trainer_state.json

{
  "best_metric": 0.8317058086395264,
  "best_model_checkpoint": "output/gpt2-multiquestions/checkpoint-1636",
  "epoch": 2.0,
  "global_step": 1636,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.61,
      "learning_rate": 8.777506112469438e-05,
      "loss": 0.4879,
      "step": 500
    },
    {
      "epoch": 1.0,
      "eval_loss": 0.8343050479888916,
      "eval_runtime": 2.343,
      "eval_samples_per_second": 213.404,
      "eval_steps_per_second": 26.889,
      "step": 818
    },
    {
      "epoch": 1.22,
      "learning_rate": 7.555012224938876e-05,
      "loss": 0.4298,
      "step": 1000
    },
    {
      "epoch": 1.83,
      "learning_rate": 6.332518337408312e-05,
      "loss": 0.3909,
      "step": 1500
    },
    {
      "epoch": 2.0,
      "eval_loss": 0.8317058086395264,
      "eval_runtime": 2.381,
      "eval_samples_per_second": 209.994,
      "eval_steps_per_second": 26.459,
      "step": 1636
    }
  ],
  "max_steps": 4090,
  "num_train_epochs": 5,
  "total_flos": 2562490957824000.0,
  "trial_name": null,
  "trial_params": null
}