Edit model card

hBERTv2_new_pretrain_48_qqp

This model is a fine-tuned version of gokuls/bert_12_layer_model_v2_complete_training_new_48 on the GLUE QQP dataset. It achieves the following results on the evaluation set:

  • Loss: 0.4029
  • Accuracy: 0.8216
  • F1: 0.7562
  • Combined Score: 0.7889

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: 4e-05
  • train_batch_size: 128
  • eval_batch_size: 128
  • 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

Training results

Training Loss Epoch Step Validation Loss Accuracy F1 Combined Score
0.5044 1.0 2843 0.4468 0.7865 0.6961 0.7413
0.4102 2.0 5686 0.4359 0.7992 0.6935 0.7464
0.3553 3.0 8529 0.4127 0.8080 0.7105 0.7592
0.3122 4.0 11372 0.4029 0.8216 0.7562 0.7889
0.2756 5.0 14215 0.4481 0.8228 0.7518 0.7873
0.2479 6.0 17058 0.4778 0.8268 0.7633 0.7951
0.223 7.0 19901 0.4425 0.8158 0.7721 0.7939
0.2028 8.0 22744 0.4705 0.8267 0.7686 0.7977
0.183 9.0 25587 0.4908 0.8301 0.7659 0.7980

Framework versions

  • Transformers 4.29.2
  • Pytorch 1.14.0a0+410ce96
  • Datasets 2.12.0
  • Tokenizers 0.13.3
Downloads last month
3
Inference API
This model can be loaded on Inference API (serverless).

Dataset used to train gokuls/hBERTv2_new_pretrain_48_qqp

Evaluation results