trainer_state.json

{
  "best_metric": 0.3441010117530823,
  "best_model_checkpoint": "bert_base_lda_5_v1_book_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": 2.5492544174194336,
      "learning_rate": 4.9e-05,
      "loss": 0.4848,
      "step": 410
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.8343401061687717,
      "eval_loss": 0.3843107223510742,
      "eval_runtime": 3.5761,
      "eval_samples_per_second": 1527.636,
      "eval_steps_per_second": 6.152,
      "step": 410
    },
    {
      "epoch": 2.0,
      "grad_norm": 4.246053218841553,
      "learning_rate": 4.8e-05,
      "loss": 0.3412,
      "step": 820
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.8544755628775398,
      "eval_loss": 0.3441010117530823,
      "eval_runtime": 3.6463,
      "eval_samples_per_second": 1498.245,
      "eval_steps_per_second": 6.034,
      "step": 820
    },
    {
      "epoch": 3.0,
      "grad_norm": 5.30092716217041,
      "learning_rate": 4.7e-05,
      "loss": 0.2379,
      "step": 1230
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.85667215815486,
      "eval_loss": 0.3484192490577698,
      "eval_runtime": 3.5654,
      "eval_samples_per_second": 1532.217,
      "eval_steps_per_second": 6.17,
      "step": 1230
    },
    {
      "epoch": 4.0,
      "grad_norm": 4.396666526794434,
      "learning_rate": 4.600000000000001e-05,
      "loss": 0.1558,
      "step": 1640
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.8389163463298553,
      "eval_loss": 0.4953833520412445,
      "eval_runtime": 3.6897,
      "eval_samples_per_second": 1480.61,
      "eval_steps_per_second": 5.963,
      "step": 1640
    },
    {
      "epoch": 5.0,
      "grad_norm": 5.491185665130615,
      "learning_rate": 4.5e-05,
      "loss": 0.1041,
      "step": 2050
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.8376349990847519,
      "eval_loss": 0.500628650188446,
      "eval_runtime": 3.8009,
      "eval_samples_per_second": 1437.283,
      "eval_steps_per_second": 5.788,
      "step": 2050
    },
    {
      "epoch": 6.0,
      "grad_norm": 10.016534805297852,
      "learning_rate": 4.4000000000000006e-05,
      "loss": 0.0824,
      "step": 2460
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.8531942156324364,
      "eval_loss": 0.5768154263496399,
      "eval_runtime": 3.5644,
      "eval_samples_per_second": 1532.654,
      "eval_steps_per_second": 6.172,
      "step": 2460
    },
    {
      "epoch": 7.0,
      "grad_norm": 4.23161506652832,
      "learning_rate": 4.3e-05,
      "loss": 0.0601,
      "step": 2870
    },
    {
      "epoch": 7.0,
      "eval_accuracy": 0.8546586124839831,
      "eval_loss": 0.5504027009010315,
      "eval_runtime": 3.4714,
      "eval_samples_per_second": 1573.719,
      "eval_steps_per_second": 6.338,
      "step": 2870
    },
    {
      "epoch": 7.0,
      "step": 2870,
      "total_flos": 9.645664445050368e+16,
      "train_loss": 0.20949012344307186,
      "train_runtime": 1250.1394,
      "train_samples_per_second": 4189.253,
      "train_steps_per_second": 16.398
    }
  ],
  "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
}