Convert dataset to Parquet

#7
README.md CHANGED
@@ -1,52 +1,39 @@
1
  ---
2
- license: cc-by-sa-4.0
3
- task_categories:
4
- - multiple-choice
5
- language:
6
- - en
7
- - zh
8
- pretty_name: LogiQA2.0
9
- data_splits:
10
- - train
11
- - validation
12
- - test
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
13
  ---
14
-
15
- # Dataset Card for Dataset Name
16
-
17
- ## Dataset Description
18
-
19
- - **Homepage:** https://github.com/csitfun/LogiQA2.0, https://github.com/csitfun/LogiEval
20
- - **Repository:** https://github.com/csitfun/LogiQA2.0, https://github.com/csitfun/LogiEval
21
- - **Paper:** https://ieeexplore.ieee.org/abstract/document/10174688
22
-
23
- ### Dataset Summary
24
-
25
- Logiqa2.0 dataset - logical reasoning in MRC and NLI tasks
26
-
27
- LogiEval: a benchmark suite for testing logical reasoning abilities of instruct-prompt large language models
28
-
29
- ### Licensing Information
30
-
31
- Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License.
32
-
33
- ### Citation Information
34
-
35
- @ARTICLE{10174688,
36
- author={Liu, Hanmeng and Liu, Jian and Cui, Leyang and Teng, Zhiyang and Duan, Nan and Zhou, Ming and Zhang, Yue},
37
- journal={IEEE/ACM Transactions on Audio, Speech, and Language Processing},
38
- title={LogiQA 2.0 — An Improved Dataset for Logical Reasoning in Natural Language Understanding},
39
- year={2023},
40
- volume={},
41
- number={},
42
- pages={1-16},
43
- doi={10.1109/TASLP.2023.3293046}}
44
-
45
- @misc{liu2023evaluating,
46
- title={Evaluating the Logical Reasoning Ability of ChatGPT and GPT-4},
47
- author={Hanmeng Liu and Ruoxi Ning and Zhiyang Teng and Jian Liu and Qiji Zhou and Yue Zhang},
48
- year={2023},
49
- eprint={2304.03439},
50
- archivePrefix={arXiv},
51
- primaryClass={cs.CL}
52
- }
 
1
  ---
2
+ dataset_info:
3
+ config_name: logiqa2
4
+ features:
5
+ - name: id
6
+ dtype: int32
7
+ - name: answer
8
+ dtype: int32
9
+ - name: text
10
+ dtype: string
11
+ - name: type
12
+ dtype: string
13
+ - name: question
14
+ dtype: string
15
+ - name: options
16
+ sequence: string
17
+ splits:
18
+ - name: train
19
+ num_bytes: 13374480
20
+ num_examples: 12567
21
+ - name: test
22
+ num_bytes: 1656534
23
+ num_examples: 1572
24
+ - name: validation
25
+ num_bytes: 1687080
26
+ num_examples: 1569
27
+ download_size: 8538515
28
+ dataset_size: 16718094
29
+ configs:
30
+ - config_name: logiqa2
31
+ data_files:
32
+ - split: train
33
+ path: logiqa2/train-*
34
+ - split: test
35
+ path: logiqa2/test-*
36
+ - split: validation
37
+ path: logiqa2/validation-*
38
+ default: true
39
  ---
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
logiqa2.py DELETED
@@ -1,230 +0,0 @@
1
- # Copyright 2020 The HuggingFace Datasets Authors and the current dataset script contributor.
2
- #
3
- # Licensed under the Apache License, Version 2.0 (the "License");
4
- # you may not use this file except in compliance with the License.
5
- # You may obtain a copy of the License at
6
- #
7
- # http://www.apache.org/licenses/LICENSE-2.0
8
- #
9
- # Unless required by applicable law or agreed to in writing, software
10
- # distributed under the License is distributed on an "AS IS" BASIS,
11
- # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
12
- # See the License for the specific language governing permissions and
13
- # limitations under the License.
14
- """LogiQA dataset."""
15
-
16
- import datasets
17
- import json
18
- import ast
19
-
20
- _CITATION = """\
21
- @ARTICLE{10174688,
22
- author={Liu, Hanmeng and Liu, Jian and Cui, Leyang and Teng, Zhiyang and Duan, Nan and Zhou, Ming and Zhang, Yue},
23
- journal={IEEE/ACM Transactions on Audio, Speech, and Language Processing},
24
- title={LogiQA 2.0 — An Improved Dataset for Logical Reasoning in Natural Language Understanding},
25
- year={2023},
26
- volume={},
27
- number={},
28
- pages={1-16},
29
- doi={10.1109/TASLP.2023.3293046}}
30
- """
31
-
32
- _DESCRIPTION = """\
33
- The dataset is an amendment and re-annotation of LogiQA in 2020, a large-scale logical reasoning reading comprehension dataset adapted from the Chinese Civil Service Examination. We increase the data size, refine the texts with manual translation by professionals, and improve the quality by removing items with distinctive cultural features like Chinese idioms. Furthermore, we conduct a fine-grained annotation on the dataset and turn it into a two-way natural language inference (NLI) task, resulting in 35k premise-hypothesis pairs with gold labels, making it the first large-scale NLI dataset for complex logical reasoning
34
- """
35
-
36
- _HOMEPAGE = "https://github.com/csitfun/LogiQA2.0/tree/main"
37
-
38
- _LICENSE = (
39
- "Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License"
40
- )
41
-
42
- _URLS = {
43
- "logiqa2": {
44
- "train": "https://raw.githubusercontent.com/csitfun/LogiQA2.0/main/logiqa/DATA/LOGIQA/train.txt",
45
- "validation": "https://raw.githubusercontent.com/csitfun/LogiQA2.0/main/logiqa/DATA/LOGIQA/dev.txt",
46
- "test": "https://raw.githubusercontent.com/csitfun/LogiQA2.0/main/logiqa/DATA/LOGIQA/test.txt",
47
- },
48
- "logiqa2_zh": {
49
- "train": "https://raw.githubusercontent.com/csitfun/LogiQA2.0/main/logiqa/DATA/LOGIQA/train_zh.txt",
50
- "validation": "https://raw.githubusercontent.com/csitfun/LogiQA2.0/main/logiqa/DATA/LOGIQA/dev_zh.txt",
51
- "test": "https://raw.githubusercontent.com/csitfun/LogiQA2.0/main/logiqa/DATA/LOGIQA/test_zh.txt",
52
- },
53
- "logiqa2_nli": {
54
- "train": "https://raw.githubusercontent.com/csitfun/LogiQA2.0/main/logiqa2nli/DATA/QA2NLI/train.txt",
55
- "validation": "https://raw.githubusercontent.com/csitfun/LogiQA2.0/main/logiqa2nli/DATA/QA2NLI/dev.txt",
56
- "test": "https://raw.githubusercontent.com/csitfun/LogiQA2.0/main/logiqa2nli/DATA/QA2NLI/test.txt",
57
- },
58
- "logieval": {
59
- "train": "https://raw.githubusercontent.com/csitfun/LogiEval/main/Data/logiqa_ood.jsonl",
60
- "test": "https://raw.githubusercontent.com/csitfun/LogiEval/main/Data/logiqa.jsonl",
61
- },
62
- }
63
-
64
-
65
- class LogiQA2(datasets.GeneratorBasedBuilder):
66
- """TODO: Short description of my dataset."""
67
-
68
- VERSION = datasets.Version("2.0.0")
69
-
70
- # This is an example of a dataset with multiple configurations.
71
- # If you don't want/need to define several sub-sets in your dataset,
72
- # just remove the BUILDER_CONFIG_CLASS and the BUILDER_CONFIGS attributes.
73
-
74
- # If you need to make complex sub-parts in the datasets with configurable options
75
- # You can create your own builder configuration class to store attribute, inheriting from datasets.BuilderConfig
76
- # BUILDER_CONFIG_CLASS = MyBuilderConfig
77
-
78
- # You will be able to load one or the other configurations in the following list with
79
- # data = datasets.load_dataset('my_dataset', 'first_domain')
80
- # data = datasets.load_dataset('my_dataset', 'second_domain')
81
- BUILDER_CONFIGS = [
82
- datasets.BuilderConfig(
83
- name="logiqa2",
84
- version=VERSION,
85
- description="The LogiQA multiple answer dataset translated in English from Chinese.",
86
- ),
87
- datasets.BuilderConfig(
88
- name="logiqa2_zh",
89
- version=VERSION,
90
- description="The original LogiQA multiple answer dataset in Chinese.",
91
- ),
92
- datasets.BuilderConfig(
93
- name="logiqa2_nli",
94
- version=VERSION,
95
- description="The NLI part of LogiQA2.0 dataset",
96
- ),
97
- datasets.BuilderConfig(
98
- name="logieval",
99
- version=VERSION,
100
- description="Instruction based MRC task",
101
- ),
102
- ]
103
- DEFAULT_CONFIG_NAME = "logiqa2"
104
-
105
- def _info(self):
106
-
107
- if self.config.name == "logiqa2_zh":
108
- features = datasets.Features(
109
- {
110
- "answer": datasets.Value("int32"),
111
- "text": datasets.Value("string"),
112
- "question": datasets.Value("string"),
113
- "options": datasets.features.Sequence(datasets.Value("string")),
114
- }
115
- )
116
- # # major_premise (maybe minor) is sometimes str, sometimes list
117
- # # can't get it to work.
118
- elif self.config.name == "logiqa2_nli":
119
- features = datasets.Features(
120
- {
121
- "label": datasets.ClassLabel(
122
- num_classes=2,
123
- names=["not entailed", "entailed"],
124
- names_file=None,
125
- id=None,
126
- ),
127
- "major_premise": datasets.features.Sequence(
128
- datasets.Value("string")
129
- ),
130
- "minor_premise": datasets.Value("string"),
131
- "conclusion": datasets.Value("string"),
132
- }
133
- )
134
- elif self.config.name in ("logiqa2_nli", "logieval"):
135
- features = datasets.Features(
136
- {"content": datasets.Value("string"), "ideal": datasets.Value("string")}
137
- )
138
- else:
139
- features = datasets.Features(
140
- {
141
- "id": datasets.Value("int32"),
142
- "answer": datasets.Value("int32"),
143
- "text": datasets.Value("string"),
144
- # "type" is a dict with arbitrary keys and values
145
- "type": datasets.Value("string"),
146
- "question": datasets.Value("string"),
147
- "options": datasets.features.Sequence(datasets.Value("string")),
148
- }
149
- )
150
- return datasets.DatasetInfo(
151
- description=_DESCRIPTION,
152
- features=features,
153
- homepage=_HOMEPAGE,
154
- license=_LICENSE,
155
- citation=_CITATION,
156
- )
157
-
158
- def _split_generators(self, dl_manager):
159
- _urls = _URLS[self.config.name]
160
- urls = {
161
- "train": _urls["train"],
162
- "test": _urls["test"],
163
- }
164
- if "validation" in _urls:
165
- urls["validation"] = _urls["validation"]
166
- data_dir = dl_manager.download_and_extract(urls)
167
- splits = [
168
- datasets.SplitGenerator(
169
- name=datasets.Split.TRAIN,
170
- # These kwargs will be passed to _generate_examples
171
- gen_kwargs={
172
- "filepath": data_dir["train"],
173
- "split": "train",
174
- },
175
- ),
176
- datasets.SplitGenerator(
177
- name=datasets.Split.TEST,
178
- # These kwargs will be passed to _generate_examples
179
- gen_kwargs={"filepath": data_dir["test"], "split": "test"},
180
- ),
181
- ]
182
- if "validation" in _urls:
183
- splits.append(
184
- datasets.SplitGenerator(
185
- name=datasets.Split.VALIDATION,
186
- # These kwargs will be passed to _generate_examples
187
- gen_kwargs={
188
- "filepath": data_dir["validation"],
189
- "split": "validation",
190
- },
191
- )
192
- )
193
- return splits
194
-
195
- def _generate_examples(self, filepath, split):
196
- with open(filepath, encoding="utf-8") as f:
197
- for key, row in enumerate(f):
198
- data = json.loads(row)
199
-
200
- if self.config.name == "logiqa2_zh":
201
- yield key, {
202
- "answer": data["answer"],
203
- "text": data["text"],
204
- "question": data["question"],
205
- "options": data["options"],
206
- }
207
- elif self.config.name == "logiqa2_nli":
208
- if isinstance(data["major_premise"], str):
209
- data["major_premise"] = [data["major_premise"]]
210
- data["minor_premise"] = data["minor_premise"].strip()
211
- yield key, {
212
- "label": data["label"],
213
- "major_premise": data["major_premise"],
214
- "minor_premise": data["minor_premise"],
215
- "conclusion": data["conclusion"],
216
- }
217
- elif self.config.name == "logieval":
218
- yield key, {
219
- "content": data["input"][1]["content"],
220
- "ideal": data["ideal"],
221
- }
222
- else:
223
- yield key, {
224
- "id": data["id"],
225
- "answer": data["answer"],
226
- "text": data["text"].strip(),
227
- "type": data["type"],
228
- "question": data["question"].strip(),
229
- "options": [x.strip() for x in data["options"]],
230
- }
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
logiqa2/test-00000-of-00001.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:af5e1c9d0213fe3ae16e6bda1aa6ed7d7ee5846841ac0bb8c31c34fe55af8af8
3
+ size 851288
logiqa2/train-00000-of-00001.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:b9b579c7ceff4ed578b43a3016bd98745b28be0d8adabe747d312b4ba6f6399f
3
+ size 6824963
logiqa2/validation-00000-of-00001.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:d05e2e119ba223e0896276c6ba79ffeb727c006791e2673f00ff8524ea9927e4
3
+ size 862264