trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 5.0,
  "eval_steps": 500,
  "global_step": 17835,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.0,
      "grad_norm": 4.056117534637451,
      "learning_rate": 4e-05,
      "loss": 0.6327,
      "step": 3567
    },
    {
      "epoch": 1.0,
      "eval_gen_len": 1.0,
      "eval_loss": 0.4634159207344055,
      "eval_rouge1": 0.3795,
      "eval_rouge2": 0.0,
      "eval_rougeL": 0.3806,
      "eval_rougeLsum": 0.3802,
      "eval_runtime": 331.4212,
      "eval_samples_per_second": 2.254,
      "eval_steps_per_second": 0.284,
      "step": 3567
    },
    {
      "epoch": 2.0,
      "grad_norm": 3.9469492435455322,
      "learning_rate": 3e-05,
      "loss": 0.4349,
      "step": 7134
    },
    {
      "epoch": 2.0,
      "eval_gen_len": 1.0,
      "eval_loss": 0.45409438014030457,
      "eval_rouge1": 0.3596,
      "eval_rouge2": 0.0,
      "eval_rougeL": 0.3615,
      "eval_rougeLsum": 0.3631,
      "eval_runtime": 336.4749,
      "eval_samples_per_second": 2.22,
      "eval_steps_per_second": 0.279,
      "step": 7134
    },
    {
      "epoch": 3.0,
      "grad_norm": 2.7169110774993896,
      "learning_rate": 2e-05,
      "loss": 0.3376,
      "step": 10701
    },
    {
      "epoch": 3.0,
      "eval_gen_len": 1.0,
      "eval_loss": 0.4561801552772522,
      "eval_rouge1": 0.4121,
      "eval_rouge2": 0.0,
      "eval_rougeL": 0.4123,
      "eval_rougeLsum": 0.4116,
      "eval_runtime": 332.2435,
      "eval_samples_per_second": 2.248,
      "eval_steps_per_second": 0.283,
      "step": 10701
    },
    {
      "epoch": 4.0,
      "grad_norm": 5.345052719116211,
      "learning_rate": 1e-05,
      "loss": 0.2683,
      "step": 14268
    },
    {
      "epoch": 4.0,
      "eval_gen_len": 1.0,
      "eval_loss": 0.474356472492218,
      "eval_rouge1": 0.3886,
      "eval_rouge2": 0.0,
      "eval_rougeL": 0.3869,
      "eval_rougeLsum": 0.388,
      "eval_runtime": 329.9317,
      "eval_samples_per_second": 2.264,
      "eval_steps_per_second": 0.285,
      "step": 14268
    },
    {
      "epoch": 5.0,
      "grad_norm": 3.9704811573028564,
      "learning_rate": 0.0,
      "loss": 0.2208,
      "step": 17835
    },
    {
      "epoch": 5.0,
      "eval_gen_len": 1.0,
      "eval_loss": 0.4922512173652649,
      "eval_rouge1": 0.3905,
      "eval_rouge2": 0.0,
      "eval_rougeL": 0.3898,
      "eval_rougeLsum": 0.3916,
      "eval_runtime": 332.1945,
      "eval_samples_per_second": 2.249,
      "eval_steps_per_second": 0.283,
      "step": 17835
    },
    {
      "epoch": 5.0,
      "step": 17835,
      "total_flos": 6.557265830209536e+16,
      "train_loss": 0.37884998882989906,
      "train_runtime": 9312.1131,
      "train_samples_per_second": 7.66,
      "train_steps_per_second": 1.915
    }
  ],
  "logging_steps": 500,
  "max_steps": 17835,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 5,
  "save_steps": 500,
  "total_flos": 6.557265830209536e+16,
  "train_batch_size": 4,
  "trial_name": null,
  "trial_params": null
}