0-00005/trainer_state.json

{
  "best_metric": 17.2837,
  "best_model_checkpoint": "/local1/hfs/gs_stuff/ft-wmt14/0-00005/checkpoint-10000",
  "epoch": 0.2777777777777778,
  "eval_steps": 10000,
  "global_step": 10000,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.2777777777777778,
      "grad_norm": 1.3463367223739624,
      "learning_rate": 0.0,
      "loss": 2.1463,
      "step": 10000
    },
    {
      "epoch": 0.2777777777777778,
      "eval_bleu": 17.2837,
      "eval_gen_len": 28.3707,
      "eval_loss": 2.0151596069335938,
      "eval_runtime": 244.4053,
      "eval_samples_per_second": 12.275,
      "eval_steps_per_second": 1.534,
      "step": 10000
    },
    {
      "epoch": 0.2777777777777778,
      "step": 10000,
      "total_flos": 3807008386449408.0,
      "train_loss": 2.146338671875,
      "train_runtime": 1617.1798,
      "train_samples_per_second": 98.938,
      "train_steps_per_second": 6.184
    }
  ],
  "logging_steps": 10000,
  "max_steps": 10000,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 1,
  "save_steps": 10000,
  "total_flos": 3807008386449408.0,
  "train_batch_size": 8,
  "trial_name": null,
  "trial_params": null
}