trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 10.0,
  "eval_steps": 500,
  "global_step": 3460,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.45,
      "learning_rate": 4.2774566473988445e-05,
      "loss": 1.6654,
      "step": 500
    },
    {
      "epoch": 2.89,
      "learning_rate": 3.554913294797688e-05,
      "loss": 0.9341,
      "step": 1000
    },
    {
      "epoch": 4.34,
      "learning_rate": 2.832369942196532e-05,
      "loss": 0.6779,
      "step": 1500
    },
    {
      "epoch": 5.78,
      "learning_rate": 2.1098265895953757e-05,
      "loss": 0.5258,
      "step": 2000
    },
    {
      "epoch": 7.23,
      "learning_rate": 1.3872832369942197e-05,
      "loss": 0.4303,
      "step": 2500
    },
    {
      "epoch": 8.67,
      "learning_rate": 6.647398843930635e-06,
      "loss": 0.4153,
      "step": 3000
    },
    {
      "epoch": 10.0,
      "step": 3460,
      "total_flos": 8.674451270424576e+16,
      "train_loss": 0.7383389687951589,
      "train_runtime": 2418.3642,
      "train_samples_per_second": 366.049,
      "train_steps_per_second": 1.431
    }
  ],
  "logging_steps": 500,
  "max_steps": 3460,
  "num_train_epochs": 10,
  "save_steps": 500,
  "total_flos": 8.674451270424576e+16,
  "trial_name": null,
  "trial_params": null
}