last-checkpoint/trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 0.011142061281337047,
  "eval_steps": 3,
  "global_step": 3,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.003714020427112349,
      "grad_norm": 0.5405333042144775,
      "learning_rate": 2e-05,
      "loss": 2.7717,
      "step": 1
    },
    {
      "epoch": 0.003714020427112349,
      "eval_loss": 2.7589731216430664,
      "eval_runtime": 7.8472,
      "eval_samples_per_second": 14.528,
      "eval_steps_per_second": 7.264,
      "step": 1
    },
    {
      "epoch": 0.007428040854224698,
      "grad_norm": 0.5611510276794434,
      "learning_rate": 4e-05,
      "loss": 2.8894,
      "step": 2
    },
    {
      "epoch": 0.011142061281337047,
      "grad_norm": 0.5387593507766724,
      "learning_rate": 6e-05,
      "loss": 2.8945,
      "step": 3
    },
    {
      "epoch": 0.011142061281337047,
      "eval_loss": 2.75407338142395,
      "eval_runtime": 6.482,
      "eval_samples_per_second": 17.587,
      "eval_steps_per_second": 8.794,
      "step": 3
    }
  ],
  "logging_steps": 1,
  "max_steps": 10,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 1,
  "save_steps": 3,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": false
      },
      "attributes": {}
    }
  },
  "total_flos": 1109738548887552.0,
  "train_batch_size": 2,
  "trial_name": null,
  "trial_params": null
}