trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 0.5116588940706196,
  "eval_steps": 500,
  "global_step": 6,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.09,
      "learning_rate": 1e-05,
      "loss": 0.867,
      "step": 1
    },
    {
      "epoch": 0.43,
      "learning_rate": 1.5000000000000002e-05,
      "loss": 0.8622,
      "step": 5
    },
    {
      "epoch": 0.51,
      "eval_loss": 0.9330160021781921,
      "eval_runtime": 4125.2443,
      "eval_samples_per_second": 5.632,
      "eval_steps_per_second": 0.352,
      "step": 6
    },
    {
      "epoch": 0.51,
      "step": 6,
      "total_flos": 6.133270851249766e+17,
      "train_loss": 0.9835741718610128,
      "train_runtime": 9800.1157,
      "train_samples_per_second": 1.225,
      "train_steps_per_second": 0.001
    }
  ],
  "logging_steps": 5,
  "max_steps": 11,
  "num_train_epochs": 1,
  "save_steps": 500,
  "total_flos": 6.133270851249766e+17,
  "trial_name": null,
  "trial_params": null
}