trainer_state.json

{
  "best_metric": 0.3569449782371521,
  "best_model_checkpoint": "bert_base_lda_100_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": 3.559905529022217,
      "learning_rate": 4.9e-05,
      "loss": 0.4858,
      "step": 410
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.8255537250594911,
      "eval_loss": 0.4045892357826233,
      "eval_runtime": 4.1253,
      "eval_samples_per_second": 1324.281,
      "eval_steps_per_second": 5.333,
      "step": 410
    },
    {
      "epoch": 2.0,
      "grad_norm": 4.1515703201293945,
      "learning_rate": 4.8e-05,
      "loss": 0.3624,
      "step": 820
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.8480688266520227,
      "eval_loss": 0.3569449782371521,
      "eval_runtime": 4.0812,
      "eval_samples_per_second": 1338.562,
      "eval_steps_per_second": 5.391,
      "step": 820
    },
    {
      "epoch": 3.0,
      "grad_norm": 5.852153301239014,
      "learning_rate": 4.7e-05,
      "loss": 0.2686,
      "step": 1230
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.8511806699615596,
      "eval_loss": 0.363815575838089,
      "eval_runtime": 4.0619,
      "eval_samples_per_second": 1344.949,
      "eval_steps_per_second": 5.416,
      "step": 1230
    },
    {
      "epoch": 4.0,
      "grad_norm": 6.573513507843018,
      "learning_rate": 4.600000000000001e-05,
      "loss": 0.1868,
      "step": 1640
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.8284825187625846,
      "eval_loss": 0.47949424386024475,
      "eval_runtime": 4.0219,
      "eval_samples_per_second": 1358.314,
      "eval_steps_per_second": 5.47,
      "step": 1640
    },
    {
      "epoch": 5.0,
      "grad_norm": 9.13730525970459,
      "learning_rate": 4.5e-05,
      "loss": 0.128,
      "step": 2050
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.8396485447556288,
      "eval_loss": 0.49180516600608826,
      "eval_runtime": 3.9428,
      "eval_samples_per_second": 1385.574,
      "eval_steps_per_second": 5.58,
      "step": 2050
    },
    {
      "epoch": 6.0,
      "grad_norm": 5.090904712677002,
      "learning_rate": 4.4000000000000006e-05,
      "loss": 0.0923,
      "step": 2460
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.8460552809811459,
      "eval_loss": 0.5746628046035767,
      "eval_runtime": 3.9796,
      "eval_samples_per_second": 1372.74,
      "eval_steps_per_second": 5.528,
      "step": 2460
    },
    {
      "epoch": 7.0,
      "grad_norm": 2.029569149017334,
      "learning_rate": 4.3e-05,
      "loss": 0.0706,
      "step": 2870
    },
    {
      "epoch": 7.0,
      "eval_accuracy": 0.8416620904265056,
      "eval_loss": 0.5756229758262634,
      "eval_runtime": 3.9553,
      "eval_samples_per_second": 1381.196,
      "eval_steps_per_second": 5.562,
      "step": 2870
    },
    {
      "epoch": 7.0,
      "step": 2870,
      "total_flos": 9.645664445050368e+16,
      "train_loss": 0.22778432626757472,
      "train_runtime": 1251.5289,
      "train_samples_per_second": 4184.602,
      "train_steps_per_second": 16.38
    }
  ],
  "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
}