Datasets:
Tasks:
Text Classification
Modalities:
Text
Sub-tasks:
natural-language-inference
Languages:
Croatian
Size:
1K - 10K
ArXiv:
License:
SuzanaB
commited on
Commit
·
d6f6dec
1
Parent(s):
d71a4a8
Add files
Browse files- copa_hr.py +100 -0
- test.jsonl +3 -0
- train.jsonl +3 -0
- val.jsonl +3 -0
copa_hr.py
ADDED
@@ -0,0 +1,100 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
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 |
+
|
16 |
+
|
17 |
+
import json
|
18 |
+
|
19 |
+
import datasets
|
20 |
+
|
21 |
+
|
22 |
+
_CITATION = ''
|
23 |
+
_DESCRIPTION = """The COPA-HR dataset (Choice of plausible alternatives in Croatian) is a translation
|
24 |
+
of the English COPA dataset (https://people.ict.usc.edu/~gordon/copa.html) by following the
|
25 |
+
XCOPA dataset translation methodology (https://arxiv.org/abs/2005.00333). The dataset consists of 1000 premises
|
26 |
+
(My body cast a shadow over the grass), each given a question (What is the cause?), and two choices
|
27 |
+
(The sun was rising; The grass was cut), with a label encoding which of the choices is more plausible
|
28 |
+
given the annotator or translator (The sun was rising).
|
29 |
+
|
30 |
+
The dataset is split into 400 training samples, 100 validation samples, and 500 test samples. It includes the
|
31 |
+
following features: 'premise', 'choice1', 'choice2', 'label', 'question', 'changed' (boolean).
|
32 |
+
"""
|
33 |
+
_HOMEPAGE = 'https://www.clarin.si/repository/xmlui/handle/11356/1404'
|
34 |
+
_LICENSE = ''
|
35 |
+
|
36 |
+
_TRAINING_FILE = 'train.jsonl'
|
37 |
+
_DEV_FILE = 'val.jsonl'
|
38 |
+
_TEST_FILE = 'test.jsonl'
|
39 |
+
|
40 |
+
|
41 |
+
class CopaHr(datasets.GeneratorBasedBuilder):
|
42 |
+
VERSION = datasets.Version('1.1.0')
|
43 |
+
|
44 |
+
BUILDER_CONFIGS = [
|
45 |
+
datasets.BuilderConfig(
|
46 |
+
name='copa_hr',
|
47 |
+
version=VERSION,
|
48 |
+
data_files=[_TRAINING_FILE, _DEV_FILE, _TEST_FILE],
|
49 |
+
description=''
|
50 |
+
)
|
51 |
+
]
|
52 |
+
|
53 |
+
def _info(self):
|
54 |
+
features = datasets.Features(
|
55 |
+
{
|
56 |
+
'premise': datasets.Value('string'),
|
57 |
+
'choice1': datasets.Value('string'),
|
58 |
+
'choice2': datasets.Value('string'),
|
59 |
+
'label': datasets.features.ClassLabel(names=['0', '1']),
|
60 |
+
'question': datasets.Value('string'),
|
61 |
+
'idx': datasets.Value('int64'),
|
62 |
+
'changed': datasets.Value('bool')
|
63 |
+
}
|
64 |
+
)
|
65 |
+
|
66 |
+
return datasets.DatasetInfo(
|
67 |
+
description=_DESCRIPTION,
|
68 |
+
features=features,
|
69 |
+
supervised_keys=None,
|
70 |
+
homepage=_HOMEPAGE,
|
71 |
+
license=_LICENSE,
|
72 |
+
citation=_CITATION,
|
73 |
+
)
|
74 |
+
|
75 |
+
def _split_generators(self, dl_manager):
|
76 |
+
return [
|
77 |
+
datasets.SplitGenerator(
|
78 |
+
name=datasets.Split.TRAIN, gen_kwargs={'filepath': _TRAINING_FILE, 'split': 'train'}
|
79 |
+
),
|
80 |
+
datasets.SplitGenerator(
|
81 |
+
name=datasets.Split.VALIDATION, gen_kwargs={'filepath': _DEV_FILE, 'split': 'dev'}
|
82 |
+
),
|
83 |
+
datasets.SplitGenerator(
|
84 |
+
name=datasets.Split.TEST, gen_kwargs={'filepath': _TEST_FILE, 'split': 'test'}
|
85 |
+
),
|
86 |
+
]
|
87 |
+
|
88 |
+
def _generate_examples(self, filepath, split):
|
89 |
+
with open(filepath, encoding='utf-8') as f:
|
90 |
+
for i, line in enumerate(f):
|
91 |
+
data = json.loads(line)
|
92 |
+
yield i, {
|
93 |
+
'premise': data['premise'],
|
94 |
+
'choice1': data['choice1'],
|
95 |
+
'choice2': data['choice2'],
|
96 |
+
'question': data['question'],
|
97 |
+
'label': str(data['label']),
|
98 |
+
'idx': data['idx'],
|
99 |
+
'changed': data['changed']
|
100 |
+
}
|
test.jsonl
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:809967a09c31862d51efde4ae6ae0ed7d36d0e5179e260a786fc267569326875
|
3 |
+
size 98496
|
train.jsonl
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:a96f4bbcfb3d7dac287b31231b61a73f080b9cc9d68496cb6aaaa927f735230c
|
3 |
+
size 80381
|
val.jsonl
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:20b4348af5cfaf4be074a57e7b9e5c5d6fdb23eb0530681c7cd0295341a163a8
|
3 |
+
size 19982
|