tmp-checkpoint-1/trainer_state.json

{
  "best_metric": 2.5646207332611084,
  "best_model_checkpoint": "models/yo_nllb_lr2e-05_b48/checkpoint-1",
  "epoch": 1.0,
  "eval_steps": 500,
  "global_step": 1,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.0,
      "grad_norm": 2.999908208847046,
      "learning_rate": 2e-05,
      "loss": 1.8917,
      "step": 1
    },
    {
      "epoch": 1.0,
      "eval_bleu": 10.826,
      "eval_comet": 0.2716,
      "eval_gen_len": 29.25,
      "eval_loss": 2.5646207332611084,
      "eval_runtime": 5.8004,
      "eval_samples_per_second": 2.758,
      "eval_steps_per_second": 0.172,
      "step": 1
    }
  ],
  "logging_steps": 1,
  "max_steps": 4,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 4,
  "save_steps": 500,
  "total_flos": 2437992677376.0,
  "train_batch_size": 48,
  "trial_name": null,
  "trial_params": null
}