last-checkpoint/trainer_state.json

{
  "best_metric": 1.0592867136001587,
  "best_model_checkpoint": "./outputs/checkpoint-200",
  "epoch": 0.2840909090909091,
  "eval_steps": 100,
  "global_step": 200,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.14,
      "learning_rate": 0.0002,
      "loss": 1.5808,
      "step": 100
    },
    {
      "epoch": 0.14,
      "eval_loss": 1.0812042951583862,
      "eval_runtime": 127.4925,
      "eval_samples_per_second": 11.703,
      "eval_steps_per_second": 1.467,
      "step": 100
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0002,
      "loss": 1.2145,
      "step": 200
    },
    {
      "epoch": 0.28,
      "eval_loss": 1.0592867136001587,
      "eval_runtime": 124.4418,
      "eval_samples_per_second": 11.99,
      "eval_steps_per_second": 1.503,
      "step": 200
    }
  ],
  "logging_steps": 100,
  "max_steps": 2112,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 3,
  "save_steps": 100,
  "total_flos": 1.11990315515904e+16,
  "trial_name": null,
  "trial_params": null
}