checkpoint-20/trainer_state.json

{
  "best_global_step": null,
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 2.0,
  "eval_steps": 500,
  "global_step": 20,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.5,
      "grad_norm": 1.6837834119796753,
      "learning_rate": 0.00016842105263157895,
      "loss": 1.0551,
      "step": 5
    },
    {
      "epoch": 1.0,
      "grad_norm": 0.9896103739738464,
      "learning_rate": 0.00011578947368421053,
      "loss": 0.3065,
      "step": 10
    },
    {
      "epoch": 1.5,
      "grad_norm": 1.1433913707733154,
      "learning_rate": 6.31578947368421e-05,
      "loss": 0.214,
      "step": 15
    },
    {
      "epoch": 2.0,
      "grad_norm": 1.2518914937973022,
      "learning_rate": 1.0526315789473684e-05,
      "loss": 0.1968,
      "step": 20
    }
  ],
  "logging_steps": 5,
  "max_steps": 20,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 2,
  "save_steps": 500,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": true
      },
      "attributes": {}
    }
  },
  "total_flos": 1.861822192447008e+16,
  "train_batch_size": 1,
  "trial_name": null,
  "trial_params": null
}