{
  "best_global_step": null,
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 1.999158337716991,
  "eval_steps": 2000,
  "global_step": 18994,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.00010529640939243972,
      "grad_norm": 6.751945495605469,
      "learning_rate": 0.0,
      "loss": 2.8442,
      "step": 1
    },
    {
      "epoch": 0.21059281878487943,
      "grad_norm": 4.618383884429932,
      "learning_rate": 6.315290648694187e-06,
      "loss": 2.5015,
      "step": 2000
    },
    {
      "epoch": 0.21059281878487943,
      "eval_cosine_accuracy": 0.9469176530838013,
      "eval_loss": 0.7108581066131592,
      "eval_runtime": 24.9028,
      "eval_samples_per_second": 379.756,
      "eval_steps_per_second": 1.486,
      "step": 2000
    },
    {
      "epoch": 0.42118563756975885,
      "grad_norm": 4.100494861602783,
      "learning_rate": 1.2630581297388374e-05,
      "loss": 1.8564,
      "step": 4000
    },
    {
      "epoch": 0.42118563756975885,
      "eval_cosine_accuracy": 0.9481865167617798,
      "eval_loss": 0.6932559013366699,
      "eval_runtime": 25.2131,
      "eval_samples_per_second": 375.083,
      "eval_steps_per_second": 1.467,
      "step": 4000
    },
    {
      "epoch": 0.6317784563546383,
      "grad_norm": 4.006742477416992,
      "learning_rate": 1.8949031171019377e-05,
      "loss": 1.504,
      "step": 6000
    },
    {
      "epoch": 0.6317784563546383,
      "eval_cosine_accuracy": 0.9519932270050049,
      "eval_loss": 0.6632474064826965,
      "eval_runtime": 25.5415,
      "eval_samples_per_second": 370.261,
      "eval_steps_per_second": 1.449,
      "step": 6000
    },
    {
      "epoch": 0.8423712751395177,
      "grad_norm": 3.7652177810668945,
      "learning_rate": 2.5264321819713565e-05,
      "loss": 1.2874,
      "step": 8000
    },
    {
      "epoch": 0.8423712751395177,
      "eval_cosine_accuracy": 0.9577032923698425,
      "eval_loss": 0.6434259414672852,
      "eval_runtime": 25.6225,
      "eval_samples_per_second": 369.089,
      "eval_steps_per_second": 1.444,
      "step": 8000
    },
    {
      "epoch": 1.0529195160441873,
      "grad_norm": 3.5891263484954834,
      "learning_rate": 2.9795333581180082e-05,
      "loss": 1.1986,
      "step": 10000
    },
    {
      "epoch": 1.0529195160441873,
      "eval_cosine_accuracy": 0.9589722156524658,
      "eval_loss": 0.6360689401626587,
      "eval_runtime": 25.8862,
      "eval_samples_per_second": 365.33,
      "eval_steps_per_second": 1.429,
      "step": 10000
    },
    {
      "epoch": 1.263335086796423,
      "grad_norm": 3.1852927207946777,
      "learning_rate": 2.516095252036212e-05,
      "loss": 1.0185,
      "step": 12000
    },
    {
      "epoch": 1.263335086796423,
      "eval_cosine_accuracy": 0.961192786693573,
      "eval_loss": 0.6109231114387512,
      "eval_runtime": 25.7475,
      "eval_samples_per_second": 367.298,
      "eval_steps_per_second": 1.437,
      "step": 12000
    },
    {
      "epoch": 1.4737506575486585,
      "grad_norm": 3.5387814044952393,
      "learning_rate": 1.624389485225718e-05,
      "loss": 0.945,
      "step": 14000
    },
    {
      "epoch": 1.4737506575486585,
      "eval_cosine_accuracy": 0.9624616503715515,
      "eval_loss": 0.5934072732925415,
      "eval_runtime": 25.7367,
      "eval_samples_per_second": 367.451,
      "eval_steps_per_second": 1.438,
      "step": 14000
    },
    {
      "epoch": 1.6841662283008942,
      "grad_norm": 3.443584680557251,
      "learning_rate": 6.802555688395712e-06,
      "loss": 0.9076,
      "step": 16000
    },
    {
      "epoch": 1.6841662283008942,
      "eval_cosine_accuracy": 0.9616157412528992,
      "eval_loss": 0.5919129252433777,
      "eval_runtime": 25.6697,
      "eval_samples_per_second": 368.411,
      "eval_steps_per_second": 1.441,
      "step": 16000
    },
    {
      "epoch": 1.89458179905313,
      "grad_norm": 3.5262327194213867,
      "learning_rate": 8.163062833430229e-07,
      "loss": 0.8964,
      "step": 18000
    },
    {
      "epoch": 1.89458179905313,
      "eval_cosine_accuracy": 0.9633076190948486,
      "eval_loss": 0.5839827060699463,
      "eval_runtime": 25.6367,
      "eval_samples_per_second": 368.885,
      "eval_steps_per_second": 1.443,
      "step": 18000
    }
  ],
  "logging_steps": 2000,
  "max_steps": 18994,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 2,
  "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
}