Datasets:

ArXiv:
hugo commited on
Commit
63de5fa
1 Parent(s): 94ff39a

Update README.md

Browse files
Files changed (2) hide show
  1. README.md +58 -12
  2. mmarco.py +5 -0
README.md CHANGED
@@ -1,21 +1,67 @@
1
  # Dataset Summary
2
- This repository provides the mMARCO dataset.
3
- This dataset was created by translating the MS MARCO Passage Ranking dataset into 8 different languages.
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
4
 
5
  # Dataset Structure
6
- You can load mMARCO dataset by choosing a specific language.
7
- For each language, there are four splits. The first split contains all the passages from the dataset. The last two splits are the training and development queries.
8
- An example of 'collection' from portuguese looks as follows.
 
 
 
9
  ```
10
- {
11
- "doc_id": "0",
12
- "document": "A presença da comunicação em meio às mentes científicas era igualmente importante para o sucesso do Projeto Manhattan como o intelecto científico era. A única nuvem pendurada sobre a impressionante conquista dos pesquisadores e engenheiros atômicos é o que seu sucesso realmente significava; centenas de milhares de vidas inocentes destruídas."
13
- }
 
 
 
 
 
 
 
14
  ```
15
- # Load Dataset
16
- An example to load the dataset:
 
17
  ```
18
- dataset = load_dataset('unicamp-dl/mmarco', language='portuguese')
 
 
19
  ```
20
 
21
 
 
1
  # Dataset Summary
2
+
3
+ **mMARCO** is a multilingual version of the [MS MARCO passage ranking dataset](https://microsoft.github.io/msmarco/).
4
+ For more information, checkout our papers:
5
+ * [**mMARCO: A Multilingual Version of the MS MARCO Passage Ranking Dataset**](https://arxiv.org/abs/2108.13897)
6
+ * [**A cost-benefit analysis of cross-lingual transfer methods**](https://arxiv.org/abs/2105.06813)
7
+
8
+
9
+ There are two translated versions of mMARCO.
10
+
11
+ * **v1**
12
+ In v1 version, we use MarianNMT an open-source neural machine translation framework [made available](https://huggingface.co/Helsinki-NLP) by the Language Technology Research Group at the University of Helsinki for more than a thousand language pairs for translation. This version comprises 8 languages: Chinese, French, German, Indonesian, Italian, Portuguese, Russian and Spanish. In the paper, we refer to these models as "Helsinki".
13
+
14
+ * **v2 (Recommended)**
15
+ In v2 version, we use Google Translate to translate the dataset. In this commercial translation version, besides the 8 languages from v1, we add other 5 languages: Japanese, Dutch, Vietnamese, Hindi and Arabic.
16
+
17
+
18
+ ### Supported languages
19
+
20
+ | Language name | Language code | v1 | v2 |
21
+ |---------------|---------------| ✓ | ✓ |
22
+ | English | english | ✓ | ✓ |
23
+ | Chinese | chinese | ✓ | ✓ |
24
+ | French | french | ✓ | ✓ |
25
+ | German | german | ✓ | ✓ |
26
+ | Indonesian | indonesian | ✓ | ✓ |
27
+ | Italian | italian | ✓ | ✓ |
28
+ | Portuguese | portuguese | ✓ | ✓ |
29
+ | Russian | russian | ✓ | ✓ |
30
+ | Spanish | spanish | ✓ | ✓ |
31
+ | Arabic | arabic | | ✓ |
32
+ | Dutch | dutch | | ✓ |
33
+ | Hindi | hindi | | ✓ |
34
+ | Japanese | japanese | | ✓ |
35
+ | Vietnamese | vietnamese | | ✓ |
36
+
37
 
38
  # Dataset Structure
39
+
40
+ You can load mMARCO dataset by choosing a specific language. We include training triples (query, positive and negative example), the translated collections of documents and queries.
41
+
42
+
43
+ #### Training triples
44
+
45
  ```
46
+ >>> dataset = load_dataset('mmarco', 'english')
47
+ >>> dataset['train'][1]
48
+ {'query': 'what fruit is native to australia', 'positive': 'Passiflora herbertiana. A rare passion fruit native to Australia. Fruits are green-skinned, white fleshed, with an unknown edible rating. Some sources list the fruit as edible, sweet and tasty, while others list the fruits as being bitter and inedible.assiflora herbertiana. A rare passion fruit native to Australia. Fruits are green-skinned, white fleshed, with an unknown edible rating. Some sources list the fruit as edible, sweet and tasty, while others list the fruits as being bitter and inedible.', 'negative': 'The kola nut is the fruit of the kola tree, a genus (Cola) of trees that are native to the tropical rainforests of Africa.'}
49
+ ```
50
+
51
+ #### Queries
52
+
53
+ ```
54
+ >>> dataset = load_dataset('mmarco', 'queries-spanish')
55
+ >>> dataset['train'][1]
56
+ {'id': 634306, 'text': '¿Qué significa Chattel en el historial de crédito'}
57
  ```
58
+
59
+ #### Collection
60
+
61
  ```
62
+ >>> dataset = load_dataset('mmarco', 'collection-portuguese')
63
+ >>> dataset['collection'][100]
64
+ {'id': 100, 'text': 'Antonín Dvorák (1841-1904) Antonin Dvorak era filho de açougueiro, mas ele não seguiu o negócio de seu pai. Enquanto ajudava seu pai a meio tempo, estudou música e se formou na Escola de Órgãos de Praga em 1859.'}
65
  ```
66
 
67
 
mmarco.py CHANGED
@@ -117,15 +117,20 @@ _BASE_URLS = {
117
  }
118
 
119
  LANGUAGES = [
 
120
  "chinese",
 
121
  "english",
122
  "french",
123
  "german",
 
124
  "indonesian",
125
  "italian",
 
126
  "portuguese",
127
  "russian",
128
  "spanish",
 
129
  ]
130
 
131
 
 
117
  }
118
 
119
  LANGUAGES = [
120
+ "arabic",
121
  "chinese",
122
+ "dutch",
123
  "english",
124
  "french",
125
  "german",
126
+ "hindi",
127
  "indonesian",
128
  "italian",
129
+ "japanese",
130
  "portuguese",
131
  "russian",
132
  "spanish",
133
+ "vietnamese",
134
  ]
135
 
136