trainer_state.json

{
  "best_global_step": 820,
  "best_metric": 0.6360426545143127,
  "best_model_checkpoint": "bert_base_km_20_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": 1.8109616041183472,
      "learning_rate": 4.9e-05,
      "loss": 0.6649,
      "step": 410
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.628775398132894,
      "eval_loss": 0.6421313285827637,
      "eval_runtime": 3.4615,
      "eval_samples_per_second": 1578.236,
      "eval_steps_per_second": 6.356,
      "step": 410
    },
    {
      "epoch": 2.0,
      "grad_norm": 1.7537949085235596,
      "learning_rate": 4.8e-05,
      "loss": 0.6275,
      "step": 820
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.6421380194032583,
      "eval_loss": 0.6360426545143127,
      "eval_runtime": 3.4526,
      "eval_samples_per_second": 1582.295,
      "eval_steps_per_second": 6.372,
      "step": 820
    },
    {
      "epoch": 3.0,
      "grad_norm": 2.509582996368408,
      "learning_rate": 4.7e-05,
      "loss": 0.5638,
      "step": 1230
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.6313380926231008,
      "eval_loss": 0.6697062253952026,
      "eval_runtime": 3.4731,
      "eval_samples_per_second": 1572.931,
      "eval_steps_per_second": 6.334,
      "step": 1230
    },
    {
      "epoch": 4.0,
      "grad_norm": 4.651081562042236,
      "learning_rate": 4.600000000000001e-05,
      "loss": 0.4536,
      "step": 1640
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.6465312099578986,
      "eval_loss": 0.6980166435241699,
      "eval_runtime": 3.4663,
      "eval_samples_per_second": 1576.024,
      "eval_steps_per_second": 6.347,
      "step": 1640
    },
    {
      "epoch": 5.0,
      "grad_norm": 5.6104278564453125,
      "learning_rate": 4.5e-05,
      "loss": 0.3276,
      "step": 2050
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.6423210690097017,
      "eval_loss": 0.8002541065216064,
      "eval_runtime": 3.4467,
      "eval_samples_per_second": 1584.977,
      "eval_steps_per_second": 6.383,
      "step": 2050
    },
    {
      "epoch": 6.0,
      "grad_norm": 9.524043083190918,
      "learning_rate": 4.4000000000000006e-05,
      "loss": 0.2264,
      "step": 2460
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.6353651839648545,
      "eval_loss": 0.9973970055580139,
      "eval_runtime": 3.4687,
      "eval_samples_per_second": 1574.955,
      "eval_steps_per_second": 6.342,
      "step": 2460
    },
    {
      "epoch": 7.0,
      "grad_norm": 9.233513832092285,
      "learning_rate": 4.3e-05,
      "loss": 0.1588,
      "step": 2870
    },
    {
      "epoch": 7.0,
      "eval_accuracy": 0.6368295808164013,
      "eval_loss": 1.213170051574707,
      "eval_runtime": 3.4569,
      "eval_samples_per_second": 1580.335,
      "eval_steps_per_second": 6.364,
      "step": 2870
    },
    {
      "epoch": 7.0,
      "step": 2870,
      "total_flos": 9.645664445050368e+16,
      "train_loss": 0.4317959802075961,
      "train_runtime": 1214.1945,
      "train_samples_per_second": 4313.271,
      "train_steps_per_second": 16.884
    }
  ],
  "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
}