{
  "best_global_step": null,
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 3.0,
  "eval_steps": 1000,
  "global_step": 9745,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.0003079765937788728,
      "grad_norm": 5.27286434173584,
      "learning_rate": 0.0,
      "loss": 3.17,
      "step": 1
    },
    {
      "epoch": 0.3079765937788728,
      "grad_norm": 3.784266710281372,
      "learning_rate": 2.30715935334873e-05,
      "loss": 2.2231,
      "step": 1000
    },
    {
      "epoch": 0.3079765937788728,
      "eval_cosine_accuracy": 0.960388720035553,
      "eval_loss": 1.1210521459579468,
      "eval_runtime": 24.6573,
      "eval_samples_per_second": 383.944,
      "eval_steps_per_second": 1.501,
      "step": 1000
    },
    {
      "epoch": 0.6159531875577456,
      "grad_norm": 4.136064052581787,
      "learning_rate": 2.8203439130806744e-05,
      "loss": 1.623,
      "step": 2000
    },
    {
      "epoch": 0.6159531875577456,
      "eval_cosine_accuracy": 0.9700010418891907,
      "eval_loss": 1.0202350616455078,
      "eval_runtime": 23.9452,
      "eval_samples_per_second": 395.361,
      "eval_steps_per_second": 1.545,
      "step": 2000
    },
    {
      "epoch": 0.9239297813366184,
      "grad_norm": 3.366196632385254,
      "learning_rate": 2.56369236033878e-05,
      "loss": 1.6346,
      "step": 3000
    },
    {
      "epoch": 0.9239297813366184,
      "eval_cosine_accuracy": 0.9733812212944031,
      "eval_loss": 0.8886165022850037,
      "eval_runtime": 23.9611,
      "eval_samples_per_second": 395.099,
      "eval_steps_per_second": 1.544,
      "step": 3000
    },
    {
      "epoch": 1.231763619575254,
      "grad_norm": 3.6002681255340576,
      "learning_rate": 2.307297459149628e-05,
      "loss": 1.7279,
      "step": 4000
    },
    {
      "epoch": 1.231763619575254,
      "eval_cosine_accuracy": 0.9750713109970093,
      "eval_loss": 0.9043466448783875,
      "eval_runtime": 23.7609,
      "eval_samples_per_second": 398.427,
      "eval_steps_per_second": 1.557,
      "step": 4000
    },
    {
      "epoch": 1.5395506309633733,
      "grad_norm": 3.5627574920654297,
      "learning_rate": 2.050645906407734e-05,
      "loss": 1.3841,
      "step": 5000
    },
    {
      "epoch": 1.5395506309633733,
      "eval_cosine_accuracy": 0.9782401919364929,
      "eval_loss": 0.8797643780708313,
      "eval_runtime": 23.6385,
      "eval_samples_per_second": 400.491,
      "eval_steps_per_second": 1.565,
      "step": 5000
    },
    {
      "epoch": 1.847337642351493,
      "grad_norm": 3.2826766967773438,
      "learning_rate": 1.7939943536658397e-05,
      "loss": 1.3355,
      "step": 6000
    },
    {
      "epoch": 1.847337642351493,
      "eval_cosine_accuracy": 0.9783458113670349,
      "eval_loss": 0.8727800250053406,
      "eval_runtime": 23.4386,
      "eval_samples_per_second": 403.907,
      "eval_steps_per_second": 1.579,
      "step": 6000
    },
    {
      "epoch": 2.155124653739612,
      "grad_norm": 3.485886812210083,
      "learning_rate": 1.5373428009239457e-05,
      "loss": 1.2823,
      "step": 7000
    },
    {
      "epoch": 2.155124653739612,
      "eval_cosine_accuracy": 0.9784514904022217,
      "eval_loss": 0.8731770515441895,
      "eval_runtime": 23.265,
      "eval_samples_per_second": 406.92,
      "eval_steps_per_second": 1.59,
      "step": 7000
    },
    {
      "epoch": 2.4629116651277316,
      "grad_norm": 3.2091927528381348,
      "learning_rate": 1.2809478997347934e-05,
      "loss": 1.2592,
      "step": 8000
    },
    {
      "epoch": 2.4629116651277316,
      "eval_cosine_accuracy": 0.9802471995353699,
      "eval_loss": 0.8694875836372375,
      "eval_runtime": 23.2395,
      "eval_samples_per_second": 407.367,
      "eval_steps_per_second": 1.592,
      "step": 8000
    },
    {
      "epoch": 2.7706986765158508,
      "grad_norm": 6.691135883331299,
      "learning_rate": 1.0242963469928994e-05,
      "loss": 1.2272,
      "step": 9000
    },
    {
      "epoch": 2.7706986765158508,
      "eval_cosine_accuracy": 0.9792965054512024,
      "eval_loss": 0.8561711311340332,
      "eval_runtime": 23.6674,
      "eval_samples_per_second": 400.001,
      "eval_steps_per_second": 1.563,
      "step": 9000
    }
  ],
  "logging_steps": 1000,
  "max_steps": 12988,
  "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
}