trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 10.0,
  "eval_steps": 2.0,
  "global_step": 200,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.25,
      "learning_rate": 1.7500000000000002e-05,
      "loss": 0.6018,
      "step": 25
    },
    {
      "epoch": 2.5,
      "learning_rate": 1.5000000000000002e-05,
      "loss": 0.4881,
      "step": 50
    },
    {
      "epoch": 3.75,
      "learning_rate": 1.25e-05,
      "loss": 0.3612,
      "step": 75
    },
    {
      "epoch": 5.0,
      "learning_rate": 1e-05,
      "loss": 0.2575,
      "step": 100
    },
    {
      "epoch": 6.25,
      "learning_rate": 7.500000000000001e-06,
      "loss": 0.1875,
      "step": 125
    },
    {
      "epoch": 7.5,
      "learning_rate": 5e-06,
      "loss": 0.133,
      "step": 150
    },
    {
      "epoch": 8.75,
      "learning_rate": 2.5e-06,
      "loss": 0.108,
      "step": 175
    },
    {
      "epoch": 10.0,
      "learning_rate": 0.0,
      "loss": 0.088,
      "step": 200
    },
    {
      "epoch": 10.0,
      "step": 200,
      "total_flos": 2412728377651200.0,
      "train_loss": 0.27814194798469544,
      "train_runtime": 79.562,
      "train_samples_per_second": 461.024,
      "train_steps_per_second": 2.514
    }
  ],
  "logging_steps": 25,
  "max_steps": 200,
  "num_train_epochs": 10,
  "save_steps": 500,
  "total_flos": 2412728377651200.0,
  "trial_name": null,
  "trial_params": null
}