trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 2.7272727272727275,
  "eval_steps": 500,
  "global_step": 15,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.18,
      "grad_norm": 7.792786415646724,
      "learning_rate": 2.5e-06,
      "loss": 1.4406,
      "step": 1
    },
    {
      "epoch": 0.91,
      "grad_norm": 4.589951830420465,
      "learning_rate": 4.3712768704277535e-06,
      "loss": 1.3896,
      "step": 5
    },
    {
      "epoch": 0.91,
      "eval_loss": 1.2333842515945435,
      "eval_runtime": 1.3546,
      "eval_samples_per_second": 10.335,
      "eval_steps_per_second": 0.738,
      "step": 5
    },
    {
      "epoch": 1.82,
      "grad_norm": 2.4335206553862827,
      "learning_rate": 1.613487782393661e-06,
      "loss": 1.1706,
      "step": 10
    },
    {
      "epoch": 2.0,
      "eval_loss": 1.1097596883773804,
      "eval_runtime": 1.363,
      "eval_samples_per_second": 10.272,
      "eval_steps_per_second": 0.734,
      "step": 11
    },
    {
      "epoch": 2.73,
      "grad_norm": 1.832707625574532,
      "learning_rate": 0.0,
      "loss": 1.0868,
      "step": 15
    },
    {
      "epoch": 2.73,
      "eval_loss": 1.0922778844833374,
      "eval_runtime": 1.3688,
      "eval_samples_per_second": 10.228,
      "eval_steps_per_second": 0.731,
      "step": 15
    },
    {
      "epoch": 2.73,
      "step": 15,
      "total_flos": 3088349921280.0,
      "train_loss": 1.2190755208333333,
      "train_runtime": 247.7641,
      "train_samples_per_second": 4.165,
      "train_steps_per_second": 0.061
    }
  ],
  "logging_steps": 5,
  "max_steps": 15,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 3,
  "save_steps": 500,
  "total_flos": 3088349921280.0,
  "train_batch_size": 4,
  "trial_name": null,
  "trial_params": null
}