multilingual-microsoft/deberta-v3-base-hausa-ner-v1
This model is a fine-tuned version of microsoft/deberta-v3-base on the Beijuka/Multilingual_PII_NER_dataset dataset. It achieves the following results on the evaluation set:
- Loss: 0.0729
- Precision: 0.9414
- Recall: 0.9395
- F1: 0.9405
- Accuracy: 0.9835
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: 5e-05
- train_batch_size: 8
- eval_batch_size: 8
- seed: 42
- optimizer: Use OptimizerNames.ADAMW_TORCH_FUSED with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
- lr_scheduler_type: linear
- num_epochs: 20
Training results
| Training Loss | Epoch | Step | Validation Loss | Precision | Recall | F1 | Accuracy |
|---|---|---|---|---|---|---|---|
| No log | 1.0 | 301 | 0.0915 | 0.8913 | 0.9010 | 0.8961 | 0.9739 |
| 0.16 | 2.0 | 602 | 0.0919 | 0.8879 | 0.9251 | 0.9061 | 0.9755 |
| 0.16 | 3.0 | 903 | 0.0760 | 0.8694 | 0.9429 | 0.9047 | 0.9758 |
| 0.0638 | 4.0 | 1204 | 0.0954 | 0.8875 | 0.9365 | 0.9113 | 0.9782 |
| 0.0475 | 5.0 | 1505 | 0.0770 | 0.9158 | 0.9257 | 0.9207 | 0.9784 |
| 0.0475 | 6.0 | 1806 | 0.0911 | 0.9120 | 0.9283 | 0.9201 | 0.9795 |
| 0.0355 | 7.0 | 2107 | 0.0878 | 0.8870 | 0.9371 | 0.9114 | 0.9771 |
| 0.0355 | 8.0 | 2408 | 0.1145 | 0.8882 | 0.9435 | 0.9150 | 0.9788 |
Framework versions
- Transformers 4.55.4
- Pytorch 2.8.0+cu126
- Datasets 4.0.0
- Tokenizers 0.21.4