Edit model card

hBERTv1_data_aug_qqp

This model is a fine-tuned version of gokuls/bert_12_layer_model_v1 on the GLUE QQP dataset. It achieves the following results on the evaluation set:

  • Loss: 0.5769
  • Accuracy: 0.8162
  • F1: 0.7679
  • Combined Score: 0.7920

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 F1 Combined Score
0.2419 1.0 29671 0.5769 0.8162 0.7679 0.7920
0.104 2.0 59342 0.6327 0.8272 0.7769 0.8020
0.0911 3.0 89013 nan 0.6318 0.0 0.3159
0.0 4.0 118684 nan 0.6318 0.0 0.3159
0.0 5.0 148355 nan 0.6318 0.0 0.3159
0.0 6.0 178026 nan 0.6318 0.0 0.3159

Framework versions

  • Transformers 4.26.1
  • Pytorch 1.14.0a0+410ce96
  • Datasets 2.10.1
  • Tokenizers 0.13.2
Downloads last month
2
Inference Examples
This model does not have enough activity to be deployed to Inference API (serverless) yet. Increase its social visibility and check back later, or deploy to Inference Endpoints (dedicated) instead.

Dataset used to train gokuls/hBERTv1_data_aug_qqp

Evaluation results