derek-thomas HF staff commited on
Commit
a538510
1 Parent(s): e5adc1c

Adding dataset generation files

Browse files
Model_finetuning_scores.png ADDED

Git LFS Details

  • SHA256: c4acc7a86af51219ffe546d6c2e8b32d53317a27106a2ee30b5f5e776feeb57e
  • Pointer size: 130 Bytes
  • Size of remote file: 34.9 kB
Squad_V1_Question_Generation.ipynb ADDED
The diff for this file is too large to render. See raw diff
 
squad_modified_for_t5_qg.py ADDED
@@ -0,0 +1,120 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ # coding=utf-8
2
+ # Copyright 2020 The TensorFlow Datasets Authors and the HuggingFace Datasets Authors.
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
+ # Lint as: python3
17
+ """SQUAD: The Stanford Question Answering Dataset."""
18
+ """Modified version for fine tuning T5 on Question Generation """
19
+
20
+ import json
21
+
22
+ import datasets
23
+
24
+ # from datasets.tasks import QuestionAnsweringExtractive
25
+
26
+ logger = datasets.logging.get_logger(__name__)
27
+
28
+ _CITATION = """\
29
+ @article{2016arXiv160605250R,
30
+ author = {{Rajpurkar}, Pranav and {Zhang}, Jian and {Lopyrev},
31
+ Konstantin and {Liang}, Percy},
32
+ title = "{SQuAD: 100,000+ Questions for Machine Comprehension of Text}",
33
+ journal = {arXiv e-prints},
34
+ year = 2016,
35
+ eid = {arXiv:1606.05250},
36
+ pages = {arXiv:1606.05250},
37
+ archivePrefix = {arXiv},
38
+ eprint = {1606.05250},
39
+ }
40
+ """
41
+
42
+ _DESCRIPTION = """\
43
+ Stanford Question Answering Dataset (SQuAD) is a reading comprehension \
44
+ dataset, consisting of questions posed by crowdworkers on a set of Wikipedia \
45
+ articles, where the answer to every question is a segment of text, or span, \
46
+ from the corresponding reading passage, or the question might be unanswerable.
47
+ """
48
+
49
+ _URL = "https://rajpurkar.github.io/SQuAD-explorer/dataset/"
50
+ _URLS = {
51
+ "train": _URL + "train-v1.1.json",
52
+ "dev": _URL + "dev-v1.1.json",
53
+ }
54
+
55
+
56
+ class SquadConfig(datasets.BuilderConfig):
57
+ """BuilderConfig for SQUAD."""
58
+
59
+ def __init__(self, **kwargs):
60
+ """BuilderConfig for SQUAD.
61
+ Args:
62
+ **kwargs: keyword arguments forwarded to super.
63
+ """
64
+ super(SquadConfig, self).__init__(**kwargs)
65
+
66
+
67
+ class Squad(datasets.GeneratorBasedBuilder):
68
+ """SQUAD: The Stanford Question Answering Dataset. Version 1.1."""
69
+
70
+ BUILDER_CONFIGS = [
71
+ SquadConfig(
72
+ name="plain_text",
73
+ version=datasets.Version("2.9.0", ""),
74
+ description="Plain text",
75
+ ),
76
+ ]
77
+
78
+ def _info(self):
79
+ return datasets.DatasetInfo(
80
+ description=_DESCRIPTION,
81
+ features=datasets.Features(
82
+ {
83
+ "context": datasets.Value("string"),
84
+ "questions": datasets.Value("string"),
85
+ }
86
+ ),
87
+ # No default supervised_keys (as we have to pass both question
88
+ # and context as input).
89
+ supervised_keys=None,
90
+ homepage="https://rajpurkar.github.io/SQuAD-explorer/",
91
+ citation=_CITATION,
92
+ task_templates=[
93
+
94
+ ],
95
+ )
96
+
97
+ def _split_generators(self, dl_manager):
98
+ downloaded_files = dl_manager.download_and_extract(_URLS)
99
+
100
+ return [
101
+ datasets.SplitGenerator(name=datasets.Split.TRAIN, gen_kwargs={"filepath": downloaded_files["train"]}),
102
+ datasets.SplitGenerator(name=datasets.Split.VALIDATION, gen_kwargs={"filepath": downloaded_files["dev"]}),
103
+ ]
104
+
105
+ def _generate_examples(self, filepath):
106
+ """This function returns the examples in the raw (text) form."""
107
+ logger.info("generating examples from = %s", filepath)
108
+ key = 0
109
+ with open(filepath, encoding="utf-8") as f:
110
+ squad = json.load(f)
111
+ for article in squad["data"]:
112
+ for paragraph in article["paragraphs"]:
113
+ source_text = f"generate questions: {paragraph['context'].strip()}"
114
+ questions = [qas['question'].strip() for qas in paragraph['qas']]
115
+ target_text = " {sep_token} ".join(questions)
116
+ target_text = f"{target_text} {{sep_token}}"
117
+ yield key, {
118
+ "context": source_text,
119
+ "questions": target_text}
120
+ key += 1