multilingual-microsoft/deberta-v3-base-kanuri-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: 0.0800
- Precision: 0.9415
- Recall: 0.9415
- F1: 0.9415
- Accuracy: 0.9855
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 | 0.0974 | 0.8852 | 0.8857 | 0.8854 | 0.9718 |
| 0.1637 | 2.0 | 602 | 0.0877 | 0.8894 | 0.9194 | 0.9042 | 0.9756 |
| 0.1637 | 3.0 | 903 | 0.0788 | 0.8860 | 0.9276 | 0.9063 | 0.9758 |
| 0.0643 | 4.0 | 1204 | 0.1024 | 0.8899 | 0.9238 | 0.9065 | 0.9772 |
| 0.0463 | 5.0 | 1505 | 0.0785 | 0.9248 | 0.9130 | 0.9188 | 0.9774 |
| 0.0463 | 6.0 | 1806 | 0.0940 | 0.9132 | 0.9289 | 0.9210 | 0.9795 |
| 0.0316 | 7.0 | 2107 | 0.1033 | 0.8974 | 0.9276 | 0.9123 | 0.9770 |
| 0.0316 | 8.0 | 2408 | 0.1152 | 0.8884 | 0.9302 | 0.9088 | 0.9781 |
| 0.0179 | 9.0 | 2709 | 0.1308 | 0.8975 | 0.9289 | 0.9129 | 0.9782 |
Framework versions
- Transformers 4.55.4
- Pytorch 2.8.0+cu126
- Datasets 4.0.0
- Tokenizers 0.21.4