multilingual-google-bert/bert-base-multilingual-cased-hausa-ner-v1
This model is a fine-tuned version of google-bert/bert-base-multilingual-cased on the Beijuka/Multilingual_PII_NER_dataset dataset. It achieves the following results on the evaluation set:
- Loss: 0.1237
- Precision: 0.9530
- Recall: 0.9237
- F1: 0.9381
- Accuracy: 0.9789
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 | 0.1502 | 0.8451 | 0.8843 | 0.8643 | 0.9526 |
| 0.2112 | 2.0 | 602 | 0.1347 | 0.8573 | 0.9393 | 0.8964 | 0.9604 |
| 0.2112 | 3.0 | 903 | 0.1241 | 0.8813 | 0.9398 | 0.9096 | 0.9668 |
| 0.0847 | 4.0 | 1204 | 0.1770 | 0.8589 | 0.9460 | 0.9004 | 0.9640 |
| 0.0619 | 5.0 | 1505 | 0.1295 | 0.9012 | 0.9146 | 0.9078 | 0.9673 |
| 0.0619 | 6.0 | 1806 | 0.1502 | 0.9018 | 0.9254 | 0.9134 | 0.9683 |
| 0.0394 | 7.0 | 2107 | 0.1801 | 0.8729 | 0.9506 | 0.9101 | 0.9661 |
| 0.0394 | 8.0 | 2408 | 0.1807 | 0.9119 | 0.9321 | 0.9219 | 0.9705 |
| 0.0236 | 9.0 | 2709 | 0.1660 | 0.9259 | 0.9187 | 0.9223 | 0.9719 |
| 0.0124 | 10.0 | 3010 | 0.1878 | 0.8939 | 0.9496 | 0.9209 | 0.9705 |
| 0.0124 | 11.0 | 3311 | 0.2095 | 0.8874 | 0.9486 | 0.9170 | 0.9693 |
| 0.01 | 12.0 | 3612 | 0.2370 | 0.8814 | 0.9480 | 0.9135 | 0.9664 |
Framework versions
- Transformers 4.55.4
- Pytorch 2.8.0+cu126
- Datasets 4.0.0
- Tokenizers 0.21.4