trainer_state.json

{
  "best_global_step": 30,
  "best_metric": 0.5974180698394775,
  "best_model_checkpoint": "distilbert_km_50_v2_mrpc/checkpoint-30",
  "epoch": 7.0,
  "eval_steps": 500,
  "global_step": 105,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.0,
      "grad_norm": 1.1845976114273071,
      "learning_rate": 4.9e-05,
      "loss": 0.6312,
      "step": 15
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.7009803921568627,
      "eval_combined_score": 0.7580659536541889,
      "eval_f1": 0.8151515151515152,
      "eval_loss": 0.6015780568122864,
      "eval_runtime": 0.1836,
      "eval_samples_per_second": 2222.494,
      "eval_steps_per_second": 10.895,
      "step": 15
    },
    {
      "epoch": 2.0,
      "grad_norm": 1.5021694898605347,
      "learning_rate": 4.8e-05,
      "loss": 0.5856,
      "step": 30
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.7009803921568627,
      "eval_combined_score": 0.7577850289051791,
      "eval_f1": 0.8145896656534954,
      "eval_loss": 0.5974180698394775,
      "eval_runtime": 0.2164,
      "eval_samples_per_second": 1885.045,
      "eval_steps_per_second": 9.24,
      "step": 30
    },
    {
      "epoch": 3.0,
      "grad_norm": 1.575913667678833,
      "learning_rate": 4.7e-05,
      "loss": 0.5365,
      "step": 45
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.7083333333333334,
      "eval_combined_score": 0.762203760947965,
      "eval_f1": 0.8160741885625966,
      "eval_loss": 0.6000779271125793,
      "eval_runtime": 0.2131,
      "eval_samples_per_second": 1914.515,
      "eval_steps_per_second": 9.385,
      "step": 45
    },
    {
      "epoch": 4.0,
      "grad_norm": 1.5832329988479614,
      "learning_rate": 4.600000000000001e-05,
      "loss": 0.4884,
      "step": 60
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.696078431372549,
      "eval_combined_score": 0.7457289846631722,
      "eval_f1": 0.7953795379537953,
      "eval_loss": 0.6059989333152771,
      "eval_runtime": 0.209,
      "eval_samples_per_second": 1952.373,
      "eval_steps_per_second": 9.57,
      "step": 60
    },
    {
      "epoch": 5.0,
      "grad_norm": 2.5107924938201904,
      "learning_rate": 4.5e-05,
      "loss": 0.4025,
      "step": 75
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.6764705882352942,
      "eval_combined_score": 0.7228506787330318,
      "eval_f1": 0.7692307692307693,
      "eval_loss": 0.6637485027313232,
      "eval_runtime": 0.1884,
      "eval_samples_per_second": 2165.429,
      "eval_steps_per_second": 10.615,
      "step": 75
    },
    {
      "epoch": 6.0,
      "grad_norm": 2.7780885696411133,
      "learning_rate": 4.4000000000000006e-05,
      "loss": 0.3023,
      "step": 90
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.6666666666666666,
      "eval_combined_score": 0.7123368920521945,
      "eval_f1": 0.7580071174377224,
      "eval_loss": 0.7731847763061523,
      "eval_runtime": 0.1968,
      "eval_samples_per_second": 2072.789,
      "eval_steps_per_second": 10.161,
      "step": 90
    },
    {
      "epoch": 7.0,
      "grad_norm": 2.5387532711029053,
      "learning_rate": 4.3e-05,
      "loss": 0.1918,
      "step": 105
    },
    {
      "epoch": 7.0,
      "eval_accuracy": 0.6568627450980392,
      "eval_combined_score": 0.7060537501713973,
      "eval_f1": 0.7552447552447552,
      "eval_loss": 0.9304187297821045,
      "eval_runtime": 0.2027,
      "eval_samples_per_second": 2012.998,
      "eval_steps_per_second": 9.868,
      "step": 105
    },
    {
      "epoch": 7.0,
      "step": 105,
      "total_flos": 1700616463945728.0,
      "train_loss": 0.44832047053745816,
      "train_runtime": 35.4144,
      "train_samples_per_second": 5178.678,
      "train_steps_per_second": 21.178
    }
  ],
  "logging_steps": 1,
  "max_steps": 750,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 50,
  "save_steps": 500,
  "stateful_callbacks": {
    "EarlyStoppingCallback": {
      "args": {
        "early_stopping_patience": 5,
        "early_stopping_threshold": 0.0
      },
      "attributes": {
        "early_stopping_patience_counter": 5
      }
    },
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": true
      },
      "attributes": {}
    }
  },
  "total_flos": 1700616463945728.0,
  "train_batch_size": 256,
  "trial_name": null,
  "trial_params": null
}