trainer_state.json

{
  "best_global_step": 410,
  "best_metric": 0.6928256154060364,
  "best_model_checkpoint": "bert_base_train_book_ent_15p_s_init_qnli/checkpoint-410",
  "epoch": 6.0,
  "eval_steps": 500,
  "global_step": 2460,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.0,
      "grad_norm": 2.6589386463165283,
      "learning_rate": 4.900243902439024e-05,
      "loss": 0.6973,
      "step": 410
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.5053999633900788,
      "eval_loss": 0.6928256154060364,
      "eval_runtime": 4.6142,
      "eval_samples_per_second": 1183.95,
      "eval_steps_per_second": 4.768,
      "step": 410
    },
    {
      "epoch": 2.0,
      "grad_norm": 1.1992162466049194,
      "learning_rate": 4.8002439024390246e-05,
      "loss": 0.6959,
      "step": 820
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.4946000366099213,
      "eval_loss": 0.6954562067985535,
      "eval_runtime": 4.6225,
      "eval_samples_per_second": 1181.833,
      "eval_steps_per_second": 4.759,
      "step": 820
    },
    {
      "epoch": 3.0,
      "grad_norm": 1.4169254302978516,
      "learning_rate": 4.700243902439024e-05,
      "loss": 0.6957,
      "step": 1230
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.5053999633900788,
      "eval_loss": 0.6980332136154175,
      "eval_runtime": 4.6457,
      "eval_samples_per_second": 1175.926,
      "eval_steps_per_second": 4.736,
      "step": 1230
    },
    {
      "epoch": 4.0,
      "grad_norm": 0.994229793548584,
      "learning_rate": 4.600243902439025e-05,
      "loss": 0.6956,
      "step": 1640
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.5053999633900788,
      "eval_loss": 0.6931659579277039,
      "eval_runtime": 4.6359,
      "eval_samples_per_second": 1178.403,
      "eval_steps_per_second": 4.746,
      "step": 1640
    },
    {
      "epoch": 5.0,
      "grad_norm": 1.0654411315917969,
      "learning_rate": 4.500243902439025e-05,
      "loss": 0.6954,
      "step": 2050
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.5053999633900788,
      "eval_loss": 0.6933382749557495,
      "eval_runtime": 4.6434,
      "eval_samples_per_second": 1176.507,
      "eval_steps_per_second": 4.738,
      "step": 2050
    },
    {
      "epoch": 6.0,
      "grad_norm": 1.196447491645813,
      "learning_rate": 4.400243902439024e-05,
      "loss": 0.6951,
      "step": 2460
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.5053999633900788,
      "eval_loss": 0.6932117342948914,
      "eval_runtime": 4.6641,
      "eval_samples_per_second": 1171.299,
      "eval_steps_per_second": 4.717,
      "step": 2460
    },
    {
      "epoch": 6.0,
      "step": 2460,
      "total_flos": 8.267712381471744e+16,
      "train_loss": 0.6958341645031441,
      "train_runtime": 1153.5282,
      "train_samples_per_second": 4540.114,
      "train_steps_per_second": 17.772
    }
  ],
  "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": 8.267712381471744e+16,
  "train_batch_size": 256,
  "trial_name": null,
  "trial_params": null
}