checkpoint-579/trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 0.9987063389391979,
  "eval_steps": 500,
  "global_step": 579,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.0,
      "grad_norm": 0.8260741233825684,
      "learning_rate": 0.0001501296456352636,
      "loss": 0.9352,
      "step": 579
    },
    {
      "epoch": 1.0,
      "eval_loss": 0.8111525177955627,
      "eval_runtime": 7.5351,
      "eval_samples_per_second": 6.636,
      "eval_steps_per_second": 1.725,
      "step": 579
    }
  ],
  "logging_steps": 500,
  "max_steps": 2316,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 4,
  "save_steps": 500,
  "total_flos": 3.0632304816685056e+16,
  "train_batch_size": 4,
  "trial_name": null,
  "trial_params": null
}