trainer_state.json

{
  "best_global_step": 820,
  "best_metric": 0.6395706534385681,
  "best_model_checkpoint": "bert_base_rand_100_v2_qnli/checkpoint-820",
  "epoch": 7.0,
  "eval_steps": 500,
  "global_step": 2870,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.0,
      "grad_norm": 2.1673314571380615,
      "learning_rate": 4.9e-05,
      "loss": 0.6628,
      "step": 410
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.6247483067911404,
      "eval_loss": 0.641835629940033,
      "eval_runtime": 3.504,
      "eval_samples_per_second": 1559.097,
      "eval_steps_per_second": 6.279,
      "step": 410
    },
    {
      "epoch": 2.0,
      "grad_norm": 1.7349144220352173,
      "learning_rate": 4.8e-05,
      "loss": 0.6243,
      "step": 820
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.637195680029288,
      "eval_loss": 0.6395706534385681,
      "eval_runtime": 3.4837,
      "eval_samples_per_second": 1568.164,
      "eval_steps_per_second": 6.315,
      "step": 820
    },
    {
      "epoch": 3.0,
      "grad_norm": 3.738842010498047,
      "learning_rate": 4.7e-05,
      "loss": 0.559,
      "step": 1230
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.6322533406553176,
      "eval_loss": 0.6659859418869019,
      "eval_runtime": 3.4904,
      "eval_samples_per_second": 1565.145,
      "eval_steps_per_second": 6.303,
      "step": 1230
    },
    {
      "epoch": 4.0,
      "grad_norm": 5.834431171417236,
      "learning_rate": 4.600000000000001e-05,
      "loss": 0.4565,
      "step": 1640
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.6456159619256818,
      "eval_loss": 0.7500161528587341,
      "eval_runtime": 3.4875,
      "eval_samples_per_second": 1566.456,
      "eval_steps_per_second": 6.308,
      "step": 1640
    },
    {
      "epoch": 5.0,
      "grad_norm": 5.9190897941589355,
      "learning_rate": 4.5e-05,
      "loss": 0.3424,
      "step": 2050
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.6518396485447556,
      "eval_loss": 0.7934035062789917,
      "eval_runtime": 3.4973,
      "eval_samples_per_second": 1562.055,
      "eval_steps_per_second": 6.291,
      "step": 2050
    },
    {
      "epoch": 6.0,
      "grad_norm": 7.918484687805176,
      "learning_rate": 4.4000000000000006e-05,
      "loss": 0.2421,
      "step": 2460
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.6560497894929526,
      "eval_loss": 0.9496882557868958,
      "eval_runtime": 3.4778,
      "eval_samples_per_second": 1570.837,
      "eval_steps_per_second": 6.326,
      "step": 2460
    },
    {
      "epoch": 7.0,
      "grad_norm": 9.262678146362305,
      "learning_rate": 4.3e-05,
      "loss": 0.1726,
      "step": 2870
    },
    {
      "epoch": 7.0,
      "eval_accuracy": 0.6501922020867655,
      "eval_loss": 1.1923950910568237,
      "eval_runtime": 3.4776,
      "eval_samples_per_second": 1570.922,
      "eval_steps_per_second": 6.326,
      "step": 2870
    },
    {
      "epoch": 7.0,
      "step": 2870,
      "total_flos": 9.645664445050368e+16,
      "train_loss": 0.4371023889202689,
      "train_runtime": 1230.8438,
      "train_samples_per_second": 4254.927,
      "train_steps_per_second": 16.655
    }
  ],
  "logging_steps": 1,
  "max_steps": 20500,
  "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": 9.645664445050368e+16,
  "train_batch_size": 256,
  "trial_name": null,
  "trial_params": null
}