Update README.md
Browse files
README.md
CHANGED
@@ -1,3 +1,64 @@
|
|
1 |
-
---
|
2 |
-
|
3 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
---
|
2 |
+
language: ht
|
3 |
+
license: mit
|
4 |
+
tags:
|
5 |
+
- haitian-creole
|
6 |
+
- tokenization
|
7 |
+
- part-of-speech-tagging
|
8 |
+
- dependency-parsing
|
9 |
+
- spacy
|
10 |
+
---
|
11 |
+
|
12 |
+
# ht_core_news_sm
|
13 |
+
|
14 |
+
**Language:** Haitian Creole (ht)
|
15 |
+
**Type:** spaCy pipeline (tokenizer, POS tagger, dependency parser)
|
16 |
+
**Size:** Small (optimized for efficiency)
|
17 |
+
|
18 |
+
## Training
|
19 |
+
|
20 |
+
- Trained on ~3,300 manually corrected CoNLL-U sentences following Universal Dependencies guidelines.
|
21 |
+
- Data includes formal Haitian Creole texts (e.g., articles, religious texts, educational material).
|
22 |
+
- No pretrained word vectors were used (pure end-to-end pipeline).
|
23 |
+
- CoNLL-U Data: https://github.com/JephteyAdolphe/UD_Haitian_Creole-Adolphe
|
24 |
+
|
25 |
+
## Capabilities
|
26 |
+
|
27 |
+
- Tokenization (including contractions and informal forms common in Haitian Creole)
|
28 |
+
- Part-of-Speech (POS) tagging based on Universal POS tags
|
29 |
+
- Dependency parsing (basic syntactic parsing)
|
30 |
+
|
31 |
+
## Intended Use
|
32 |
+
|
33 |
+
- NLP research on low-resource languages
|
34 |
+
- Language technology development for Haitian Creole
|
35 |
+
- Educational or linguistic applications
|
36 |
+
|
37 |
+
## Limitations
|
38 |
+
|
39 |
+
- No named entity recognition (NER) currently included.
|
40 |
+
- Trained primarily on formal Haitian Creole; performance may vary on very informal or highly dialectal texts.
|
41 |
+
- Small dataset: best for prototyping, research, and early-stage projects.
|
42 |
+
|
43 |
+
## Example Usage
|
44 |
+
|
45 |
+
```python
|
46 |
+
import spacy
|
47 |
+
|
48 |
+
texts = [
|
49 |
+
"Si'm ka vini, m'ap pale ak li.", "M ap teste model lan (pou kounye a).",
|
50 |
+
"Map manje gato a pandan map gade televizyon lem lakay mwen.",
|
51 |
+
"M ap pale ak ou le w vini demen.", "M'ap vini, eske wap la avek lajan'm? Si ou, di'l non pou fre'w."
|
52 |
+
]
|
53 |
+
|
54 |
+
nlp = spacy.load("ht_core_news_sm")
|
55 |
+
|
56 |
+
for text in texts:
|
57 |
+
doc = nlp(text)
|
58 |
+
|
59 |
+
# Tokenization, POS tagging, Lemmatization, Dependency parsing
|
60 |
+
print("Tokens, NORM, POS (tag), Dependency:")
|
61 |
+
print(len(doc))
|
62 |
+
for token in doc:
|
63 |
+
print(f"{token.text} | {token.norm_} | {token.tag_} | {token.dep_}")
|
64 |
+
print("\n")
|