trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 3.0,
  "global_step": 11121,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.27,
      "learning_rate": 9.100800287743908e-06,
      "loss": 0.6572,
      "step": 1000
    },
    {
      "epoch": 0.54,
      "learning_rate": 8.201600575487818e-06,
      "loss": 0.6344,
      "step": 2000
    },
    {
      "epoch": 0.81,
      "learning_rate": 7.3024008632317245e-06,
      "loss": 0.6401,
      "step": 3000
    },
    {
      "epoch": 1.08,
      "learning_rate": 6.403201150975632e-06,
      "loss": 0.665,
      "step": 4000
    },
    {
      "epoch": 1.35,
      "learning_rate": 5.50400143871954e-06,
      "loss": 0.6781,
      "step": 5000
    },
    {
      "epoch": 1.62,
      "learning_rate": 4.604801726463448e-06,
      "loss": 0.726,
      "step": 6000
    },
    {
      "epoch": 1.89,
      "learning_rate": 3.705602014207356e-06,
      "loss": 0.7705,
      "step": 7000
    },
    {
      "epoch": 2.16,
      "learning_rate": 2.806402301951264e-06,
      "loss": 0.8225,
      "step": 8000
    },
    {
      "epoch": 2.43,
      "learning_rate": 1.9072025896951715e-06,
      "loss": 0.8755,
      "step": 9000
    },
    {
      "epoch": 2.7,
      "learning_rate": 1.0080028774390793e-06,
      "loss": 0.9418,
      "step": 10000
    },
    {
      "epoch": 2.97,
      "learning_rate": 1.0880316518298715e-07,
      "loss": 1.013,
      "step": 11000
    },
    {
      "epoch": 3.0,
      "step": 11121,
      "total_flos": 4.147380631930061e+16,
      "train_loss": 0.7688173565565709,
      "train_runtime": 16329.8249,
      "train_samples_per_second": 2.724,
      "train_steps_per_second": 0.681
    }
  ],
  "max_steps": 11121,
  "num_train_epochs": 3,
  "total_flos": 4.147380631930061e+16,
  "trial_name": null,
  "trial_params": null
}