Datasets:
Tasks:
Text2Text Generation
Modalities:
Text
Formats:
json
Languages:
Vietnamese
Size:
10K - 100K
DOI:
License:
Update README.md
Browse files
README.md
CHANGED
@@ -12,6 +12,11 @@ size_categories:
|
|
12 |
|
13 |
# ViQP: Dataset for Vietnamese Question Paraphrasing
|
14 |
|
|
|
|
|
|
|
|
|
|
|
15 |
An example of 'viqp_train.json' looks as follows.
|
16 |
```json
|
17 |
{
|
@@ -26,6 +31,30 @@ An example of 'viqp_train.json' looks as follows.
|
|
26 |
}
|
27 |
```
|
28 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
29 |
|
30 |
# Citation Information
|
31 |
|
@@ -34,7 +63,7 @@ An example of 'viqp_train.json' looks as follows.
|
|
34 |
title = {{ViQP: Dataset for Vietnamese Question Paraphrasing }},
|
35 |
author = {Sang Quang Nguyen, Thuc Dinh Vo, Duc Phan-Anh Nguyen, Dang Thanh Tran, Kiet Van Nguyen},
|
36 |
booktitle = {x},
|
37 |
-
year = {
|
38 |
pages = {x}
|
39 |
}
|
40 |
```
|
|
|
12 |
|
13 |
# ViQP: Dataset for Vietnamese Question Paraphrasing
|
14 |
|
15 |
+
## Dataset summary
|
16 |
+
|
17 |
+
We created a new dataset consisting of pairs of questions, where each pair contains a source question and a paraphrased question. The dataset was constructed by collecting questions from various sources and then paraphrasing them using a team of human annotators.
|
18 |
+
|
19 |
+
## Dataset sample
|
20 |
An example of 'viqp_train.json' looks as follows.
|
21 |
```json
|
22 |
{
|
|
|
31 |
}
|
32 |
```
|
33 |
|
34 |
+
# Model resource
|
35 |
+
|
36 |
+
Question paraphrasing model after training with ViQP is available at [here](https://huggingface.co/SCM-LAB/vietnamese_question_paraphrasing_ViT5_base)
|
37 |
+
|
38 |
+
```cmd
|
39 |
+
pip install simplet5
|
40 |
+
```
|
41 |
+
|
42 |
+
```python
|
43 |
+
from simplet5 import SimpleT5
|
44 |
+
|
45 |
+
model = SimpleT5()
|
46 |
+
model.load_model("t5", "SCM-LAB/vietnamese_question_paraphrasing_ViT5_base", use_gpu=True)
|
47 |
+
data = "Bài thơ Lượm do ai sáng tác?"
|
48 |
+
model.predict(f"question paraphrasing: {data}", num_return_sequences= 5, num_beams= 5)
|
49 |
+
|
50 |
+
>>> ['Người sáng tác bài thơ Lượm là ai?',
|
51 |
+
'Ai là người đã sáng tác bài thơ Lượm?',
|
52 |
+
'Tác phẩm thơ Lượm do ai sáng tác?',
|
53 |
+
'Ai là tác giả của bài thơ Lượm?',
|
54 |
+
'Tác phẩm thơ Lượm được sáng tác bởi ai?']
|
55 |
+
|
56 |
+
```
|
57 |
+
|
58 |
|
59 |
# Citation Information
|
60 |
|
|
|
63 |
title = {{ViQP: Dataset for Vietnamese Question Paraphrasing }},
|
64 |
author = {Sang Quang Nguyen, Thuc Dinh Vo, Duc Phan-Anh Nguyen, Dang Thanh Tran, Kiet Van Nguyen},
|
65 |
booktitle = {x},
|
66 |
+
year = {2023},
|
67 |
pages = {x}
|
68 |
}
|
69 |
```
|