Migrate model card from transformers-repo
Browse filesRead announcement at https://discuss.huggingface.co/t/announcement-all-model-cards-will-be-migrated-to-hf-co-model-repos/2755
Original file history: https://github.com/huggingface/transformers/commits/master/model_cards/mrm8488/distilbert-base-multi-cased-finetuned-typo-detection/README.md
README.md
ADDED
@@ -0,0 +1,53 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
---
|
2 |
+
language: multilingual
|
3 |
+
thumbnail:
|
4 |
+
---
|
5 |
+
|
6 |
+
# DISTILBERT π + Typo Detection ββββ
|
7 |
+
|
8 |
+
[distilbert-base-multilingual-cased](https://huggingface.co/distilbert-base-multilingual-cased) fine-tuned on [GitHub Typo Corpus](https://github.com/mhagiwara/github-typo-corpus) for **typo detection** (using *NER* style)
|
9 |
+
|
10 |
+
## Details of the downstream task (Typo detection as NER)
|
11 |
+
|
12 |
+
- Dataset: [GitHub Typo Corpus](https://github.com/mhagiwara/github-typo-corpus) π for 15 languages
|
13 |
+
|
14 |
+
- [Fine-tune script on NER dataset provided by Huggingface](https://github.com/huggingface/transformers/blob/master/examples/token-classification/run_ner_old.py) ποΈββοΈ
|
15 |
+
|
16 |
+
## Metrics on test set π
|
17 |
+
|
18 |
+
| Metric | # score |
|
19 |
+
| :-------: | :-------: |
|
20 |
+
| F1 | **93.51** |
|
21 |
+
| Precision | **96.08** |
|
22 |
+
| Recall | **91.06** |
|
23 |
+
|
24 |
+
## Model in action π¨
|
25 |
+
|
26 |
+
Fast usage with **pipelines** π§ͺ
|
27 |
+
|
28 |
+
```python
|
29 |
+
from transformers import pipeline
|
30 |
+
|
31 |
+
typo_checker = pipeline(
|
32 |
+
"ner",
|
33 |
+
model="mrm8488/distilbert-base-multi-cased-finetuned-typo-detection",
|
34 |
+
tokenizer="mrm8488/distilbert-base-multi-cased-finetuned-typo-detection"
|
35 |
+
)
|
36 |
+
|
37 |
+
result = typo_checker("Adddd validation midelware")
|
38 |
+
result[1:-1]
|
39 |
+
|
40 |
+
# Output:
|
41 |
+
[{'entity': 'ok', 'score': 0.7128152847290039, 'word': 'add'},
|
42 |
+
{'entity': 'typo', 'score': 0.5388424396514893, 'word': '##dd'},
|
43 |
+
{'entity': 'ok', 'score': 0.94792640209198, 'word': 'validation'},
|
44 |
+
{'entity': 'typo', 'score': 0.5839331746101379, 'word': 'mid'},
|
45 |
+
{'entity': 'ok', 'score': 0.5195121765136719, 'word': '##el'},
|
46 |
+
{'entity': 'ok', 'score': 0.7222476601600647, 'word': '##ware'}]
|
47 |
+
```
|
48 |
+
It worksπ! We typed wrong ```Add and middleware```
|
49 |
+
|
50 |
+
|
51 |
+
> Created by [Manuel Romero/@mrm8488](https://twitter.com/mrm8488)
|
52 |
+
|
53 |
+
> Made with <span style="color: #e25555;">♥</span> in Spain
|