norm": 4.619586944580078,
      "learning_rate": 3.21703074433657e-05,
      "loss": 0.9562,
      "step": 6000
    },
    {
      "epoch": 1.941138421733506,
      "eval_cosine_accuracy": 0.9676096439361572,
      "eval_loss": 0.4301552176475525,
      "eval_runtime": 35.1089,
      "eval_samples_per_second": 270.843,
      "eval_steps_per_second": 1.082,
      "step": 6000
    },
    {
      "epoch": 2.2645536869340233,
      "grad_norm": 3.7584993839263916,
      "learning_rate": 2.711367313915858e-05,
      "loss": 0.8554,
      "step": 7000
    },
    {
      "epoch": 2.2645536869340233,
      "eval_cosine_accuracy": 0.9667683243751526,
      "eval_loss": 0.43318116664886475,
      "eval_runtime": 33.7513,
      "eval_samples_per_second": 281.738,
      "eval_steps_per_second": 1.126,
      "step": 7000
    },
    {
      "epoch": 2.5879689521345406,
      "grad_norm": 4.243616580963135,
      "learning_rate": 2.2062095469255663e-05,
      "loss": 0.812,
      "step": 8000
    },
    {
      "epoch": 2.5879689521345406,
      "eval_cosine_accuracy": 0.9670838117599487,
      "eval_loss": 0.4286963939666748,
      "eval_runtime": 34.2095,
      "eval_samples_per_second": 277.964,
      "eval_steps_per_second": 1.111,
      "step": 8000
    },
    {
      "epoch": 2.9113842173350584,
      "grad_norm": 4.7793097496032715,
      "learning_rate": 1.7005461165048546e-05,
      "loss": 0.7919,
      "step": 9000
    },
    {
      "epoch": 2.9113842173350584,
      "eval_cosine_accuracy": 0.9696077108383179,
      "eval_loss": 0.4297306537628174,
      "eval_runtime": 33.8897,
      "eval_samples_per_second": 280.586,
      "eval_steps_per_second": 1.121,
      "step": 9000
    }
  ],
  "logging_steps": 1000,
  "max_steps": 12360,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 4,
  "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
}