trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 0.8532423208191127,
  "global_step": 2000,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.21,
      "learning_rate": 3.9334470989761094e-05,
      "loss": 1.7796,
      "step": 500
    },
    {
      "epoch": 0.43,
      "learning_rate": 2.8668941979522183e-05,
      "loss": 1.1378,
      "step": 1000
    },
    {
      "epoch": 0.64,
      "learning_rate": 1.800341296928328e-05,
      "loss": 1.0339,
      "step": 1500
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.337883959044369e-06,
      "loss": 0.9884,
      "step": 2000
    }
  ],
  "max_steps": 2344,
  "num_train_epochs": 1,
  "total_flos": 5005798872711168.0,
  "trial_name": null,
  "trial_params": null
}