system HF staff commited on
Commit
50ce175
0 Parent(s):

Update files from the datasets library (from 1.2.0)

Browse files

Release notes: https://github.com/huggingface/datasets/releases/tag/1.2.0

.gitattributes ADDED
@@ -0,0 +1,27 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ *.7z filter=lfs diff=lfs merge=lfs -text
2
+ *.arrow filter=lfs diff=lfs merge=lfs -text
3
+ *.bin filter=lfs diff=lfs merge=lfs -text
4
+ *.bin.* filter=lfs diff=lfs merge=lfs -text
5
+ *.bz2 filter=lfs diff=lfs merge=lfs -text
6
+ *.ftz filter=lfs diff=lfs merge=lfs -text
7
+ *.gz filter=lfs diff=lfs merge=lfs -text
8
+ *.h5 filter=lfs diff=lfs merge=lfs -text
9
+ *.joblib filter=lfs diff=lfs merge=lfs -text
10
+ *.lfs.* filter=lfs diff=lfs merge=lfs -text
11
+ *.model filter=lfs diff=lfs merge=lfs -text
12
+ *.msgpack filter=lfs diff=lfs merge=lfs -text
13
+ *.onnx filter=lfs diff=lfs merge=lfs -text
14
+ *.ot filter=lfs diff=lfs merge=lfs -text
15
+ *.parquet filter=lfs diff=lfs merge=lfs -text
16
+ *.pb filter=lfs diff=lfs merge=lfs -text
17
+ *.pt filter=lfs diff=lfs merge=lfs -text
18
+ *.pth filter=lfs diff=lfs merge=lfs -text
19
+ *.rar filter=lfs diff=lfs merge=lfs -text
20
+ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
21
+ *.tar.* filter=lfs diff=lfs merge=lfs -text
22
+ *.tflite filter=lfs diff=lfs merge=lfs -text
23
+ *.tgz filter=lfs diff=lfs merge=lfs -text
24
+ *.xz filter=lfs diff=lfs merge=lfs -text
25
+ *.zip filter=lfs diff=lfs merge=lfs -text
26
+ *.zstandard filter=lfs diff=lfs merge=lfs -text
27
+ *tfevents* filter=lfs diff=lfs merge=lfs -text
README.md ADDED
@@ -0,0 +1,196 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ annotations_creators:
3
+ labeled_final:
4
+ - expert-generated
5
+ labeled_swap:
6
+ - expert-generated
7
+ unlabeled_final:
8
+ - machine-generated
9
+ language_creators:
10
+ - machine-generated
11
+ languages:
12
+ - en
13
+ licenses:
14
+ - other
15
+ multilinguality:
16
+ - monolingual
17
+ size_categories:
18
+ labeled_final:
19
+ - 10K<n<100K
20
+ labeled_swap:
21
+ - 10K<n<100K
22
+ unlabeled_final:
23
+ - 100K<n<1M
24
+ source_datasets:
25
+ - original
26
+ task_categories:
27
+ - text-classification
28
+ - text-scoring
29
+ task_ids:
30
+ - semantic-similarity-classification
31
+ - semantic-similarity-scoring
32
+ - text-scoring-other-paraphrase-identification
33
+ ---
34
+
35
+ # Dataset Card Creation Guide
36
+
37
+ ## Table of Contents
38
+ - [Dataset Description](#dataset-description)
39
+ - [Dataset Summary](#dataset-summary)
40
+ - [Supported Tasks](#supported-tasks-and-leaderboards)
41
+ - [Languages](#languages)
42
+ - [Dataset Structure](#dataset-structure)
43
+ - [Data Instances](#data-instances)
44
+ - [Data Fields](#data-fields)
45
+ - [Data Splits](#data-splits)
46
+ - [Dataset Creation](#dataset-creation)
47
+ - [Curation Rationale](#curation-rationale)
48
+ - [Source Data](#source-data)
49
+ - [Annotations](#annotations)
50
+ - [Personal and Sensitive Information](#personal-and-sensitive-information)
51
+ - [Considerations for Using the Data](#considerations-for-using-the-data)
52
+ - [Social Impact of Dataset](#social-impact-of-dataset)
53
+ - [Discussion of Biases](#discussion-of-biases)
54
+ - [Other Known Limitations](#other-known-limitations)
55
+ - [Additional Information](#additional-information)
56
+ - [Dataset Curators](#dataset-curators)
57
+ - [Licensing Information](#licensing-information)
58
+ - [Citation Information](#citation-information)
59
+
60
+ ## Dataset Description
61
+
62
+ - **Homepage:** [PAWS](https://github.com/google-research-datasets/paws)
63
+ - **Repository:** [PAWS](https://github.com/google-research-datasets/paws)
64
+ - **Paper:** [PAWS: Paraphrase Adversaries from Word Scrambling](https://arxiv.org/abs/1904.01130)
65
+ - **Point of Contact:** [Yuan Zhang](zhangyua@google.com)
66
+
67
+ ### Dataset Summary
68
+
69
+ PAWS: Paraphrase Adversaries from Word Scrambling
70
+
71
+ This dataset contains 108,463 human-labeled and 656k noisily labeled pairs that feature the importance of modeling structure, context, and word order information for the problem of paraphrase identification. The dataset has two subsets, one based on Wikipedia and the other one based on the Quora Question Pairs (QQP) dataset.
72
+
73
+ For further details, see the accompanying paper: PAWS: Paraphrase Adversaries from Word Scrambling (https://arxiv.org/abs/1904.01130)
74
+
75
+ PAWS-QQP is not available due to license of QQP. It must be reconstructed by downloading the original data and then running our scripts to produce the data and attach the labels.
76
+
77
+ ### Supported Tasks and Leaderboards
78
+
79
+ [More Information Needed]
80
+
81
+ ### Languages
82
+
83
+ The text in the dataset is in English.
84
+
85
+ ## Dataset Structure
86
+
87
+ ### Data Instances
88
+
89
+ Below are two examples from the dataset:
90
+
91
+ | | Sentence 1 | Sentence 2 | Label |
92
+ | :-- | :---------------------------- | :---------------------------- | :---- |
93
+ | (1) | Although interchangeable, the body pieces on the 2 cars are not similar. | Although similar, the body parts are not interchangeable on the 2 cars. | 0 |
94
+ | (2) | Katz was born in Sweden in 1947 and moved to New York City at the age of 1. | Katz was born in 1947 in Sweden and moved to New York at the age of one. | 1 |
95
+
96
+ The first pair has different semantic meaning while the second pair is a paraphrase. State-of-the-art models trained on existing datasets have dismal performance on PAWS (<40% accuracy); however, including PAWS training data for these models improves their accuracy to 85% while maintaining performance on existing datasets such as the [Quora Question Pairs](https://data.quora.com/First-Quora-Dataset-Release-Question-Pairs).
97
+
98
+
99
+ ### Data Fields
100
+
101
+ This corpus contains pairs generated from Wikipedia pages, and can be downloaded
102
+ here:
103
+
104
+ * **PAWS-Wiki Labeled (Final)**: containing pairs that are generated from both word swapping and back translation methods. All pairs have human judgements on both paraphrasing and fluency and they are split into Train/Dev/Test sections.
105
+
106
+ * **PAWS-Wiki Labeled (Swap-only)**: containing pairs that have no back translation counterparts and therefore they are not included in the first set. Nevertheless, they are high-quality pairs with human judgements on both paraphrasing and fluency, and they can be included as an auxiliary training set.
107
+
108
+ * **PAWS-Wiki Unlabeled (Final)**: Pairs in this set have noisy labels without human judgments and can also be used as an auxiliary training set. They are generated from both word swapping and back translation methods.
109
+
110
+ All files are in the tsv format with four columns:
111
+
112
+ Column Name | Data
113
+ :------------ | :--------------------------
114
+ id | A unique id for each pair
115
+ sentence1 | The first sentence
116
+ sentence2 | The second sentence
117
+ (noisy_)label | (Noisy) label for each pair
118
+
119
+ Each label has two possible values: `0` indicates the pair has different meaning, while `1` indicates the pair is a paraphrase.
120
+
121
+
122
+ ### Data Splits
123
+
124
+ The number of examples and the proportion of paraphrase (Yes%) pairs are shown
125
+ below:
126
+
127
+ Data | Train | Dev | Test | Yes%
128
+ :------------------ | ------: | -----: | ----: | ----:
129
+ Labeled (Final) | 49,401 | 8,000 | 8,000 | 44.2%
130
+ Labeled (Swap-only) | 30,397 | -- | -- | 9.6%
131
+ Unlabeled (Final) | 645,652 | 10,000 | -- | 50.0%
132
+
133
+ ## Dataset Creation
134
+
135
+ ### Curation Rationale
136
+
137
+ Existing paraphrase identification datasets lack sentence pairs that have high lexical overlap without being paraphrases. Models trained on such data fail to distinguish pairs like *flights from New York to Florida* and *flights from Florida to New York*.
138
+
139
+ ### Source Data
140
+
141
+ #### Initial Data Collection and Normalization
142
+
143
+ Their automatic generation method is based on two ideas. The first swaps words to generate a sentence pair with the same BOW, controlled by a language model. The second uses back translation to generate paraphrases with high BOW overlap but different word order. These two strategies generate high-quality, diverse PAWS pairs, balanced evenly between paraphrases and non-paraphrases.
144
+
145
+ #### Who are the source language producers?
146
+
147
+ Mentioned above.
148
+
149
+ ### Annotations
150
+
151
+ #### Annotation process
152
+
153
+ Sentence pairs are presented to five annotators, each of which gives a binary judgment as to whether they are paraphrases or not. They chose binary judgments to make dataset have the same label schema as the QQP corpus. Overall, human agreement is high on both Quora (92.0%) and Wikipedia (94.7%) and each label only takes about 24 seconds. As such, answers are usually straight-forward to human raters.
154
+
155
+ #### Who are the annotators?
156
+
157
+ [More Information Needed]
158
+
159
+ ### Personal and Sensitive Information
160
+
161
+ [More Information Needed]
162
+
163
+ ## Considerations for Using the Data
164
+
165
+ ### Social Impact of Dataset
166
+
167
+ [More Information Needed]
168
+
169
+ ### Discussion of Biases
170
+
171
+ [More Information Needed]
172
+
173
+ ### Other Known Limitations
174
+
175
+ [More Information Needed]
176
+
177
+ ## Additional Information
178
+
179
+ ### Dataset Curators
180
+
181
+ List the people involved in collecting the dataset and their affiliation(s). If funding information is known, include it here.
182
+
183
+ ### Licensing Information
184
+
185
+ The dataset may be freely used for any purpose, although acknowledgement of Google LLC ("Google") as the data source would be appreciated. The dataset is provided "AS IS" without any warranty, express or implied. Google disclaims all liability for any damages, direct or indirect, resulting from the use of the dataset.
186
+
187
+ ### Citation Information
188
+
189
+ ```
190
+ @InProceedings{paws2019naacl,
191
+ title = {{PAWS: Paraphrase Adversaries from Word Scrambling}},
192
+ author = {Zhang, Yuan and Baldridge, Jason and He, Luheng},
193
+ booktitle = {Proc. of NAACL},
194
+ year = {2019}
195
+ }
196
+ ```
dataset_infos.json ADDED
@@ -0,0 +1 @@
 
1
+ {"labeled_final": {"description": "PAWS: Paraphrase Adversaries from Word Scrambling\n\nThis dataset contains 108,463 human-labeled and 656k noisily labeled pairs that feature\nthe importance of modeling structure, context, and word order information for the problem\nof paraphrase identification. The dataset has two subsets, one based on Wikipedia and the\nother one based on the Quora Question Pairs (QQP) dataset.\n\nFor further details, see the accompanying paper: PAWS: Paraphrase Adversaries from Word Scrambling\n(https://arxiv.org/abs/1904.01130)\n\nPAWS-QQP is not available due to license of QQP. It must be reconstructed by downloading the original\ndata and then running our scripts to produce the data and attach the labels.\n\nNOTE: There might be some missing or wrong labels in the dataset and we have replaced them with -1.\n", "citation": "@InProceedings{paws2019naacl,\n title = {{PAWS: Paraphrase Adversaries from Word Scrambling}},\n author = {Zhang, Yuan and Baldridge, Jason and He, Luheng},\n booktitle = {Proc. of NAACL},\n year = {2019}\n}\n", "homepage": "https://github.com/google-research-datasets/paws", "license": "The dataset may be freely used for any purpose, although acknowledgement of Google LLC (\"Google\") as the data source would be appreciated. The dataset is provided \"AS IS\" without any warranty, express or implied. Google disclaims all liability for any damages, direct or indirect, resulting from the use of the dataset.", "features": {"id": {"dtype": "int32", "id": null, "_type": "Value"}, "sentence1": {"dtype": "string", "id": null, "_type": "Value"}, "sentence2": {"dtype": "string", "id": null, "_type": "Value"}, "label": {"num_classes": 2, "names": ["0", "1"], "names_file": null, "id": null, "_type": "ClassLabel"}}, "post_processed": null, "supervised_keys": null, "builder_name": "paws", "config_name": "labeled_final", "version": {"version_str": "1.1.0", "description": "", "major": 1, "minor": 1, "patch": 0}, "splits": {"train": {"name": "train", "num_bytes": 12239978, "num_examples": 49401, "dataset_name": "paws"}, "test": {"name": "test", "num_bytes": 1987802, "num_examples": 8000, "dataset_name": "paws"}, "validation": {"name": "validation", "num_bytes": 1975870, "num_examples": 8000, "dataset_name": "paws"}}, "download_checksums": {"https://storage.googleapis.com/paws/english/paws_wiki_labeled_final.tar.gz": {"num_bytes": 4687157, "checksum": "1aad6cbb8a90b15563a0c154752c2b2c8e3bc5bdaa125172214d598bc76bc9fd"}}, "download_size": 4687157, "post_processing_size": null, "dataset_size": 16203650, "size_in_bytes": 20890807}, "labeled_swap": {"description": "PAWS: Paraphrase Adversaries from Word Scrambling\n\nThis dataset contains 108,463 human-labeled and 656k noisily labeled pairs that feature\nthe importance of modeling structure, context, and word order information for the problem\nof paraphrase identification. The dataset has two subsets, one based on Wikipedia and the\nother one based on the Quora Question Pairs (QQP) dataset.\n\nFor further details, see the accompanying paper: PAWS: Paraphrase Adversaries from Word Scrambling\n(https://arxiv.org/abs/1904.01130)\n\nPAWS-QQP is not available due to license of QQP. It must be reconstructed by downloading the original\ndata and then running our scripts to produce the data and attach the labels.\n\nNOTE: There might be some missing or wrong labels in the dataset and we have replaced them with -1.\n", "citation": "@InProceedings{paws2019naacl,\n title = {{PAWS: Paraphrase Adversaries from Word Scrambling}},\n author = {Zhang, Yuan and Baldridge, Jason and He, Luheng},\n booktitle = {Proc. of NAACL},\n year = {2019}\n}\n", "homepage": "https://github.com/google-research-datasets/paws", "license": "The dataset may be freely used for any purpose, although acknowledgement of Google LLC (\"Google\") as the data source would be appreciated. The dataset is provided \"AS IS\" without any warranty, express or implied. Google disclaims all liability for any damages, direct or indirect, resulting from the use of the dataset.", "features": {"id": {"dtype": "int32", "id": null, "_type": "Value"}, "sentence1": {"dtype": "string", "id": null, "_type": "Value"}, "sentence2": {"dtype": "string", "id": null, "_type": "Value"}, "label": {"num_classes": 2, "names": ["0", "1"], "names_file": null, "id": null, "_type": "ClassLabel"}}, "post_processed": null, "supervised_keys": null, "builder_name": "paws", "config_name": "labeled_swap", "version": {"version_str": "1.1.0", "description": "", "major": 1, "minor": 1, "patch": 0}, "splits": {"train": {"name": "train", "num_bytes": 7963651, "num_examples": 30397, "dataset_name": "paws"}}, "download_checksums": {"https://storage.googleapis.com/paws/english/paws_wiki_labeled_swap.tar.gz": {"num_bytes": 2257283, "checksum": "886ddb2f7f7499b2f64d260956ebbd6e14fc436eadac56cdbb966831b00d7861"}}, "download_size": 2257283, "post_processing_size": null, "dataset_size": 7963651, "size_in_bytes": 10220934}, "unlabeled_final": {"description": "PAWS: Paraphrase Adversaries from Word Scrambling\n\nThis dataset contains 108,463 human-labeled and 656k noisily labeled pairs that feature\nthe importance of modeling structure, context, and word order information for the problem\nof paraphrase identification. The dataset has two subsets, one based on Wikipedia and the\nother one based on the Quora Question Pairs (QQP) dataset.\n\nFor further details, see the accompanying paper: PAWS: Paraphrase Adversaries from Word Scrambling\n(https://arxiv.org/abs/1904.01130)\n\nPAWS-QQP is not available due to license of QQP. It must be reconstructed by downloading the original\ndata and then running our scripts to produce the data and attach the labels.\n\nNOTE: There might be some missing or wrong labels in the dataset and we have replaced them with -1.\n", "citation": "@InProceedings{paws2019naacl,\n title = {{PAWS: Paraphrase Adversaries from Word Scrambling}},\n author = {Zhang, Yuan and Baldridge, Jason and He, Luheng},\n booktitle = {Proc. of NAACL},\n year = {2019}\n}\n", "homepage": "https://github.com/google-research-datasets/paws", "license": "The dataset may be freely used for any purpose, although acknowledgement of Google LLC (\"Google\") as the data source would be appreciated. The dataset is provided \"AS IS\" without any warranty, express or implied. Google disclaims all liability for any damages, direct or indirect, resulting from the use of the dataset.", "features": {"id": {"dtype": "int32", "id": null, "_type": "Value"}, "sentence1": {"dtype": "string", "id": null, "_type": "Value"}, "sentence2": {"dtype": "string", "id": null, "_type": "Value"}, "label": {"num_classes": 2, "names": ["0", "1"], "names_file": null, "id": null, "_type": "ClassLabel"}}, "post_processed": null, "supervised_keys": null, "builder_name": "paws", "config_name": "unlabeled_final", "version": {"version_str": "1.1.0", "description": "", "major": 1, "minor": 1, "patch": 0}, "splits": {"train": {"name": "train", "num_bytes": 157806996, "num_examples": 645652, "dataset_name": "paws"}, "validation": {"name": "validation", "num_bytes": 2442173, "num_examples": 10000, "dataset_name": "paws"}}, "download_checksums": {"https://storage.googleapis.com/paws/english/paws_wiki_unlabeled_final.tar.gz": {"num_bytes": 47393331, "checksum": "c70222d390ece5218e397b3ea4b3797212ffe945fe1eae088fa6cb317c2ca3c6"}}, "download_size": 47393331, "post_processing_size": null, "dataset_size": 160249169, "size_in_bytes": 207642500}}
dummy/labeled_final/1.1.0/dummy_data.zip ADDED
@@ -0,0 +1,3 @@
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:dd79a3019efdeaff75ffd8e48e0457b307b4f65ac0d6c896b9d6410f3ac5266c
3
+ size 2306
dummy/labeled_swap/1.1.0/dummy_data.zip ADDED
@@ -0,0 +1,3 @@
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:08140f0e1728e33b840fa54bda8d7776e44496b2f5a2899aa2b91e1f753bc0a0
3
+ size 995
dummy/unlabeled_final/1.1.0/dummy_data.zip ADDED
@@ -0,0 +1,3 @@
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:2dc312b5b0350443ef4ca377748b76f05ee806752aad3550818fd8ede9f41161
3
+ size 1867
paws.py ADDED
@@ -0,0 +1,209 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
+ """PAWS, a dataset for paraphrase identification"""
16
+
17
+ from __future__ import absolute_import, division, print_function
18
+
19
+ import csv
20
+ import os
21
+
22
+ import datasets
23
+
24
+
25
+ _CITATION = """\
26
+ @InProceedings{paws2019naacl,
27
+ title = {{PAWS: Paraphrase Adversaries from Word Scrambling}},
28
+ author = {Zhang, Yuan and Baldridge, Jason and He, Luheng},
29
+ booktitle = {Proc. of NAACL},
30
+ year = {2019}
31
+ }
32
+ """
33
+
34
+ _DESCRIPTION = """\
35
+ PAWS: Paraphrase Adversaries from Word Scrambling
36
+
37
+ This dataset contains 108,463 human-labeled and 656k noisily labeled pairs that feature
38
+ the importance of modeling structure, context, and word order information for the problem
39
+ of paraphrase identification. The dataset has two subsets, one based on Wikipedia and the
40
+ other one based on the Quora Question Pairs (QQP) dataset.
41
+
42
+ For further details, see the accompanying paper: PAWS: Paraphrase Adversaries from Word Scrambling
43
+ (https://arxiv.org/abs/1904.01130)
44
+
45
+ PAWS-QQP is not available due to license of QQP. It must be reconstructed by downloading the original
46
+ data and then running our scripts to produce the data and attach the labels.
47
+
48
+ NOTE: There might be some missing or wrong labels in the dataset and we have replaced them with -1.
49
+ """
50
+
51
+ _HOMEPAGE = "https://github.com/google-research-datasets/paws"
52
+
53
+ _LICENSE = 'The dataset may be freely used for any purpose, although acknowledgement of Google LLC ("Google") as the data source would be appreciated. The dataset is provided "AS IS" without any warranty, express or implied. Google disclaims all liability for any damages, direct or indirect, resulting from the use of the dataset.'
54
+
55
+ _DATA_OPTIONS = [
56
+ "labeled_final",
57
+ "labeled_swap",
58
+ "unlabeled_final",
59
+ ]
60
+
61
+
62
+ class PAWSConfig(datasets.BuilderConfig):
63
+ """BuilderConfig for PAWS."""
64
+
65
+ def __init__(self, **kwargs):
66
+ """Constructs a PAWSConfig.
67
+ Args:
68
+ **kwargs: keyword arguments forwarded to super.
69
+ """
70
+ super(PAWSConfig, self).__init__(version=datasets.Version("1.1.0", ""), **kwargs),
71
+
72
+
73
+ class PAWS(datasets.GeneratorBasedBuilder):
74
+ """PAWS, a dataset for paraphrase identification"""
75
+
76
+ VERSION = datasets.Version("1.1.0")
77
+
78
+ BUILDER_CONFIGS = [
79
+ PAWSConfig(
80
+ name=config_name,
81
+ description=(f"This config contains samples of {config_name}."),
82
+ )
83
+ for config_name in _DATA_OPTIONS
84
+ ]
85
+
86
+ def _info(self):
87
+ features = datasets.Features(
88
+ {
89
+ "id": datasets.Value("int32"),
90
+ "sentence1": datasets.Value("string"),
91
+ "sentence2": datasets.Value("string"),
92
+ "label": datasets.features.ClassLabel(names=["0", "1"]),
93
+ }
94
+ )
95
+ return datasets.DatasetInfo(
96
+ # This is the description that will appear on the datasets page.
97
+ description=_DESCRIPTION,
98
+ # This defines the different columns of the dataset and their types
99
+ features=features, # Here we define them above because they are different between the two configurations
100
+ # If there's a common (input, target) tuple from the features,
101
+ # specify them here. They'll be used if as_supervised=True in
102
+ # builder.as_dataset.
103
+ supervised_keys=None,
104
+ # Homepage of the dataset for documentation
105
+ homepage=_HOMEPAGE,
106
+ # License for the dataset if available
107
+ license=_LICENSE,
108
+ # Citation for the dataset
109
+ citation=_CITATION,
110
+ )
111
+
112
+ def _split_generators(self, dl_manager):
113
+ """Returns SplitGenerators."""
114
+
115
+ _DATA_URL = f"https://storage.googleapis.com/paws/english/paws_wiki_{self.config.name}.tar.gz"
116
+ data_dir = dl_manager.download_and_extract(_DATA_URL)
117
+
118
+ if self.config.name == "labeled_final":
119
+ _TRAIN_FILE_NAME = os.path.join(data_dir, "final", "train.tsv")
120
+ _VAL_FILE_NAME = os.path.join(data_dir, "final", "dev.tsv")
121
+ _TEST_FILE_NAME = os.path.join(data_dir, "final", "test.tsv")
122
+ return [
123
+ datasets.SplitGenerator(
124
+ name=datasets.Split.TRAIN,
125
+ # These kwargs will be passed to _generate_examples
126
+ gen_kwargs={
127
+ "filepath": _TRAIN_FILE_NAME,
128
+ "split": datasets.Split.TRAIN,
129
+ },
130
+ ),
131
+ datasets.SplitGenerator(
132
+ name=datasets.Split.TEST,
133
+ # These kwargs will be passed to _generate_examples
134
+ gen_kwargs={
135
+ "filepath": _TEST_FILE_NAME,
136
+ "split": datasets.Split.TEST,
137
+ },
138
+ ),
139
+ datasets.SplitGenerator(
140
+ name=datasets.Split.VALIDATION,
141
+ # These kwargs will be passed to _generate_examples
142
+ gen_kwargs={
143
+ "filepath": _VAL_FILE_NAME,
144
+ "split": datasets.Split.VALIDATION,
145
+ },
146
+ ),
147
+ ]
148
+
149
+ elif self.config.name == "labeled_swap":
150
+ _TRAIN_FILE_NAME = os.path.join(data_dir, "swap", "train.tsv")
151
+ return [
152
+ datasets.SplitGenerator(
153
+ name=datasets.Split.TRAIN,
154
+ # These kwargs will be passed to _generate_examples
155
+ gen_kwargs={
156
+ "filepath": _TRAIN_FILE_NAME,
157
+ "split": datasets.Split.TRAIN,
158
+ },
159
+ ),
160
+ ]
161
+
162
+ elif self.config.name == "unlabeled_final":
163
+ _TRAIN_FILE_NAME = os.path.join(data_dir, "unlabeled", "final", "train.tsv")
164
+ _VAL_FILE_NAME = os.path.join(data_dir, "unlabeled", "final", "dev.tsv")
165
+ return [
166
+ datasets.SplitGenerator(
167
+ name=datasets.Split.TRAIN,
168
+ # These kwargs will be passed to _generate_examples
169
+ gen_kwargs={
170
+ "filepath": _TRAIN_FILE_NAME,
171
+ "split": datasets.Split.TRAIN,
172
+ },
173
+ ),
174
+ datasets.SplitGenerator(
175
+ name=datasets.Split.VALIDATION,
176
+ # These kwargs will be passed to _generate_examples
177
+ gen_kwargs={
178
+ "filepath": _VAL_FILE_NAME,
179
+ "split": datasets.Split.VALIDATION,
180
+ },
181
+ ),
182
+ ]
183
+ else:
184
+ raise NotImplementedError("{} does not exist".format(self.config.name))
185
+
186
+ def _generate_examples(self, filepath, split):
187
+ """ Yields examples. """
188
+
189
+ with open(filepath, encoding="utf-8") as f:
190
+ data = csv.DictReader(f, delimiter="\t")
191
+ for id_, row in enumerate(data):
192
+ if self.config.name != "unlabeled_final":
193
+ if row["label"] not in ["0", "1"]:
194
+ row["label"] = -1
195
+ yield id_, {
196
+ "id": row["id"],
197
+ "sentence1": row["sentence1"],
198
+ "sentence2": row["sentence2"],
199
+ "label": row["label"],
200
+ }
201
+ else:
202
+ if row["noisy_label"] not in ["0", "1"]:
203
+ row["noisy_label"] = -1
204
+ yield id_, {
205
+ "id": row["id"],
206
+ "sentence1": row["sentence1"],
207
+ "sentence2": row["sentence2"],
208
+ "label": row["noisy_label"],
209
+ }