checkpoint-129/trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 3.0,
  "eval_steps": 500,
  "global_step": 129,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.5813953488372093,
      "grad_norm": 1.75051748752594,
      "learning_rate": 0.0001793103448275862,
      "loss": 1.4127,
      "step": 25
    },
    {
      "epoch": 1.1627906976744187,
      "grad_norm": 0.3321884870529175,
      "learning_rate": 0.00013620689655172413,
      "loss": 0.922,
      "step": 50
    },
    {
      "epoch": 1.744186046511628,
      "grad_norm": 0.4291050434112549,
      "learning_rate": 9.310344827586207e-05,
      "loss": 0.8028,
      "step": 75
    },
    {
      "epoch": 2.3255813953488373,
      "grad_norm": 0.49715569615364075,
      "learning_rate": 5e-05,
      "loss": 0.838,
      "step": 100
    },
    {
      "epoch": 2.9069767441860463,
      "grad_norm": 0.5227110981941223,
      "learning_rate": 6.896551724137932e-06,
      "loss": 0.7422,
      "step": 125
    }
  ],
  "logging_steps": 25,
  "max_steps": 129,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 3,
  "save_steps": 500,
  "total_flos": 1.0486941166338048e+16,
  "train_batch_size": 2,
  "trial_name": null,
  "trial_params": null
}