trainer_state.json

{
  "best_metric": 8.0735,
  "best_model_checkpoint": "outputs/checkpoint-50000",
  "epoch": 1.815051943543868,
  "global_step": 50000,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.0,
      "learning_rate": 1.0000000000000001e-07,
      "loss": 35.9847,
      "step": 1
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.0002494777777777778,
      "loss": 3.7902,
      "step": 27547
    },
    {
      "epoch": 1.0,
      "eval_gen_len": 16.5788,
      "eval_loss": 2.6182119846343994,
      "eval_runtime": 183.9602,
      "eval_sacrebleu": 7.7572,
      "eval_samples_per_second": 13.59,
      "eval_steps_per_second": 1.701,
      "step": 27547
    },
    {
      "epoch": 1.82,
      "learning_rate": 0.0,
      "loss": 3.0818,
      "step": 50000
    },
    {
      "epoch": 1.82,
      "eval_gen_len": 16.856,
      "eval_loss": 2.577773094177246,
      "eval_runtime": 186.4212,
      "eval_sacrebleu": 8.0735,
      "eval_samples_per_second": 13.41,
      "eval_steps_per_second": 1.679,
      "step": 50000
    },
    {
      "epoch": 1.82,
      "step": 50000,
      "total_flos": 2.0533127458240512e+17,
      "train_loss": 3.472727038040161,
      "train_runtime": 55623.1595,
      "train_samples_per_second": 28.765,
      "train_steps_per_second": 0.899
    }
  ],
  "max_steps": 50000,
  "num_train_epochs": 2,
  "total_flos": 2.0533127458240512e+17,
  "trial_name": null,
  "trial_params": null
}