ljvmiranda921
commited on
Update README.md
Browse files
README.md
CHANGED
@@ -5,8 +5,25 @@ tags:
|
|
5 |
language:
|
6 |
- tl
|
7 |
license: mit
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
8 |
---
|
9 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
10 |
|
11 |
| Feature | Description |
|
12 |
| --- | --- |
|
@@ -33,4 +50,23 @@ calamanCy: Tagalog NLP pipelines in spaCy
|
|
33 |
| **`parser`** | `ROOT`, `acl`, `acl:relcl`, `advcl`, `advmod`, `amod`, `appos`, `case`, `cc`, `ccomp`, `compound`, `compound:redup`, `conj`, `dep`, `det`, `discourse`, `dislocated`, `fixed`, `flat`, `goeswith`, `list`, `mark`, `nmod`, `nmod:poss`, `nsubj`, `nummod`, `obj`, `obj:agent`, `obl`, `orphan`, `parataxis`, `punct`, `vocative`, `xcomp` |
|
34 |
| **`ner`** | `LOC`, `ORG`, `PER` |
|
35 |
|
36 |
-
</details>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
5 |
language:
|
6 |
- tl
|
7 |
license: mit
|
8 |
+
datasets:
|
9 |
+
- UD-Filipino/UD_Tagalog-NewsCrawl
|
10 |
+
- ljvmiranda921/tlunified-ner
|
11 |
+
- SEACrowd/tlunified_ner
|
12 |
+
base_model:
|
13 |
+
- microsoft/mdeberta-v3-base
|
14 |
+
pipeline_tag: token-classification
|
15 |
+
library_name: spacy
|
16 |
---
|
17 |
+
|
18 |
+
<img src="https://raw.githubusercontent.com/ljvmiranda921/calamanCy/refs/heads/master/logo.png" width="130" height="130" align="right" />
|
19 |
+
|
20 |
+
# calamanCy: Tagalog NLP pipelines in spaCy
|
21 |
+
|
22 |
+
This is the latest **transformer-based pipeline** for [calamanCy](https://arxiv.org/abs/2311.07171).
|
23 |
+
Compared to the 0.1.0 version, this pipeline is trained on a larger treebank ([UD-NewsCrawl](https://huggingface.co/datasets/UD-Filipino/UD_Tagalog-NewsCrawl)), with large improvements in dependency parsing, morphological annotation, and POS tagging.
|
24 |
+
This pipeline also implements a neural edit-tree lemmatizer, allowing better lemmatization than the previous model.
|
25 |
+
The training code can be found [in GitHub](https://github.com/ljvmiranda921/calamanCy/tree/master/models/v0.1.0).
|
26 |
+
|
27 |
|
28 |
| Feature | Description |
|
29 |
| --- | --- |
|
|
|
50 |
| **`parser`** | `ROOT`, `acl`, `acl:relcl`, `advcl`, `advmod`, `amod`, `appos`, `case`, `cc`, `ccomp`, `compound`, `compound:redup`, `conj`, `dep`, `det`, `discourse`, `dislocated`, `fixed`, `flat`, `goeswith`, `list`, `mark`, `nmod`, `nmod:poss`, `nsubj`, `nummod`, `obj`, `obj:agent`, `obl`, `orphan`, `parataxis`, `punct`, `vocative`, `xcomp` |
|
51 |
| **`ner`** | `LOC`, `ORG`, `PER` |
|
52 |
|
53 |
+
</details>
|
54 |
+
|
55 |
+
### Citation
|
56 |
+
|
57 |
+
If you're using this model, please cite:
|
58 |
+
|
59 |
+
```
|
60 |
+
@inproceedings{miranda-2023-calamancy,
|
61 |
+
title = "calaman{C}y: A {T}agalog Natural Language Processing Toolkit",
|
62 |
+
author = "Miranda, Lester James",
|
63 |
+
booktitle = "Proceedings of the 3rd Workshop for Natural Language Processing Open Source Software (NLP-OSS 2023)",
|
64 |
+
month = dec,
|
65 |
+
year = "2023",
|
66 |
+
address = "Singapore",
|
67 |
+
publisher = "Association for Computational Linguistics",
|
68 |
+
url = "https://aclanthology.org/2023.nlposs-1.1/",
|
69 |
+
doi = "10.18653/v1/2023.nlposs-1.1",
|
70 |
+
pages = "1--7",
|
71 |
+
}
|
72 |
+
```
|