hBERTv2_qnli / README.md
gokuls's picture
End of training
e95d10f
metadata
language:
  - en
tags:
  - generated_from_trainer
datasets:
  - glue
metrics:
  - accuracy
model-index:
  - name: hBERTv2_qnli
    results:
      - task:
          name: Text Classification
          type: text-classification
        dataset:
          name: GLUE QNLI
          type: glue
          config: qnli
          split: validation
          args: qnli
        metrics:
          - name: Accuracy
            type: accuracy
            value: 0.5053999633900788

hBERTv2_qnli

This model is a fine-tuned version of gokuls/bert_12_layer_model_v2 on the GLUE QNLI dataset. It achieves the following results on the evaluation set:

  • Loss: 0.6930
  • Accuracy: 0.5054

Model description

More information needed

Intended uses & limitations

More information needed

Training and evaluation data

More information needed

Training procedure

Training hyperparameters

The following hyperparameters were used during training:

  • learning_rate: 5e-05
  • train_batch_size: 256
  • eval_batch_size: 256
  • seed: 10
  • distributed_type: multi-GPU
  • optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
  • lr_scheduler_type: linear
  • num_epochs: 50
  • mixed_precision_training: Native AMP

Training results

Training Loss Epoch Step Validation Loss Accuracy
0.6968 1.0 410 0.6952 0.5054
0.6943 2.0 820 0.6932 0.4946
0.6937 3.0 1230 0.6933 0.5054
0.6934 4.0 1640 0.6931 0.5054
0.6934 5.0 2050 0.6931 0.5054
0.6933 6.0 2460 0.6930 0.5054
0.6933 7.0 2870 0.6931 0.5054
0.6932 8.0 3280 0.6930 0.5054
0.6932 9.0 3690 0.6934 0.4946
0.6932 10.0 4100 0.6930 0.5054
0.6932 11.0 4510 0.6931 0.4946
0.6933 12.0 4920 0.6934 0.4946
0.6932 13.0 5330 0.6931 0.4946

Framework versions

  • Transformers 4.26.1
  • Pytorch 1.14.0a0+410ce96
  • Datasets 2.10.1
  • Tokenizers 0.13.2