trainer_state.json

{
  "best_global_step": 820,
  "best_metric": 0.6435126066207886,
  "best_model_checkpoint": "tiny_bert_km_20_v1_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": 1.436014175415039,
      "learning_rate": 4.9e-05,
      "loss": 0.6683,
      "step": 410
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.6099212886692293,
      "eval_loss": 0.6518413424491882,
      "eval_runtime": 1.6737,
      "eval_samples_per_second": 3264.029,
      "eval_steps_per_second": 13.145,
      "step": 410
    },
    {
      "epoch": 2.0,
      "grad_norm": 1.2714914083480835,
      "learning_rate": 4.8e-05,
      "loss": 0.6405,
      "step": 820
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.6223686619073769,
      "eval_loss": 0.6435126066207886,
      "eval_runtime": 1.5977,
      "eval_samples_per_second": 3419.381,
      "eval_steps_per_second": 13.77,
      "step": 820
    },
    {
      "epoch": 3.0,
      "grad_norm": 2.3162267208099365,
      "learning_rate": 4.7e-05,
      "loss": 0.6003,
      "step": 1230
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.6284092989200073,
      "eval_loss": 0.6470852494239807,
      "eval_runtime": 1.6026,
      "eval_samples_per_second": 3408.837,
      "eval_steps_per_second": 13.728,
      "step": 1230
    },
    {
      "epoch": 4.0,
      "grad_norm": 2.326233386993408,
      "learning_rate": 4.600000000000001e-05,
      "loss": 0.5376,
      "step": 1640
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.6183415705656233,
      "eval_loss": 0.6951343417167664,
      "eval_runtime": 1.6123,
      "eval_samples_per_second": 3388.237,
      "eval_steps_per_second": 13.645,
      "step": 1640
    },
    {
      "epoch": 5.0,
      "grad_norm": 4.237625598907471,
      "learning_rate": 4.5e-05,
      "loss": 0.4637,
      "step": 2050
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.6168771737140765,
      "eval_loss": 0.7689837217330933,
      "eval_runtime": 1.6005,
      "eval_samples_per_second": 3413.229,
      "eval_steps_per_second": 13.745,
      "step": 2050
    },
    {
      "epoch": 6.0,
      "grad_norm": 6.307756423950195,
      "learning_rate": 4.4000000000000006e-05,
      "loss": 0.3921,
      "step": 2460
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.6075416437854658,
      "eval_loss": 0.9081873893737793,
      "eval_runtime": 1.6248,
      "eval_samples_per_second": 3362.28,
      "eval_steps_per_second": 13.54,
      "step": 2460
    },
    {
      "epoch": 7.0,
      "grad_norm": 7.064264297485352,
      "learning_rate": 4.3e-05,
      "loss": 0.3267,
      "step": 2870
    },
    {
      "epoch": 7.0,
      "eval_accuracy": 0.6036976020501555,
      "eval_loss": 1.0488042831420898,
      "eval_runtime": 1.6117,
      "eval_samples_per_second": 3389.489,
      "eval_steps_per_second": 13.65,
      "step": 2870
    },
    {
      "epoch": 7.0,
      "step": 2870,
      "total_flos": 1.922715369529651e+16,
      "train_loss": 0.518453240477665,
      "train_runtime": 405.2301,
      "train_samples_per_second": 12923.892,
      "train_steps_per_second": 50.589
    }
  ],
  "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
}