bert-small-finetuned-ner-to-multilabel-xglue-ner-new
This model is a fine-tuned version of google/bert_uncased_L-4_H-512_A-8 on the None dataset. It achieves the following results on the evaluation set:
- Loss: 0.0525
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: 3e-05
- train_batch_size: 8
- eval_batch_size: 8
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: constant
- num_epochs: 40
Training results
| Training Loss | Epoch | Step | Validation Loss |
|---|---|---|---|
| 0.1842 | 0.28 | 500 | 0.0937 |
| 0.0858 | 0.57 | 1000 | 0.0666 |
| 0.07 | 0.85 | 1500 | 0.0586 |
| 0.0535 | 1.14 | 2000 | 0.0482 |
| 0.0413 | 1.42 | 2500 | 0.0541 |
| 0.0445 | 1.71 | 3000 | 0.0451 |
| 0.0378 | 1.99 | 3500 | 0.0531 |
| 0.0248 | 2.28 | 4000 | 0.0501 |
| 0.0255 | 2.56 | 4500 | 0.0525 |
Framework versions
- Transformers 4.21.2
- Pytorch 1.12.1+cu113
- Datasets 2.4.0
- Tokenizers 0.12.1