trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 66.66666666666667,
  "eval_steps": 500,
  "global_step": 100,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 6.67,
      "learning_rate": 0.0002,
      "loss": 1.8198,
      "step": 10
    },
    {
      "epoch": 13.33,
      "learning_rate": 0.0002,
      "loss": 0.5831,
      "step": 20
    },
    {
      "epoch": 20.0,
      "learning_rate": 0.0002,
      "loss": 0.1285,
      "step": 30
    },
    {
      "epoch": 26.67,
      "learning_rate": 0.0002,
      "loss": 0.0797,
      "step": 40
    },
    {
      "epoch": 33.33,
      "learning_rate": 0.0002,
      "loss": 0.0755,
      "step": 50
    },
    {
      "epoch": 40.0,
      "learning_rate": 0.0002,
      "loss": 0.0748,
      "step": 60
    },
    {
      "epoch": 46.67,
      "learning_rate": 0.0002,
      "loss": 0.0739,
      "step": 70
    },
    {
      "epoch": 53.33,
      "learning_rate": 0.0002,
      "loss": 0.074,
      "step": 80
    },
    {
      "epoch": 60.0,
      "learning_rate": 0.0002,
      "loss": 0.074,
      "step": 90
    },
    {
      "epoch": 66.67,
      "learning_rate": 0.0002,
      "loss": 0.0735,
      "step": 100
    }
  ],
  "logging_steps": 10,
  "max_steps": 100,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 100,
  "save_steps": 100,
  "total_flos": 616647229145088.0,
  "train_batch_size": 4,
  "trial_name": null,
  "trial_params": null
}