trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 0.947680157946693,
  "eval_steps": 500,
  "global_step": 15,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.06,
      "learning_rate": 1.9781476007338058e-05,
      "loss": 1.8063,
      "step": 1
    },
    {
      "epoch": 0.32,
      "learning_rate": 1.5000000000000002e-05,
      "loss": 1.7809,
      "step": 5
    },
    {
      "epoch": 0.63,
      "learning_rate": 5.000000000000003e-06,
      "loss": 1.7536,
      "step": 10
    },
    {
      "epoch": 0.95,
      "learning_rate": 0.0,
      "loss": 1.7926,
      "step": 15
    },
    {
      "epoch": 0.95,
      "eval_loss": 1.7090861797332764,
      "eval_runtime": 2328.2925,
      "eval_samples_per_second": 0.466,
      "eval_steps_per_second": 0.466,
      "step": 15
    },
    {
      "epoch": 0.95,
      "step": 15,
      "total_flos": 1.6904795211798938e+17,
      "train_loss": 1.777411683400472,
      "train_runtime": 7576.8776,
      "train_samples_per_second": 0.267,
      "train_steps_per_second": 0.002
    }
  ],
  "logging_steps": 5,
  "max_steps": 15,
  "num_train_epochs": 1,
  "save_steps": 500,
  "total_flos": 1.6904795211798938e+17,
  "trial_name": null,
  "trial_params": null
}