checkpoint-160/trainer_state.json

{
  "best_global_step": null,
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 4.0,
  "eval_steps": 400,
  "global_step": 160,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.2543720190779015,
      "grad_norm": 0.06228283420205116,
      "learning_rate": 9.8e-05,
      "loss": 0.6457,
      "step": 50
    },
    {
      "epoch": 2.508744038155803,
      "grad_norm": 0.027570705860853195,
      "learning_rate": 0.00019800000000000002,
      "loss": 0.3064,
      "step": 100
    },
    {
      "epoch": 3.763116057233704,
      "grad_norm": 0.030717678368091583,
      "learning_rate": 3.6666666666666666e-05,
      "loss": 0.2606,
      "step": 150
    }
  ],
  "logging_steps": 50,
  "max_steps": 160,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 4,
  "save_steps": 400,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": true
      },
      "attributes": {}
    }
  },
  "total_flos": 1.6520324775873085e+18,
  "train_batch_size": 4,
  "trial_name": null,
  "trial_params": null
}