last-checkpoint/trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 0.25,
  "eval_steps": 3,
  "global_step": 3,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.08333333333333333,
      "grad_norm": NaN,
      "learning_rate": 0.0,
      "loss": 1.4924,
      "step": 1
    },
    {
      "epoch": 0.08333333333333333,
      "eval_loss": 1.4975985288619995,
      "eval_runtime": 0.1911,
      "eval_samples_per_second": 15.701,
      "eval_steps_per_second": 15.701,
      "step": 1
    },
    {
      "epoch": 0.16666666666666666,
      "grad_norm": 5.183752536773682,
      "learning_rate": 2e-05,
      "loss": 1.3074,
      "step": 2
    },
    {
      "epoch": 0.25,
      "grad_norm": 6.239664077758789,
      "learning_rate": 4e-05,
      "loss": 1.5739,
      "step": 3
    },
    {
      "epoch": 0.25,
      "eval_loss": 1.4466005563735962,
      "eval_runtime": 0.211,
      "eval_samples_per_second": 14.216,
      "eval_steps_per_second": 14.216,
      "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": 554869274443776.0,
  "train_batch_size": 1,
  "trial_name": null,
  "trial_params": null
}