trainer_state.json

{
  "best_global_step": null,
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 2.7058823529411766,
  "eval_steps": 500,
  "global_step": 24,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.5882352941176471,
      "grad_norm": 0.8477006798047029,
      "learning_rate": 3.6190476190476195e-05,
      "loss": 1.1601,
      "mean_token_accuracy": 0.7532356321811676,
      "num_tokens": 2505918.0,
      "step": 5
    },
    {
      "epoch": 1.1176470588235294,
      "grad_norm": 0.6034108550215781,
      "learning_rate": 2.6666666666666667e-05,
      "loss": 0.9765,
      "mean_token_accuracy": 0.7825209730201297,
      "num_tokens": 4791423.0,
      "step": 10
    },
    {
      "epoch": 1.7058823529411766,
      "grad_norm": 0.3511641082101572,
      "learning_rate": 1.7142857142857142e-05,
      "loss": 0.8292,
      "mean_token_accuracy": 0.8093127250671387,
      "num_tokens": 7330272.0,
      "step": 15
    },
    {
      "epoch": 2.235294117647059,
      "grad_norm": 0.2702266691100785,
      "learning_rate": 7.61904761904762e-06,
      "loss": 0.8318,
      "mean_token_accuracy": 0.8098728888564639,
      "num_tokens": 9618952.0,
      "step": 20
    },
    {
      "epoch": 2.7058823529411766,
      "mean_token_accuracy": 0.8229228667914867,
      "num_tokens": 11607611.0,
      "step": 24,
      "total_flos": 17258225598464.0,
      "train_loss": 0.8854026099046072,
      "train_runtime": 158.7818,
      "train_samples_per_second": 10.184,
      "train_steps_per_second": 0.151
    }
  ],
  "logging_steps": 5,
  "max_steps": 24,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 3,
  "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": 17258225598464.0,
  "train_batch_size": 2,
  "trial_name": null,
  "trial_params": null
}