trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 0.32,
  "eval_steps": 500,
  "global_step": 800,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.04,
      "learning_rate": 9.765191054744305e-05,
      "loss": 1.4073,
      "step": 100
    },
    {
      "epoch": 0.08,
      "learning_rate": 8.783564079088477e-05,
      "loss": 0.9097,
      "step": 200
    },
    {
      "epoch": 0.12,
      "learning_rate": 7.190238384283412e-05,
      "loss": 0.8875,
      "step": 300
    },
    {
      "epoch": 0.16,
      "learning_rate": 5.242811110572242e-05,
      "loss": 0.9049,
      "step": 400
    },
    {
      "epoch": 0.2,
      "learning_rate": 3.2561279283505883e-05,
      "loss": 0.8747,
      "step": 500
    },
    {
      "epoch": 0.24,
      "learning_rate": 1.5513811136094787e-05,
      "loss": 0.8431,
      "step": 600
    },
    {
      "epoch": 0.28,
      "learning_rate": 4.041815510209396e-06,
      "loss": 0.8692,
      "step": 700
    },
    {
      "epoch": 0.32,
      "learning_rate": 4.0974693947259944e-10,
      "loss": 0.8736,
      "step": 800
    }
  ],
  "logging_steps": 100,
  "max_steps": 800,
  "num_train_epochs": 1,
  "save_steps": 500,
  "total_flos": 1.293820624896e+16,
  "trial_name": null,
  "trial_params": null
}