Update README.md
Browse files
README.md
CHANGED
@@ -1,67 +1,95 @@
|
|
1 |
-
---
|
2 |
-
language:
|
3 |
-
- multilingual
|
4 |
-
- pl
|
5 |
-
- ru
|
6 |
-
- uk
|
7 |
-
- bg
|
8 |
-
- cs
|
9 |
-
- sl
|
10 |
-
datasets:
|
11 |
-
- SlavicNER
|
12 |
-
license: apache-2.0
|
13 |
-
library_name: transformers
|
14 |
-
pipeline_tag: text2text-generation
|
15 |
-
tags:
|
16 |
-
- lemmatization
|
17 |
-
widget:
|
18 |
-
- text: "pl:Polsce"
|
19 |
-
- text: "cs:Velké Británii"
|
20 |
-
- text: "bg:българите"
|
21 |
-
- text: "ru:Великобританию"
|
22 |
-
- text: "sl:evropske komisije"
|
23 |
-
- text: "uk:Європейського агентства лікарських засобів"
|
24 |
-
---
|
25 |
-
|
26 |
-
# Model description
|
27 |
-
|
28 |
-
This is a baseline model for named entity **lemmatization** trained on the single-out topic split of the
|
29 |
-
[SlavicNER corpus](https://github.com/SlavicNLP/SlavicNER).
|
30 |
-
|
31 |
-
|
32 |
-
# Resources and Technical Documentation
|
33 |
-
|
34 |
-
- Paper: [Cross-lingual Named Entity Corpus for Slavic Languages](https://arxiv.org/pdf/2404.00482), to appear in LREC-COLING 2024.
|
35 |
-
- Annotation guidelines: https://arxiv.org/pdf/2404.00482
|
36 |
-
- SlavicNER Corpus: https://github.com/SlavicNLP/SlavicNER
|
37 |
-
|
38 |
-
|
39 |
-
# Evaluation
|
40 |
-
|
41 |
-
*Will appear soon*
|
42 |
-
|
43 |
-
|
44 |
-
# Usage
|
45 |
-
|
46 |
-
You can use this model directly with a pipeline for text2text generation:
|
47 |
-
|
48 |
-
```python
|
49 |
-
from transformers import pipeline
|
50 |
-
|
51 |
-
model_name = "SlavicNLP/slavicner-lemma-cross-topic-large"
|
52 |
-
pipe = pipeline("text2text-generation", model_name)
|
53 |
-
|
54 |
-
texts = ["pl:Polsce", "cs:Velké Británii", "bg:българите", "ru:Великобританию",
|
55 |
-
"sl:evropske komisije", "uk:Європейського агентства лікарських засобів"]
|
56 |
-
|
57 |
-
outputs = pipe(texts)
|
58 |
-
|
59 |
-
ids = [o['generated_text'] for o in outputs]
|
60 |
-
print(ids)
|
61 |
-
# ['GPE-Poland', 'GPE-Great-Britain', 'GPE-Bulgaria', 'GPE-Great-Britain',
|
62 |
-
# 'ORG-European-Commission', 'ORG-EMA-European-Medicines-Agency']
|
63 |
-
```
|
64 |
-
|
65 |
-
# Citation
|
66 |
-
|
67 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
---
|
2 |
+
language:
|
3 |
+
- multilingual
|
4 |
+
- pl
|
5 |
+
- ru
|
6 |
+
- uk
|
7 |
+
- bg
|
8 |
+
- cs
|
9 |
+
- sl
|
10 |
+
datasets:
|
11 |
+
- SlavicNER
|
12 |
+
license: apache-2.0
|
13 |
+
library_name: transformers
|
14 |
+
pipeline_tag: text2text-generation
|
15 |
+
tags:
|
16 |
+
- lemmatization
|
17 |
+
widget:
|
18 |
+
- text: "pl:Polsce"
|
19 |
+
- text: "cs:Velké Británii"
|
20 |
+
- text: "bg:българите"
|
21 |
+
- text: "ru:Великобританию"
|
22 |
+
- text: "sl:evropske komisije"
|
23 |
+
- text: "uk:Європейського агентства лікарських засобів"
|
24 |
+
---
|
25 |
+
|
26 |
+
# Model description
|
27 |
+
|
28 |
+
This is a baseline model for named entity **lemmatization** trained on the single-out topic split of the
|
29 |
+
[SlavicNER corpus](https://github.com/SlavicNLP/SlavicNER).
|
30 |
+
|
31 |
+
|
32 |
+
# Resources and Technical Documentation
|
33 |
+
|
34 |
+
- Paper: [Cross-lingual Named Entity Corpus for Slavic Languages](https://arxiv.org/pdf/2404.00482), to appear in LREC-COLING 2024.
|
35 |
+
- Annotation guidelines: https://arxiv.org/pdf/2404.00482
|
36 |
+
- SlavicNER Corpus: https://github.com/SlavicNLP/SlavicNER
|
37 |
+
|
38 |
+
|
39 |
+
# Evaluation
|
40 |
+
|
41 |
+
*Will appear soon*
|
42 |
+
|
43 |
+
|
44 |
+
# Usage
|
45 |
+
|
46 |
+
You can use this model directly with a pipeline for text2text generation:
|
47 |
+
|
48 |
+
```python
|
49 |
+
from transformers import pipeline
|
50 |
+
|
51 |
+
model_name = "SlavicNLP/slavicner-lemma-cross-topic-large"
|
52 |
+
pipe = pipeline("text2text-generation", model_name)
|
53 |
+
|
54 |
+
texts = ["pl:Polsce", "cs:Velké Británii", "bg:българите", "ru:Великобританию",
|
55 |
+
"sl:evropske komisije", "uk:Європейського агентства лікарських засобів"]
|
56 |
+
|
57 |
+
outputs = pipe(texts)
|
58 |
+
|
59 |
+
ids = [o['generated_text'] for o in outputs]
|
60 |
+
print(ids)
|
61 |
+
# ['GPE-Poland', 'GPE-Great-Britain', 'GPE-Bulgaria', 'GPE-Great-Britain',
|
62 |
+
# 'ORG-European-Commission', 'ORG-EMA-European-Medicines-Agency']
|
63 |
+
```
|
64 |
+
|
65 |
+
# Citation
|
66 |
+
|
67 |
+
```latex
|
68 |
+
@inproceedings{piskorski-etal-2024-cross-lingual,
|
69 |
+
title = "Cross-lingual Named Entity Corpus for {S}lavic Languages",
|
70 |
+
author = "Piskorski, Jakub and
|
71 |
+
Marci{\'n}czuk, Micha{\l} and
|
72 |
+
Yangarber, Roman",
|
73 |
+
editor = "Calzolari, Nicoletta and
|
74 |
+
Kan, Min-Yen and
|
75 |
+
Hoste, Veronique and
|
76 |
+
Lenci, Alessandro and
|
77 |
+
Sakti, Sakriani and
|
78 |
+
Xue, Nianwen",
|
79 |
+
booktitle = "Proceedings of the 2024 Joint International Conference on Computational Linguistics, Language Resources and Evaluation (LREC-COLING 2024)",
|
80 |
+
month = may,
|
81 |
+
year = "2024",
|
82 |
+
address = "Torino, Italy",
|
83 |
+
publisher = "ELRA and ICCL",
|
84 |
+
url = "https://aclanthology.org/2024.lrec-main.369",
|
85 |
+
pages = "4143--4157",
|
86 |
+
abstract = "This paper presents a corpus manually annotated with named entities for six Slavic languages {---} Bulgarian, Czech, Polish, Slovenian, Russian,
|
87 |
+
and Ukrainian. This work is the result of a series of shared tasks, conducted in 2017{--}2023 as a part of the Workshops on Slavic Natural
|
88 |
+
Language Processing. The corpus consists of 5,017 documents on seven topics. The documents are annotated with five classes of named entities.
|
89 |
+
Each entity is described by a category, a lemma, and a unique cross-lingual identifier. We provide two train-tune dataset splits
|
90 |
+
{---} single topic out and cross topics. For each split, we set benchmarks using a transformer-based neural network architecture
|
91 |
+
with the pre-trained multilingual models {---} XLM-RoBERTa-large for named entity mention recognition and categorization,
|
92 |
+
and mT5-large for named entity lemmatization and linking.",
|
93 |
+
}
|
94 |
+
```
|
95 |
+
|