trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 0.9894588896697118,
  "eval_steps": 500,
  "global_step": 11,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.09,
      "learning_rate": 0.0001,
      "loss": 1.8236,
      "step": 1
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00015000000000000001,
      "loss": 1.775,
      "step": 5
    },
    {
      "epoch": 0.9,
      "learning_rate": 6.030737921409169e-06,
      "loss": 1.5495,
      "step": 10
    },
    {
      "epoch": 0.99,
      "eval_loss": 1.506221055984497,
      "eval_runtime": 640.0172,
      "eval_samples_per_second": 0.991,
      "eval_steps_per_second": 0.125,
      "step": 11
    },
    {
      "epoch": 0.99,
      "step": 11,
      "total_flos": 9.067629788193096e+17,
      "train_loss": 1.6525719924406572,
      "train_runtime": 19611.737,
      "train_samples_per_second": 0.29,
      "train_steps_per_second": 0.001
    }
  ],
  "logging_steps": 5,
  "max_steps": 11,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 1,
  "save_steps": 100,
  "total_flos": 9.067629788193096e+17,
  "train_batch_size": 4,
  "trial_name": null,
  "trial_params": null
}