Datasets:
Xinyu Crystina ZHANG
commited on
Commit
·
691c155
1
Parent(s):
694378e
add readme
Browse files
README.md
ADDED
@@ -0,0 +1,78 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
---
|
2 |
+
language:
|
3 |
+
- Arabic
|
4 |
+
- Bengali
|
5 |
+
- English
|
6 |
+
- Finnish
|
7 |
+
- Indonesian
|
8 |
+
- Finnish
|
9 |
+
- Korean
|
10 |
+
- Russian
|
11 |
+
- Swahili
|
12 |
+
- Telugu
|
13 |
+
- Thai
|
14 |
+
- Japanese
|
15 |
+
|
16 |
+
|
17 |
+
license: "Apache License 2.0"
|
18 |
+
---
|
19 |
+
|
20 |
+
# Dataset Summary
|
21 |
+
Mr. TyDi is a multi-lingual benchmark dataset built on TyDi, covering eleven typologically diverse languages. It is designed for monolingual retrieval, specifically to evaluate ranking with learned dense representations.
|
22 |
+
|
23 |
+
# Dataset Structure
|
24 |
+
The only configuration here is the `language`,
|
25 |
+
For each language, there are three splits: `train`, `dev`, and `test`.
|
26 |
+
The negative examples from training set are sampled from the top-30 BM25 runfiles on each language.
|
27 |
+
Specifically, we combine the **training** data for all languages under the `combined` configuration.
|
28 |
+
|
29 |
+
An example of `train` set looks as follows:
|
30 |
+
```
|
31 |
+
{
|
32 |
+
'query_id': '1',
|
33 |
+
'query': 'When was quantum field theory developed?',
|
34 |
+
'positive_passages': [
|
35 |
+
{
|
36 |
+
'docid': '25267#12',
|
37 |
+
'title': 'Quantum field theory',
|
38 |
+
'text': 'Quantum field theory naturally began with the study of electromagnetic interactions, as the electromagnetic field was the only known classical field as of the 1920s.'
|
39 |
+
},
|
40 |
+
...
|
41 |
+
]
|
42 |
+
'negative_passages': [
|
43 |
+
{
|
44 |
+
'docid': '346489#8',
|
45 |
+
'title': 'Local quantum field theory',
|
46 |
+
'text': 'More recently, the approach has been further implemented to include an algebraic version of quantum field ...'
|
47 |
+
},
|
48 |
+
...
|
49 |
+
],
|
50 |
+
}
|
51 |
+
```
|
52 |
+
|
53 |
+
An example of `dev` and `test` set looks as follows. We only provide the docid of positive passages here to save the space.
|
54 |
+
Also no candidate passages are provided at this point.
|
55 |
+
Note that to perform the retrieval, it need to be used together with [castorini/mr-tydi-corpus](https://huggingface.co/datasets/castorini/mr-tydi-corpus)
|
56 |
+
```
|
57 |
+
{
|
58 |
+
'query_id': '0',
|
59 |
+
'query': 'Is Creole a pidgin of French?',
|
60 |
+
'positive_passages': [
|
61 |
+
{
|
62 |
+
'docid': '3716905#1',
|
63 |
+
},
|
64 |
+
...
|
65 |
+
]
|
66 |
+
}
|
67 |
+
```
|
68 |
+
|
69 |
+
Citation Information
|
70 |
+
```
|
71 |
+
@article{mrtydi,
|
72 |
+
title={{Mr. TyDi}: A Multi-lingual Benchmark for Dense Retrieval},
|
73 |
+
author={Xinyu Zhang and Xueguang Ma and Peng Shi and Jimmy Lin},
|
74 |
+
year={2021},
|
75 |
+
journal={arXiv:2108.08787},
|
76 |
+
}
|
77 |
+
```
|
78 |
+
|