trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 10.0,
  "eval_steps": 500,
  "global_step": 30,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.3333333333333333,
      "grad_norm": 19.037089746841605,
      "learning_rate": 3.3333333333333333e-06,
      "loss": 0.969,
      "step": 1
    },
    {
      "epoch": 1.0,
      "eval_loss": 1.2659122943878174,
      "eval_runtime": 2.8561,
      "eval_samples_per_second": 51.469,
      "eval_steps_per_second": 1.751,
      "step": 3
    },
    {
      "epoch": 1.6666666666666665,
      "grad_norm": 17.984437460856608,
      "learning_rate": 9.86522435289912e-06,
      "loss": 1.0058,
      "step": 5
    },
    {
      "epoch": 2.0,
      "eval_loss": 0.7384828329086304,
      "eval_runtime": 2.8559,
      "eval_samples_per_second": 51.473,
      "eval_steps_per_second": 1.751,
      "step": 6
    },
    {
      "epoch": 3.0,
      "eval_loss": 0.4851093292236328,
      "eval_runtime": 2.8507,
      "eval_samples_per_second": 51.566,
      "eval_steps_per_second": 1.754,
      "step": 9
    },
    {
      "epoch": 3.3333333333333335,
      "grad_norm": 5.364449672221281,
      "learning_rate": 8.43120818934367e-06,
      "loss": 0.6526,
      "step": 10
    },
    {
      "epoch": 4.0,
      "eval_loss": 0.31103894114494324,
      "eval_runtime": 2.8499,
      "eval_samples_per_second": 51.58,
      "eval_steps_per_second": 1.754,
      "step": 12
    },
    {
      "epoch": 5.0,
      "grad_norm": 4.475074057863949,
      "learning_rate": 5.8682408883346535e-06,
      "loss": 0.3241,
      "step": 15
    },
    {
      "epoch": 5.0,
      "eval_loss": 0.16107995808124542,
      "eval_runtime": 2.8608,
      "eval_samples_per_second": 51.385,
      "eval_steps_per_second": 1.748,
      "step": 15
    },
    {
      "epoch": 6.0,
      "eval_loss": 0.0703248530626297,
      "eval_runtime": 2.8588,
      "eval_samples_per_second": 51.42,
      "eval_steps_per_second": 1.749,
      "step": 18
    },
    {
      "epoch": 6.666666666666667,
      "grad_norm": 2.7549035025775233,
      "learning_rate": 3.019601169804216e-06,
      "loss": 0.1058,
      "step": 20
    },
    {
      "epoch": 7.0,
      "eval_loss": 0.03494955599308014,
      "eval_runtime": 2.8605,
      "eval_samples_per_second": 51.389,
      "eval_steps_per_second": 1.748,
      "step": 21
    },
    {
      "epoch": 8.0,
      "eval_loss": 0.023995481431484222,
      "eval_runtime": 2.8621,
      "eval_samples_per_second": 51.36,
      "eval_steps_per_second": 1.747,
      "step": 24
    },
    {
      "epoch": 8.333333333333334,
      "grad_norm": 0.8694630292381832,
      "learning_rate": 8.225609429353187e-07,
      "loss": 0.033,
      "step": 25
    },
    {
      "epoch": 9.0,
      "eval_loss": 0.019569795578718185,
      "eval_runtime": 2.8592,
      "eval_samples_per_second": 51.412,
      "eval_steps_per_second": 1.749,
      "step": 27
    },
    {
      "epoch": 10.0,
      "grad_norm": 0.5886770375212557,
      "learning_rate": 0.0,
      "loss": 0.0202,
      "step": 30
    },
    {
      "epoch": 10.0,
      "eval_loss": 0.018579846248030663,
      "eval_runtime": 2.8517,
      "eval_samples_per_second": 51.547,
      "eval_steps_per_second": 1.753,
      "step": 30
    },
    {
      "epoch": 10.0,
      "step": 30,
      "total_flos": 6281389670400.0,
      "train_loss": 0.3556735252340635,
      "train_runtime": 237.9619,
      "train_samples_per_second": 6.177,
      "train_steps_per_second": 0.126
    }
  ],
  "logging_steps": 5,
  "max_steps": 30,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 10,
  "save_steps": 100,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": true
      },
      "attributes": {}
    }
  },
  "total_flos": 6281389670400.0,
  "train_batch_size": 16,
  "trial_name": null,
  "trial_params": null
}