trainer_state.json

{
  "best_global_step": 820,
  "best_metric": 0.6467869877815247,
  "best_model_checkpoint": "tiny_bert_rand_5_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": 1.361190676689148,
      "learning_rate": 4.9e-05,
      "loss": 0.6655,
      "step": 410
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.6207212154493867,
      "eval_loss": 0.6482405066490173,
      "eval_runtime": 1.6434,
      "eval_samples_per_second": 3324.203,
      "eval_steps_per_second": 13.387,
      "step": 410
    },
    {
      "epoch": 2.0,
      "grad_norm": 1.4431250095367432,
      "learning_rate": 4.8e-05,
      "loss": 0.636,
      "step": 820
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.6267618524620172,
      "eval_loss": 0.6467869877815247,
      "eval_runtime": 1.5553,
      "eval_samples_per_second": 3512.524,
      "eval_steps_per_second": 14.145,
      "step": 820
    },
    {
      "epoch": 3.0,
      "grad_norm": 2.969020366668701,
      "learning_rate": 4.7e-05,
      "loss": 0.5935,
      "step": 1230
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.6231008603331503,
      "eval_loss": 0.6598222255706787,
      "eval_runtime": 1.5702,
      "eval_samples_per_second": 3479.09,
      "eval_steps_per_second": 14.011,
      "step": 1230
    },
    {
      "epoch": 4.0,
      "grad_norm": 3.04614520072937,
      "learning_rate": 4.600000000000001e-05,
      "loss": 0.5345,
      "step": 1640
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.624565257184697,
      "eval_loss": 0.6995676159858704,
      "eval_runtime": 1.5541,
      "eval_samples_per_second": 3515.33,
      "eval_steps_per_second": 14.157,
      "step": 1640
    },
    {
      "epoch": 5.0,
      "grad_norm": 3.5794517993927,
      "learning_rate": 4.5e-05,
      "loss": 0.4697,
      "step": 2050
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.6143144792238696,
      "eval_loss": 0.7826966047286987,
      "eval_runtime": 1.5785,
      "eval_samples_per_second": 3460.942,
      "eval_steps_per_second": 13.938,
      "step": 2050
    },
    {
      "epoch": 6.0,
      "grad_norm": 6.8822221755981445,
      "learning_rate": 4.4000000000000006e-05,
      "loss": 0.4077,
      "step": 2460
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.6040637012630423,
      "eval_loss": 0.8890290260314941,
      "eval_runtime": 1.5923,
      "eval_samples_per_second": 3430.785,
      "eval_steps_per_second": 13.816,
      "step": 2460
    },
    {
      "epoch": 7.0,
      "grad_norm": 6.8207197189331055,
      "learning_rate": 4.3e-05,
      "loss": 0.3487,
      "step": 2870
    },
    {
      "epoch": 7.0,
      "eval_accuracy": 0.6042467508694856,
      "eval_loss": 1.0397076606750488,
      "eval_runtime": 1.5548,
      "eval_samples_per_second": 3513.73,
      "eval_steps_per_second": 14.15,
      "step": 2870
    },
    {
      "epoch": 7.0,
      "step": 2870,
      "total_flos": 1.922715369529651e+16,
      "train_loss": 0.522213314468437,
      "train_runtime": 400.0579,
      "train_samples_per_second": 13090.981,
      "train_steps_per_second": 51.243
    }
  ],
  "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": 1.922715369529651e+16,
  "train_batch_size": 256,
  "trial_name": null,
  "trial_params": null
}