Datasets:
Tasks:
Question Answering
Modalities:
Text
Formats:
parquet
Sub-tasks:
multiple-choice-qa
Languages:
Chinese
Size:
10K - 100K
ArXiv:
License:
Commit
•
7aa834e
1
Parent(s):
48d5f78
Delete loading script
Browse files
c3.py
DELETED
@@ -1,149 +0,0 @@
|
|
1 |
-
# coding=utf-8
|
2 |
-
# Copyright 2020 The HuggingFace Datasets Authors and the current dataset script contributor.
|
3 |
-
#
|
4 |
-
# Licensed under the Apache License, Version 2.0 (the "License");
|
5 |
-
# you may not use this file except in compliance with the License.
|
6 |
-
# You may obtain a copy of the License at
|
7 |
-
#
|
8 |
-
# http://www.apache.org/licenses/LICENSE-2.0
|
9 |
-
#
|
10 |
-
# Unless required by applicable law or agreed to in writing, software
|
11 |
-
# distributed under the License is distributed on an "AS IS" BASIS,
|
12 |
-
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
|
13 |
-
# See the License for the specific language governing permissions and
|
14 |
-
# limitations under the License.
|
15 |
-
"""C3 Parallel Corpora"""
|
16 |
-
|
17 |
-
|
18 |
-
import json
|
19 |
-
|
20 |
-
import datasets
|
21 |
-
|
22 |
-
|
23 |
-
_CITATION = """\
|
24 |
-
@article{sun2019investigating,
|
25 |
-
title={Investigating Prior Knowledge for Challenging Chinese Machine Reading Comprehension},
|
26 |
-
author={Sun, Kai and Yu, Dian and Yu, Dong and Cardie, Claire},
|
27 |
-
journal={Transactions of the Association for Computational Linguistics},
|
28 |
-
year={2020},
|
29 |
-
url={https://arxiv.org/abs/1904.09679v3}
|
30 |
-
}
|
31 |
-
"""
|
32 |
-
|
33 |
-
_DESCRIPTION = """\
|
34 |
-
Machine reading comprehension tasks require a machine reader to answer questions relevant to the given document. In this paper, we present the first free-form multiple-Choice Chinese machine reading Comprehension dataset (C^3), containing 13,369 documents (dialogues or more formally written mixed-genre texts) and their associated 19,577 multiple-choice free-form questions collected from Chinese-as-a-second-language examinations.
|
35 |
-
We present a comprehensive analysis of the prior knowledge (i.e., linguistic, domain-specific, and general world knowledge) needed for these real-world problems. We implement rule-based and popular neural methods and find that there is still a significant performance gap between the best performing model (68.5%) and human readers (96.0%), especially on problems that require prior knowledge. We further study the effects of distractor plausibility and data augmentation based on translated relevant datasets for English on model performance. We expect C^3 to present great challenges to existing systems as answering 86.8% of questions requires both knowledge within and beyond the accompanying document, and we hope that C^3 can serve as a platform to study how to leverage various kinds of prior knowledge to better understand a given written or orally oriented text.
|
36 |
-
"""
|
37 |
-
|
38 |
-
_URL = "https://raw.githubusercontent.com/nlpdata/c3/master/data/"
|
39 |
-
|
40 |
-
|
41 |
-
class C3Config(datasets.BuilderConfig):
|
42 |
-
"""BuilderConfig for NewDataset"""
|
43 |
-
|
44 |
-
def __init__(self, type_, **kwargs):
|
45 |
-
"""
|
46 |
-
|
47 |
-
Args:
|
48 |
-
pair: the language pair to consider
|
49 |
-
zip_file: The location of zip file containing original data
|
50 |
-
**kwargs: keyword arguments forwarded to super.
|
51 |
-
"""
|
52 |
-
self.type_ = type_
|
53 |
-
super().__init__(**kwargs)
|
54 |
-
|
55 |
-
|
56 |
-
class C3(datasets.GeneratorBasedBuilder):
|
57 |
-
"""C3 is the first free-form multiple-Choice Chinese machine reading Comprehension dataset, containing 13,369 documents (dialogues or more formally written mixed-genre texts) and their associated 19,577 multiple-choice free-form questions collected from Chinese-as-a-second language examinations."""
|
58 |
-
|
59 |
-
VERSION = datasets.Version("1.0.0")
|
60 |
-
|
61 |
-
# This is an example of a dataset with multiple configurations.
|
62 |
-
# If you don't want/need to define several sub-sets in your dataset,
|
63 |
-
# just remove the BUILDER_CONFIG_CLASS and the BUILDER_CONFIGS attributes.
|
64 |
-
BUILDER_CONFIG_CLASS = C3Config
|
65 |
-
BUILDER_CONFIGS = [
|
66 |
-
C3Config(
|
67 |
-
name="mixed",
|
68 |
-
description="Mixed genre questions",
|
69 |
-
version=datasets.Version("1.0.0"),
|
70 |
-
type_="mixed",
|
71 |
-
),
|
72 |
-
C3Config(
|
73 |
-
name="dialog",
|
74 |
-
description="Dialog questions",
|
75 |
-
version=datasets.Version("1.0.0"),
|
76 |
-
type_="dialog",
|
77 |
-
),
|
78 |
-
]
|
79 |
-
|
80 |
-
def _info(self):
|
81 |
-
return datasets.DatasetInfo(
|
82 |
-
# This is the description that will appear on the datasets page.
|
83 |
-
description=_DESCRIPTION,
|
84 |
-
# datasets.features.FeatureConnectors
|
85 |
-
features=datasets.Features(
|
86 |
-
{
|
87 |
-
"documents": datasets.Sequence(datasets.Value("string")),
|
88 |
-
"document_id": datasets.Value("string"),
|
89 |
-
"questions": datasets.Sequence(
|
90 |
-
{
|
91 |
-
"question": datasets.Value("string"),
|
92 |
-
"answer": datasets.Value("string"),
|
93 |
-
"choice": datasets.Sequence(datasets.Value("string")),
|
94 |
-
}
|
95 |
-
),
|
96 |
-
}
|
97 |
-
),
|
98 |
-
# If there's a common (input, target) tuple from the features,
|
99 |
-
# specify them here. They'll be used if as_supervised=True in
|
100 |
-
# builder.as_dataset.
|
101 |
-
supervised_keys=None,
|
102 |
-
# Homepage of the dataset for documentation
|
103 |
-
homepage="https://github.com/nlpdata/c3",
|
104 |
-
citation=_CITATION,
|
105 |
-
)
|
106 |
-
|
107 |
-
def _split_generators(self, dl_manager):
|
108 |
-
# m or d
|
109 |
-
T = self.config.type_[0]
|
110 |
-
files = [_URL + f"c3-{T}-{split}.json" for split in ["train", "test", "dev"]]
|
111 |
-
dl_dir = dl_manager.download_and_extract(files)
|
112 |
-
|
113 |
-
return [
|
114 |
-
datasets.SplitGenerator(
|
115 |
-
name=datasets.Split.TRAIN,
|
116 |
-
# These kwargs will be passed to _generate_examples
|
117 |
-
gen_kwargs={
|
118 |
-
"filename": dl_dir[0],
|
119 |
-
"split": "train",
|
120 |
-
},
|
121 |
-
),
|
122 |
-
datasets.SplitGenerator(
|
123 |
-
name=datasets.Split.TEST,
|
124 |
-
# These kwargs will be passed to _generate_examples
|
125 |
-
gen_kwargs={
|
126 |
-
"filename": dl_dir[1],
|
127 |
-
"split": "test",
|
128 |
-
},
|
129 |
-
),
|
130 |
-
datasets.SplitGenerator(
|
131 |
-
name=datasets.Split.VALIDATION,
|
132 |
-
# These kwargs will be passed to _generate_examples
|
133 |
-
gen_kwargs={
|
134 |
-
"filename": dl_dir[2],
|
135 |
-
"split": "dev",
|
136 |
-
},
|
137 |
-
),
|
138 |
-
]
|
139 |
-
|
140 |
-
def _generate_examples(self, filename, split):
|
141 |
-
"""Yields examples."""
|
142 |
-
with open(filename, "r", encoding="utf-8") as sf:
|
143 |
-
data = json.load(sf)
|
144 |
-
for id_, (documents, questions, document_id) in enumerate(data):
|
145 |
-
yield id_, {
|
146 |
-
"documents": documents,
|
147 |
-
"questions": questions,
|
148 |
-
"document_id": document_id,
|
149 |
-
}
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|