microsoft-deberta-v3-large_ner_conll2003
This model is a fine-tuned version of microsoft/deberta-v3-large on the conll2003 dataset. It achieves the following results on the evaluation set:
- Loss: 0.0293
- Precision: 0.9667
- Recall: 0.9724
- F1: 0.9695
- Accuracy: 0.9945
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: 16
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: cosine
- num_epochs: 5
Training results
| Training Loss | Epoch | Step | Validation Loss | Precision | Recall | F1 | Accuracy |
|---|---|---|---|---|---|---|---|
| 0.0986 | 1.0 | 878 | 0.0323 | 0.9453 | 0.9596 | 0.9524 | 0.9921 |
| 0.0212 | 2.0 | 1756 | 0.0270 | 0.9571 | 0.9675 | 0.9623 | 0.9932 |
| 0.009 | 3.0 | 2634 | 0.0280 | 0.9638 | 0.9714 | 0.9676 | 0.9940 |
| 0.0035 | 4.0 | 3512 | 0.0290 | 0.9657 | 0.9712 | 0.9685 | 0.9943 |
| 0.0022 | 5.0 | 4390 | 0.0293 | 0.9667 | 0.9724 | 0.9695 | 0.9945 |
Framework versions
- Transformers 4.20.1
- Pytorch 1.11.0
- Datasets 2.1.0
- Tokenizers 0.12.1