File size: 1,449 Bytes
f565181
 
 
 
 
 
 
 
 
 
 
ae10e58
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
f565181
 
 
 
 
 
 
ae10e58
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
f565181
 
 
 
 
 
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
---
language: luo
license: apache-2.0
tags:
- pos-tagging
- token-classification
- dholuo
- afroxlmr
- masakhane
datasets:
- kencorpus

model-index:
- name: Dholuo POS Tagger (AfroXLM-R)
  results:
  - task:
      type: token-classification
      name: POS Tagging
    dataset:
      name: KenCorpus (Dholuo)
      type: kencorpus
      split: test
    metrics:
    - name: Accuracy
      type: accuracy
      value: 0.7167
    - name: Macro F1
      type: f1
      value: 0.5338
    - name: Weighted F1
      type: f1
      value: 0.7248
---

# Dholuo POS Tagger (AfroXLM-R)

This model is a fine-tuned version of `masakhane/luo-pos-tagger-afroxlmr`
trained on KenCorpus POS-tagged Dholuo data.

## Evaluation Results

The model was evaluated on the KenCorpus Dholuo POS test set.

| Metric        | Value |
|---------------|-------|
| Accuracy      | 0.7167 |
| Macro F1      | 0.5338 |
| Weighted F1   | 0.7248 |

### Per-tag Performance (Test Set)

- Strong performance on **NOUN (F1 ≈ 0.86)** and **VERB (F1 ≈ 0.84)**
- Moderate performance on function words (ADP, PRON, DET)
- Rare tags (INTJ, PART, PUNCT) have near-zero scores due to very low support

## Usage

```python
from transformers import AutoTokenizer, AutoModelForTokenClassification

model = AutoModelForTokenClassification.from_pretrained(
    "Omballa/dholuo-pos-afroxlmr-finetuned"
)
tokenizer = AutoTokenizer.from_pretrained(
    "Omballa/dholuo-pos-afroxlmr-finetuned"
)