trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 32.0,
  "global_step": 9696,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 11.24,
      "learning_rate": 1.2976485148514853e-05,
      "loss": 2.9964,
      "step": 3406
    },
    {
      "epoch": 11.24,
      "eval_bleu": 8.3668,
      "eval_gen_len": 49.1713,
      "eval_loss": 2.9263081550598145,
      "eval_meteor": 0.2788,
      "eval_runtime": 534.66,
      "eval_samples_per_second": 3.397,
      "eval_steps_per_second": 0.284,
      "step": 3406
    },
    {
      "epoch": 22.48,
      "learning_rate": 5.9529702970297035e-06,
      "loss": 1.8732,
      "step": 6812
    },
    {
      "epoch": 22.48,
      "eval_bleu": 9.2938,
      "eval_gen_len": 48.9251,
      "eval_loss": 2.9048359394073486,
      "eval_meteor": 0.2971,
      "eval_runtime": 526.9982,
      "eval_samples_per_second": 3.446,
      "eval_steps_per_second": 0.288,
      "step": 6812
    },
    {
      "epoch": 32.0,
      "step": 9696,
      "total_flos": 2.37387465916416e+16,
      "train_loss": 2.1759391255898053,
      "train_runtime": 8987.7149,
      "train_samples_per_second": 25.859,
      "train_steps_per_second": 1.079
    }
  ],
  "max_steps": 9696,
  "num_train_epochs": 32,
  "total_flos": 2.37387465916416e+16,
  "trial_name": null,
  "trial_params": null
}