roberta-base-qqp
This model is a fine-tuned version of roberta-base on the GLUE QQP dataset. It achieves the following results on the evaluation set:
- Loss: 0.4435
- Accuracy: 0.9153
- F1: 0.8867
- Combined Score: 0.9010
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: 2e-05
- train_batch_size: 16
- eval_batch_size: 8
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_ratio: 0.06
- num_epochs: 10.0
Training results
| Training Loss | Epoch | Step | Validation Loss | Accuracy | F1 | Combined Score |
|---|---|---|---|---|---|---|
| 0.2751 | 1.0 | 22741 | 0.3057 | 0.8905 | 0.8512 | 0.8709 |
| 0.2443 | 2.0 | 45482 | 0.2530 | 0.9005 | 0.8710 | 0.8857 |
| 0.2157 | 3.0 | 68223 | 0.2643 | 0.9070 | 0.8769 | 0.8919 |
| 0.1838 | 4.0 | 90964 | 0.2806 | 0.9109 | 0.8815 | 0.8962 |
| 0.146 | 5.0 | 113705 | 0.3277 | 0.9113 | 0.8809 | 0.8961 |
| 0.1262 | 6.0 | 136446 | 0.3939 | 0.9113 | 0.8812 | 0.8962 |
| 0.0867 | 7.0 | 159187 | 0.4435 | 0.9153 | 0.8867 | 0.9010 |
| 0.0757 | 8.0 | 181928 | 0.4812 | 0.9147 | 0.8844 | 0.8996 |
| 0.0479 | 9.0 | 204669 | 0.5081 | 0.9151 | 0.8871 | 0.9011 |
| 0.0379 | 10.0 | 227410 | 0.5647 | 0.9149 | 0.8858 | 0.9003 |
Framework versions
- Transformers 4.20.0.dev0
- Pytorch 1.11.0+cu113
- Datasets 2.1.0
- Tokenizers 0.12.1