---
library_name: transformers
license: afl-3.0
base_model: masakhane/afroxlmr-large-ner-masakhaner-1.0_2.0
tags:
- named-entity-recognition
- lumasaba
- african-language
- pii-detection
- token-classification
- generated_from_trainer
datasets:
- Beijuka/Multilingual_PII_NER_dataset
metrics:
- precision
- recall
- f1
- accuracy
model-index:
- name: multilingual-masakhane/afroxlmr-large-ner-masakhaner-1.0_2.0-lumasaba-ner-v1
  results:
  - task:
      name: Token Classification
      type: token-classification
    dataset:
      name: Beijuka/Multilingual_PII_NER_dataset
      type: Beijuka/Multilingual_PII_NER_dataset
      args: 'split: train+validation+test'
    metrics:
    - name: Precision
      type: precision
      value: 0.9702892885066459
    - name: Recall
      type: recall
      value: 0.9487767584097859
    - name: F1
      type: f1
      value: 0.9594124468496328
    - name: Accuracy
      type: accuracy
      value: 0.9525409491810164
---

<!-- This model card has been generated automatically according to the information the Trainer had access to. You
should probably proofread and complete it, then remove this comment. -->

# multilingual-masakhane/afroxlmr-large-ner-masakhaner-1.0_2.0-lumasaba-ner-v1

This model is a fine-tuned version of [masakhane/afroxlmr-large-ner-masakhaner-1.0_2.0](https://huggingface.co/masakhane/afroxlmr-large-ner-masakhaner-1.0_2.0) on the Beijuka/Multilingual_PII_NER_dataset dataset.
It achieves the following results on the evaluation set:
- Loss: 0.3834
- Precision: 0.9703
- Recall: 0.9488
- F1: 0.9594
- Accuracy: 0.9525

## 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: 4
- eval_batch_size: 4
- 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 |
|:-------------:|:-----:|:----:|:---------------:|:---------:|:------:|:------:|:--------:|
| 1.1185        | 1.0   | 796  | 0.5047          | 0.8681    | 0.8810 | 0.8745 | 0.8735   |
| 0.3868        | 2.0   | 1592 | 0.4627          | 0.9012    | 0.9146 | 0.9079 | 0.9108   |
| 0.2335        | 3.0   | 2388 | 0.4419          | 0.9115    | 0.9272 | 0.9193 | 0.9198   |
| 0.1462        | 4.0   | 3184 | 0.3402          | 0.9499    | 0.9507 | 0.9503 | 0.9520   |
| 0.1072        | 5.0   | 3980 | 0.2399          | 0.9560    | 0.9538 | 0.9549 | 0.9563   |
| 0.0916        | 6.0   | 4776 | 0.3072          | 0.9548    | 0.9593 | 0.9570 | 0.9588   |
| 0.0432        | 7.0   | 5572 | 0.3124          | 0.9573    | 0.9663 | 0.9618 | 0.9605   |
| 0.0383        | 8.0   | 6368 | 0.3386          | 0.9669    | 0.9608 | 0.9639 | 0.9575   |
| 0.0502        | 9.0   | 7164 | 0.4429          | 0.9644    | 0.9554 | 0.9599 | 0.9550   |
| 0.0349        | 10.0  | 7960 | 0.4191          | 0.9605    | 0.9522 | 0.9564 | 0.9481   |
| 0.039         | 11.0  | 8756 | 0.4815          | 0.9558    | 0.9648 | 0.9602 | 0.9537   |


### Framework versions

- Transformers 4.55.4
- Pytorch 2.8.0+cu126
- Datasets 4.0.0
- Tokenizers 0.21.4