{
  "best_global_step": null,
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 3.998301726577979,
  "eval_steps": 1000,
  "global_step": 14124,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.0002832058906825262,
      "grad_norm": 5.138741493225098,
      "learning_rate": 0.0,
      "loss": 2.7345,
      "step": 1
    },
    {
      "epoch": 0.2832058906825262,
      "grad_norm": 3.8642654418945312,
      "learning_rate": 1.0608849557522124e-05,
      "loss": 2.3628,
      "step": 1000
    },
    {
      "epoch": 0.2832058906825262,
      "eval_cosine_accuracy": 0.9549968242645264,
      "eval_loss": 1.1044740676879883,
      "eval_runtime": 24.643,
      "eval_samples_per_second": 384.125,
      "eval_steps_per_second": 1.501,
      "step": 1000
    },
    {
      "epoch": 0.5664117813650524,
      "grad_norm": 3.815659999847412,
      "learning_rate": 2.1228318584070797e-05,
      "loss": 1.8255,
      "step": 2000
    },
    {
      "epoch": 0.5664117813650524,
      "eval_cosine_accuracy": 0.965983510017395,
      "eval_loss": 1.0202418565750122,
      "eval_runtime": 23.6988,
      "eval_samples_per_second": 399.43,
      "eval_steps_per_second": 1.561,
      "step": 2000
    },
    {
      "epoch": 0.8496176720475785,
      "grad_norm": 3.5664591789245605,
      "learning_rate": 2.953801221347022e-05,
      "loss": 1.6176,
      "step": 3000
    },
    {
      "epoch": 0.8496176720475785,
      "eval_cosine_accuracy": 0.9692584276199341,
      "eval_loss": 0.9256343841552734,
      "eval_runtime": 27.6579,
      "eval_samples_per_second": 342.253,
      "eval_steps_per_second": 1.338,
      "step": 3000
    },
    {
      "epoch": 1.1327483724879706,
      "grad_norm": 3.3883819580078125,
      "learning_rate": 2.6882909992034692e-05,
      "loss": 1.9619,
      "step": 4000
    },
    {
      "epoch": 1.1327483724879706,
      "eval_cosine_accuracy": 0.9713712334632874,
      "eval_loss": 0.8905332088470459,
      "eval_runtime": 24.0669,
      "eval_samples_per_second": 393.32,
      "eval_steps_per_second": 1.537,
      "step": 4000
    },
    {
      "epoch": 1.4157939428247948,
      "grad_norm": 3.606387138366699,
      "learning_rate": 2.4230462872820605e-05,
      "loss": 1.4568,
      "step": 5000
    },
    {
      "epoch": 1.4157939428247948,
      "eval_cosine_accuracy": 0.9725332856178284,
      "eval_loss": 0.8822396397590637,
      "eval_runtime": 23.9729,
      "eval_samples_per_second": 394.862,
      "eval_steps_per_second": 1.543,
      "step": 5000
    },
    {
      "epoch": 1.698839513161619,
      "grad_norm": 3.6768007278442383,
      "learning_rate": 2.1575360651385078e-05,
      "loss": 1.3739,
      "step": 6000
    },
    {
      "epoch": 1.698839513161619,
      "eval_cosine_accuracy": 0.9725332856178284,
      "eval_loss": 0.8737966418266296,
      "eval_runtime": 24.139,
      "eval_samples_per_second": 392.145,
      "eval_steps_per_second": 1.533,
      "step": 6000
    },
    {
      "epoch": 1.9818850834984434,
      "grad_norm": 3.6265289783477783,
      "learning_rate": 1.8922913532170987e-05,
      "loss": 1.3245,
      "step": 7000
    },
    {
      "epoch": 1.9818850834984434,
      "eval_cosine_accuracy": 0.9744347929954529,
      "eval_loss": 0.8641142249107361,
      "eval_runtime": 24.3115,
      "eval_samples_per_second": 389.364,
      "eval_steps_per_second": 1.522,
      "step": 7000
    },
    {
      "epoch": 2.2649306538352674,
      "grad_norm": 3.438566207885742,
      "learning_rate": 1.6267811310735464e-05,
      "loss": 1.2634,
      "step": 8000
    },
    {
      "epoch": 2.2649306538352674,
      "eval_cosine_accuracy": 0.9749630093574524,
      "eval_loss": 0.8607254028320312,
      "eval_runtime": 24.1145,
      "eval_samples_per_second": 392.544,
      "eval_steps_per_second": 1.534,
      "step": 8000
    },
    {
      "epoch": 2.547976224172092,
      "grad_norm": 3.4893147945404053,
      "learning_rate": 1.3615364191521373e-05,
      "loss": 1.246,
      "step": 9000
    },
    {
      "epoch": 2.547976224172092,
      "eval_cosine_accuracy": 0.9744347929954529,
      "eval_loss": 0.8550631999969482,
      "eval_runtime": 24.3882,
      "eval_samples_per_second": 388.139,
      "eval_steps_per_second": 1.517,
      "step": 9000
    },
    {
      "epoch": 2.831021794508916,
      "grad_norm": 3.4752719402313232,
      "learning_rate": 1.0960261970085847e-05,
      "loss": 1.2144,
      "step": 10000
    },
    {
      "epoch": 2.831021794508916,
      "eval_cosine_accuracy": 0.9754912257194519,
      "eval_loss": 0.8519046306610107,
      "eval_runtime": 24.1435,
      "eval_samples_per_second": 392.072,
      "eval_steps_per_second": 1.533,
      "step": 10000
    },
    {
      "epoch": 3.11406736484574,
      "grad_norm": 3.2669014930725098,
      "learning_rate": 8.307814850871758e-06,
      "loss": 1.1947,
      "step": 11000
    },
    {
      "epoch": 3.11406736484574,
      "eval_cosine_accuracy": 0.9754912257194519,
      "eval_loss": 0.8547665476799011,
      "eval_runtime": 24.0863,
      "eval_samples_per_second": 393.004,
      "eval_steps_per_second": 1.536,
      "step": 11000
    },
    {
      "epoch": 3.397112935182564,
      "grad_norm": 3.2285778522491455,
      "learning_rate": 5.655367731657669e-06,
      "loss": 1.1763,
      "step": 12000
    },
    {
      "epoch": 3.397112935182564,
      "eval_cosine_accuracy": 0.9757025241851807,
      "eval_loss": 0.8565053343772888,
      "eval_runtime": 24.1275,
      "eval_samples_per_second": 392.332,
      "eval_steps_per_second": 1.534,
      "step": 12000
    },
    {
      "epoch": 3.6801585055193886,
      "grad_norm": 3.3000833988189697,
      "learning_rate": 3.0002655102221436e-06,
      "loss": 1.1602,
      "step": 13000
    },
    {
      "epoch": 3.6801585055193886,
      "eval_cosine_accuracy": 0.9754912257194519,
      "eval_loss": 0.8563781380653381,
      "eval_runtime": 24.4069,
      "eval_samples_per_second": 387.842,
      "eval_steps_per_second": 1.516,
      "step": 13000
    },
    {
      "epoch": 3.963204075856213,
      "grad_norm": 3.174485921859741,
      "learning_rate": 3.451632887866183e-07,
      "loss": 1.1548,
      "step": 14000
    },
    {
      "epoch": 3.963204075856213,
      "eval_cosine_accuracy": 0.9757025241851807,
      "eval_loss": 0.8540062308311462,
      "eval_runtime": 24.2517,
      "eval_samples_per_second": 390.324,
      "eval_steps_per_second": 1.526,
      "step": 14000
    }
  ],
  "logging_steps": 1000,
  "max_steps": 14124,
  "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": true
      },
      "attributes": {}
    }
  },
  "total_flos": 0.0,
  "train_batch_size": 256,
  "trial_name": null,
  "trial_params": null
}