trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 0.96,
  "eval_steps": 500,
  "global_step": 18,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.05,
      "learning_rate": 0.0001,
      "loss": 2.1555,
      "step": 1
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00018314696123025454,
      "loss": 1.69,
      "step": 5
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0001,
      "loss": 0.5879,
      "step": 10
    },
    {
      "epoch": 0.8,
      "learning_rate": 1.6853038769745467e-05,
      "loss": 0.3652,
      "step": 15
    },
    {
      "epoch": 0.96,
      "eval_loss": 0.3462468981742859,
      "eval_runtime": 5.1291,
      "eval_samples_per_second": 19.496,
      "eval_steps_per_second": 2.535,
      "step": 18
    },
    {
      "epoch": 0.96,
      "step": 18,
      "total_flos": 13668439425024.0,
      "train_loss": 0.8159759442011515,
      "train_runtime": 112.608,
      "train_samples_per_second": 2.664,
      "train_steps_per_second": 0.16
    }
  ],
  "logging_steps": 5,
  "max_steps": 18,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 1,
  "save_steps": 500,
  "total_flos": 13668439425024.0,
  "train_batch_size": 4,
  "trial_name": null,
  "trial_params": null
}