asawczyn commited on
Commit
3d5bb85
1 Parent(s): 8d641e6

Upload folder using huggingface_hub

Browse files
.gitattributes CHANGED
@@ -53,3 +53,4 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
53
  *.jpg filter=lfs diff=lfs merge=lfs -text
54
  *.jpeg filter=lfs diff=lfs merge=lfs -text
55
  *.webp filter=lfs diff=lfs merge=lfs -text
 
 
53
  *.jpg filter=lfs diff=lfs merge=lfs -text
54
  *.jpeg filter=lfs diff=lfs merge=lfs -text
55
  *.webp filter=lfs diff=lfs merge=lfs -text
56
+ ir/corpus.jsonl filter=lfs diff=lfs merge=lfs -text
README.md CHANGED
@@ -1,15 +1,85 @@
1
  ---
2
- {}
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
3
  ---
4
- # PUGG: KBQA, MRC, IR dataset for Polish
5
 
6
  ## Description
7
 
8
- This repository will contain PUGG - the dataset created in the research paper titled:
 
 
 
 
 
 
 
 
9
 
10
  **"Developing PUGG for Polish: A Modern Approach to KBQA, MRC, and IR Dataset Construction"**
11
 
12
- authored by:
13
  * Albert Sawczyn
14
  * Katsiaryna Viarenich
15
  * Konrad Wojtasik
@@ -20,6 +90,22 @@ authored by:
20
 
21
  The paper was accepted for ACL 2024 (findings).
22
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
23
  ## Links
24
 
25
  * Code:
@@ -28,6 +114,44 @@ The paper was accepted for ACL 2024 (findings).
28
  * ACL - TBA
29
  * Arxiv - TBA
30
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
31
  ## Citation
32
 
33
  TBA
 
1
  ---
2
+ annotations_creators:
3
+ - expert-generated
4
+ language_creators: []
5
+ language:
6
+ - pl
7
+ license:
8
+ - cc-by-sa-4.0
9
+ multilinguality:
10
+ - monolingual
11
+ size_categories:
12
+ - 1K<n<10K
13
+ - 10K<n<100K
14
+ source_datasets:
15
+ - original
16
+ task_categories:
17
+ - question-answering
18
+ - text-retrieval
19
+ task_ids:
20
+ - extractive-qa
21
+ - document-retrieval
22
+ pretty_name: 'PUGG: KBQA, MRC, IR dataset for Polish'
23
+ tags:
24
+ - knowledge graph
25
+ - KBQA
26
+ - wikipedia
27
+ - wikidata
28
+ configs:
29
+ - config_name: kbqa_all
30
+ data_files:
31
+ - split: train
32
+ path: kbqa/*/train.jsonl
33
+ - split: test
34
+ path: kbqa/*/test.jsonl
35
+ - config_name: kbqa_natural
36
+ data_files:
37
+ - split: train
38
+ path: kbqa/natural/train.jsonl
39
+ - split: test
40
+ path: kbqa/natural/test.jsonl
41
+ - config_name: kbqa_template-based
42
+ data_files:
43
+ - split: train
44
+ path: kbqa/template-based/train.jsonl
45
+ - split: test
46
+ path: kbqa/template-based/test.jsonl
47
+ - config_name: mrc
48
+ data_files:
49
+ - split: train
50
+ path: mrc/train.jsonl
51
+ - split: test
52
+ path: mrc/test.jsonl
53
+ - config_name: ir_corpus
54
+ data_files:
55
+ - split: test
56
+ path: ir/corpus.jsonl
57
+ - config_name: ir_queries
58
+ data_files:
59
+ - split: test
60
+ path: ir/queries.jsonl
61
+ - config_name: ir_qrels
62
+ data_files:
63
+ - split: test
64
+ path: ir/qrels/test.jsonl
65
  ---
66
+ # PUGG: KBQA, MRC, IR Dataset for Polish
67
 
68
  ## Description
69
 
70
+ This repository contains the PUGG dataset designed for three NLP tasks in the Polish language:
71
+
72
+ - KBQA (Knowledge Base Question Answering)
73
+ - MRC (Machine Reading Comprehension)
74
+ - IR (Information Retrieval)
75
+
76
+ ## Paper
77
+
78
+ For more detailed information, please refer to our research paper titled:
79
 
80
  **"Developing PUGG for Polish: A Modern Approach to KBQA, MRC, and IR Dataset Construction"**
81
 
82
+ Authored by:
83
  * Albert Sawczyn
84
  * Katsiaryna Viarenich
85
  * Konrad Wojtasik
 
90
 
91
  The paper was accepted for ACL 2024 (findings).
92
 
93
+ ## Repositories
94
+
95
+ The dataset is available in the following repositories:
96
+
97
+ * [General](https://huggingface.co/datasets/clarin-pl/PUGG_KBQA) (this repository) - contains all tasks (KBQA, MRC, IR)
98
+
99
+ For more straightforward usage, the tasks are also available in separate repositories:
100
+
101
+ * [KBQA](https://huggingface.co/datasets/clarin-pl/PUGG_KBQA)
102
+ * [MRC](https://huggingface.co/datasets/clarin-pl/PUGG_MRC)
103
+ * [IR](https://huggingface.co/datasets/clarin-pl/PUGG_IR)
104
+
105
+ The knowledge graph for KBQA task is available in the following repository:
106
+
107
+ * [Knowledge Graph](https://huggingface.co/datasets/clarin-pl/PUGG_KG)
108
+
109
  ## Links
110
 
111
  * Code:
 
114
  * ACL - TBA
115
  * Arxiv - TBA
116
 
117
+ ## Usage
118
+
119
+ ```python
120
+ from datasets import load_dataset
121
+
122
+ # loading KBQA (all)
123
+ dataset = load_dataset("clarin-pl/PUGG", "kbqa_all")
124
+ print(dataset)
125
+
126
+ # loading KBQA (natural)
127
+ dataset = load_dataset("clarin-pl/PUGG", "kbqa_natural")
128
+ print(dataset)
129
+
130
+ # loading KBQA (template-based)
131
+ dataset = load_dataset("clarin-pl/PUGG", "kbqa_template-based")
132
+ print(dataset)
133
+
134
+ # loading MRC
135
+
136
+ dataset = load_dataset("clarin-pl/PUGG", "mrc")
137
+ print(dataset)
138
+
139
+ # loading IR
140
+
141
+ ## corpus
142
+ dataset = load_dataset("clarin-pl/PUGG", "ir_corpus")
143
+ print(dataset)
144
+
145
+ ## queries
146
+ dataset = load_dataset("clarin-pl/PUGG", "ir_queries")
147
+ print(dataset)
148
+
149
+ ## qrels
150
+ dataset = load_dataset("clarin-pl/PUGG", "ir_qrels")
151
+ print(dataset)
152
+ ```
153
+
154
+
155
  ## Citation
156
 
157
  TBA
ir/corpus.jsonl ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:999f7fa6e898fd1fbc2ba1548c6d24f6c47b1be8d8678d2e6c7b87ef75315895
3
+ size 301448889
ir/qrels/test.jsonl ADDED
The diff for this file is too large to render. See raw diff
 
ir/queries.jsonl ADDED
The diff for this file is too large to render. See raw diff
 
kbqa/natural/test.jsonl ADDED
The diff for this file is too large to render. See raw diff
 
kbqa/natural/train.jsonl ADDED
The diff for this file is too large to render. See raw diff
 
kbqa/template-based/test.jsonl ADDED
The diff for this file is too large to render. See raw diff
 
kbqa/template-based/train.jsonl ADDED
The diff for this file is too large to render. See raw diff
 
mrc/test.jsonl ADDED
The diff for this file is too large to render. See raw diff
 
mrc/train.jsonl ADDED
The diff for this file is too large to render. See raw diff