checkpoint-5/trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 0.3333333333333333,
  "global_step": 5,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.07,
      "learning_rate": 0.0005,
      "loss": 10.5545,
      "theoretical_loss": 20.841284642178216,
      "tokens_seen": 65536
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00046428571428571433,
      "loss": 10.5833,
      "theoretical_loss": 17.59466794495971,
      "tokens_seen": 131072
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00042857142857142855,
      "loss": 9.0487,
      "theoretical_loss": 15.967943933467634,
      "tokens_seen": 196608
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003928571428571429,
      "loss": 8.6045,
      "theoretical_loss": 14.920783596619636,
      "tokens_seen": 262144
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00035714285714285714,
      "loss": 8.612,
      "theoretical_loss": 14.164683778394041,
      "tokens_seen": 327680
    }
  ],
  "max_steps": 15,
  "num_train_epochs": 9223372036854775807,
  "total_flos": 167226900480000.0,
  "trial_name": null,
  "trial_params": null
}