trainer_state.json

{
  "best_global_step": 820,
  "best_metric": 0.6418310403823853,
  "best_model_checkpoint": "tiny_bert_km_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": 1.9984785318374634,
      "learning_rate": 4.9e-05,
      "loss": 0.6645,
      "step": 410
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.628226249313564,
      "eval_loss": 0.6445777416229248,
      "eval_runtime": 1.6666,
      "eval_samples_per_second": 3277.881,
      "eval_steps_per_second": 13.2,
      "step": 410
    },
    {
      "epoch": 2.0,
      "grad_norm": 1.9222540855407715,
      "learning_rate": 4.8e-05,
      "loss": 0.6396,
      "step": 820
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.629873695771554,
      "eval_loss": 0.6418310403823853,
      "eval_runtime": 1.5593,
      "eval_samples_per_second": 3503.409,
      "eval_steps_per_second": 14.109,
      "step": 820
    },
    {
      "epoch": 3.0,
      "grad_norm": 2.6571438312530518,
      "learning_rate": 4.7e-05,
      "loss": 0.605,
      "step": 1230
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.6154127768625297,
      "eval_loss": 0.6450832486152649,
      "eval_runtime": 1.5874,
      "eval_samples_per_second": 3441.513,
      "eval_steps_per_second": 13.859,
      "step": 1230
    },
    {
      "epoch": 4.0,
      "grad_norm": 3.3628931045532227,
      "learning_rate": 4.600000000000001e-05,
      "loss": 0.5487,
      "step": 1640
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.6291414973457807,
      "eval_loss": 0.6746199727058411,
      "eval_runtime": 1.5663,
      "eval_samples_per_second": 3487.9,
      "eval_steps_per_second": 14.046,
      "step": 1640
    },
    {
      "epoch": 5.0,
      "grad_norm": 3.4197938442230225,
      "learning_rate": 4.5e-05,
      "loss": 0.4892,
      "step": 2050
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.6426871682225883,
      "eval_loss": 0.6786310076713562,
      "eval_runtime": 1.5756,
      "eval_samples_per_second": 3467.339,
      "eval_steps_per_second": 13.963,
      "step": 2050
    },
    {
      "epoch": 6.0,
      "grad_norm": 4.647822856903076,
      "learning_rate": 4.4000000000000006e-05,
      "loss": 0.4329,
      "step": 2460
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.6465312099578986,
      "eval_loss": 0.7132157683372498,
      "eval_runtime": 1.5819,
      "eval_samples_per_second": 3453.419,
      "eval_steps_per_second": 13.907,
      "step": 2460
    },
    {
      "epoch": 7.0,
      "grad_norm": 5.704725742340088,
      "learning_rate": 4.3e-05,
      "loss": 0.3805,
      "step": 2870
    },
    {
      "epoch": 7.0,
      "eval_accuracy": 0.6437854658612484,
      "eval_loss": 0.8527612686157227,
      "eval_runtime": 1.5786,
      "eval_samples_per_second": 3460.745,
      "eval_steps_per_second": 13.937,
      "step": 2870
    },
    {
      "epoch": 7.0,
      "step": 2870,
      "total_flos": 1.922715369529651e+16,
      "train_loss": 0.5372033574439923,
      "train_runtime": 399.9234,
      "train_samples_per_second": 13095.384,
      "train_steps_per_second": 51.26
    }
  ],
  "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
}