trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 2.608695652173913,
  "eval_steps": 500,
  "global_step": 15,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.17,
      "grad_norm": 14.027121277015823,
      "learning_rate": 1e-05,
      "loss": 1.7106,
      "step": 1
    },
    {
      "epoch": 0.87,
      "grad_norm": 3.8863181388751995,
      "learning_rate": 1.7485107481711014e-05,
      "loss": 1.5113,
      "step": 5
    },
    {
      "epoch": 0.87,
      "eval_loss": 1.3346843719482422,
      "eval_runtime": 38.3647,
      "eval_samples_per_second": 11.547,
      "eval_steps_per_second": 0.73,
      "step": 5
    },
    {
      "epoch": 1.74,
      "grad_norm": 1.21094679444063,
      "learning_rate": 6.453951129574644e-06,
      "loss": 1.2262,
      "step": 10
    },
    {
      "epoch": 1.91,
      "eval_loss": 1.163671612739563,
      "eval_runtime": 38.3065,
      "eval_samples_per_second": 11.565,
      "eval_steps_per_second": 0.731,
      "step": 11
    },
    {
      "epoch": 2.61,
      "grad_norm": 0.8692774707662335,
      "learning_rate": 0.0,
      "loss": 1.1423,
      "step": 15
    },
    {
      "epoch": 2.61,
      "eval_loss": 1.1481901407241821,
      "eval_runtime": 38.4804,
      "eval_samples_per_second": 11.512,
      "eval_steps_per_second": 0.728,
      "step": 15
    },
    {
      "epoch": 2.61,
      "step": 15,
      "total_flos": 3088349921280.0,
      "train_loss": 1.3065491199493409,
      "train_runtime": 362.3528,
      "train_samples_per_second": 3.03,
      "train_steps_per_second": 0.041
    }
  ],
  "logging_steps": 5,
  "max_steps": 15,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 3,
  "save_steps": 500,
  "total_flos": 3088349921280.0,
  "train_batch_size": 4,
  "trial_name": null,
  "trial_params": null
}