distilbert-base-uncased-finetuned-squad-squad_v2
This model is a fine-tuned version of distilbert-base-uncased on the squad_v2 dataset. It achieves the following results on the evaluation set:
- Loss: 1.8785
- Exact Match: 63.9097
- F1-score: 67.6732
Overview
Language model: distilbert-base-uncased
Language: English
Downstream-task: Extractive QA
Training data: SQuAD 2.0
Eval data: SQuAD 2.0
Training procedure
Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 2e-05
- train_batch_size: 16
- eval_batch_size: 16
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 5
Training results
| Training Loss | Epoch | Step | Validation Loss |
|---|---|---|---|
| 1.2231 | 1.0 | 8235 | 1.2833 |
| 0.9337 | 2.0 | 16470 | 1.2849 |
| 0.7437 | 3.0 | 24705 | 1.4186 |
| 0.5927 | 4.0 | 32940 | 1.6308 |
| 0.4795 | 5.0 | 41175 | 1.8785 |
Framework versions
- Transformers 4.34.0
- Pytorch 1.12.1
- Datasets 2.14.5
- Tokenizers 0.14.1