trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 5.632,
  "eval_steps": 500,
  "global_step": 66,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.94,
      "grad_norm": 3.1855039143428243,
      "learning_rate": 1.9774033898178668e-05,
      "loss": 0.6049,
      "step": 11
    },
    {
      "epoch": 1.88,
      "grad_norm": 1.503490043614782,
      "learning_rate": 1.6976315211349848e-05,
      "loss": 0.5013,
      "step": 22
    },
    {
      "epoch": 2.82,
      "grad_norm": 1.0765404597934374,
      "learning_rate": 1.1852887240871145e-05,
      "loss": 0.4173,
      "step": 33
    },
    {
      "epoch": 3.75,
      "grad_norm": 1.1608136815700927,
      "learning_rate": 6.111758245266795e-06,
      "loss": 0.3491,
      "step": 44
    },
    {
      "epoch": 4.69,
      "grad_norm": 1.130960568658929,
      "learning_rate": 1.6668608091748495e-06,
      "loss": 0.3011,
      "step": 55
    },
    {
      "epoch": 5.63,
      "grad_norm": 0.8830994671304758,
      "learning_rate": 0.0,
      "loss": 0.2804,
      "step": 66
    },
    {
      "epoch": 5.63,
      "step": 66,
      "total_flos": 1.1455799142853378e+19,
      "train_loss": 0.4090242385864258,
      "train_runtime": 32683.3827,
      "train_samples_per_second": 1.101,
      "train_steps_per_second": 0.002
    }
  ],
  "logging_steps": 11,
  "max_steps": 66,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 6,
  "save_steps": 1000,
  "total_flos": 1.1455799142853378e+19,
  "train_batch_size": 1,
  "trial_name": null,
  "trial_params": null
}