trainer_state.json

{
  "best_metric": 0.5835913370480271,
  "best_model_checkpoint": "output/fine_tuned/t5-base/CoLA/checkpoint-1876",
  "epoch": 10.0,
  "eval_steps": 500,
  "global_step": 2680,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.0,
      "eval_loss": 0.47356653213500977,
      "eval_matthews_correlation": 0.48848785650378773,
      "eval_runtime": 4.2201,
      "eval_samples_per_second": 247.152,
      "eval_steps_per_second": 31.042,
      "step": 268
    },
    {
      "epoch": 1.8656716417910446,
      "grad_norm": 5.837749481201172,
      "learning_rate": 4.067164179104478e-05,
      "loss": 0.4668,
      "step": 500
    },
    {
      "epoch": 2.0,
      "eval_loss": 0.498794287443161,
      "eval_matthews_correlation": 0.5416469931221344,
      "eval_runtime": 3.8345,
      "eval_samples_per_second": 272.005,
      "eval_steps_per_second": 34.164,
      "step": 536
    },
    {
      "epoch": 3.0,
      "eval_loss": 0.5234289765357971,
      "eval_matthews_correlation": 0.5495093920893817,
      "eval_runtime": 3.8772,
      "eval_samples_per_second": 269.01,
      "eval_steps_per_second": 33.787,
      "step": 804
    },
    {
      "epoch": 3.7313432835820897,
      "grad_norm": 3.0416481494903564,
      "learning_rate": 3.1343283582089554e-05,
      "loss": 0.293,
      "step": 1000
    },
    {
      "epoch": 4.0,
      "eval_loss": 0.683355450630188,
      "eval_matthews_correlation": 0.5434890534893785,
      "eval_runtime": 3.8435,
      "eval_samples_per_second": 271.364,
      "eval_steps_per_second": 34.083,
      "step": 1072
    },
    {
      "epoch": 5.0,
      "eval_loss": 0.6669164896011353,
      "eval_matthews_correlation": 0.565579540997454,
      "eval_runtime": 3.9128,
      "eval_samples_per_second": 266.561,
      "eval_steps_per_second": 33.48,
      "step": 1340
    },
    {
      "epoch": 5.597014925373134,
      "grad_norm": 3.6730122566223145,
      "learning_rate": 2.201492537313433e-05,
      "loss": 0.2049,
      "step": 1500
    },
    {
      "epoch": 6.0,
      "eval_loss": 0.6630592346191406,
      "eval_matthews_correlation": 0.5807484475986369,
      "eval_runtime": 3.8474,
      "eval_samples_per_second": 271.094,
      "eval_steps_per_second": 34.049,
      "step": 1608
    },
    {
      "epoch": 7.0,
      "eval_loss": 0.7744859457015991,
      "eval_matthews_correlation": 0.5835913370480271,
      "eval_runtime": 2.7082,
      "eval_samples_per_second": 385.127,
      "eval_steps_per_second": 48.372,
      "step": 1876
    },
    {
      "epoch": 7.462686567164179,
      "grad_norm": 5.858553409576416,
      "learning_rate": 1.2686567164179105e-05,
      "loss": 0.1511,
      "step": 2000
    },
    {
      "epoch": 8.0,
      "eval_loss": 0.8352206349372864,
      "eval_matthews_correlation": 0.5780839016534659,
      "eval_runtime": 2.3839,
      "eval_samples_per_second": 437.524,
      "eval_steps_per_second": 54.953,
      "step": 2144
    },
    {
      "epoch": 9.0,
      "eval_loss": 0.8335957527160645,
      "eval_matthews_correlation": 0.567550266689718,
      "eval_runtime": 2.3912,
      "eval_samples_per_second": 436.19,
      "eval_steps_per_second": 54.785,
      "step": 2412
    },
    {
      "epoch": 9.328358208955224,
      "grad_norm": 2.9813477993011475,
      "learning_rate": 3.358208955223881e-06,
      "loss": 0.128,
      "step": 2500
    },
    {
      "epoch": 10.0,
      "eval_loss": 0.8593345880508423,
      "eval_matthews_correlation": 0.5624297390215861,
      "eval_runtime": 2.9658,
      "eval_samples_per_second": 351.678,
      "eval_steps_per_second": 44.171,
      "step": 2680
    },
    {
      "epoch": 10.0,
      "step": 2680,
      "total_flos": 1.30568834773248e+16,
      "train_loss": 0.05434396516031294,
      "train_runtime": 200.6705,
      "train_samples_per_second": 426.121,
      "train_steps_per_second": 13.355
    }
  ],
  "logging_steps": 500,
  "max_steps": 2680,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 10,
  "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": 1.30568834773248e+16,
  "train_batch_size": 32,
  "trial_name": null,
  "trial_params": null
}