Add `datasets.no_passages.aio_02` config
Browse files- README.md +32 -1
- quiz-datasets.py +9 -2
README.md
CHANGED
@@ -189,10 +189,41 @@ dataset_info:
|
|
189 |
num_examples: 4288198
|
190 |
download_size: 1110830651
|
191 |
dataset_size: 3054493919
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
192 |
---
|
193 |
|
194 |
# Quiz Datasets for NLP
|
195 |
|
196 |
Question answering (QA) datasets created from Japanese quiz (trivia) questions.
|
197 |
|
198 |
-
Please refer to [cl-tohoku/quiz-datasets](https://github.com/cl-tohoku/quiz-datasets) for details, as well as the licenses of the question data.
|
|
|
189 |
num_examples: 4288198
|
190 |
download_size: 1110830651
|
191 |
dataset_size: 3054493919
|
192 |
+
- config_name: datasets.no_passages.aio_02
|
193 |
+
features:
|
194 |
+
- name: qid
|
195 |
+
dtype: string
|
196 |
+
- name: competition
|
197 |
+
dtype: string
|
198 |
+
- name: timestamp
|
199 |
+
dtype: string
|
200 |
+
- name: section
|
201 |
+
dtype: string
|
202 |
+
- name: number
|
203 |
+
dtype: string
|
204 |
+
- name: original_question
|
205 |
+
dtype: string
|
206 |
+
- name: original_answer
|
207 |
+
dtype: string
|
208 |
+
- name: original_additional_info
|
209 |
+
dtype: string
|
210 |
+
- name: question
|
211 |
+
dtype: string
|
212 |
+
- name: answers
|
213 |
+
sequence: string
|
214 |
+
splits:
|
215 |
+
- name: train
|
216 |
+
num_bytes: 9464003
|
217 |
+
num_examples: 22335
|
218 |
+
- name: validation
|
219 |
+
num_bytes: 409779
|
220 |
+
num_examples: 1000
|
221 |
+
download_size: 2267163
|
222 |
+
dataset_size: 9873782
|
223 |
---
|
224 |
|
225 |
# Quiz Datasets for NLP
|
226 |
|
227 |
Question answering (QA) datasets created from Japanese quiz (trivia) questions.
|
228 |
|
229 |
+
Please refer to [cl-tohoku/quiz-datasets](https://github.com/cl-tohoku/quiz-datasets) for details, as well as the licenses of the question data.
|
quiz-datasets.py
CHANGED
@@ -34,6 +34,10 @@ Attribution-ShareAlike 3.0 Unported ライセンスおよび GFDL ライセン
|
|
34 |
|
35 |
_URL_BASE = "https://github.com/cl-tohoku/quiz-datasets/releases/download"
|
36 |
_URLS = {
|
|
|
|
|
|
|
|
|
37 |
"datasets.jawiki-20220404-c400-small.aio_02": {
|
38 |
"train": f"{_URL_BASE}/v1.0.0/datasets.jawiki-20220404-c400-small.aio_02_train.jsonl.gz",
|
39 |
"dev": f"{_URL_BASE}/v1.0.0/datasets.jawiki-20220404-c400-small.aio_02_dev.jsonl.gz",
|
@@ -79,6 +83,10 @@ class QuizDatasets(datasets.GeneratorBasedBuilder):
|
|
79 |
"original_additional_info": datasets.Value("string"),
|
80 |
"question": datasets.Value("string"),
|
81 |
"answers": datasets.Sequence(datasets.Value("string")),
|
|
|
|
|
|
|
|
|
82 |
"passages": datasets.Sequence(
|
83 |
{
|
84 |
"passage_id": datasets.Value("int32"),
|
@@ -88,8 +96,7 @@ class QuizDatasets(datasets.GeneratorBasedBuilder):
|
|
88 |
),
|
89 |
"positive_passage_indices": datasets.Sequence(datasets.Value("int32")),
|
90 |
"negative_passage_indices": datasets.Sequence(datasets.Value("int32")),
|
91 |
-
}
|
92 |
-
)
|
93 |
elif self.config.name.startswith("passages"):
|
94 |
features = datasets.Features(
|
95 |
{
|
|
|
34 |
|
35 |
_URL_BASE = "https://github.com/cl-tohoku/quiz-datasets/releases/download"
|
36 |
_URLS = {
|
37 |
+
"datasets.no_passages.aio_02": {
|
38 |
+
"train": f"{_URL_BASE}/v1.0.0/datasets.no_passages.aio_02_train.jsonl.gz",
|
39 |
+
"dev": f"{_URL_BASE}/v1.0.0/datasets.no_passages.aio_02_dev.jsonl.gz",
|
40 |
+
},
|
41 |
"datasets.jawiki-20220404-c400-small.aio_02": {
|
42 |
"train": f"{_URL_BASE}/v1.0.0/datasets.jawiki-20220404-c400-small.aio_02_train.jsonl.gz",
|
43 |
"dev": f"{_URL_BASE}/v1.0.0/datasets.jawiki-20220404-c400-small.aio_02_dev.jsonl.gz",
|
|
|
83 |
"original_additional_info": datasets.Value("string"),
|
84 |
"question": datasets.Value("string"),
|
85 |
"answers": datasets.Sequence(datasets.Value("string")),
|
86 |
+
}
|
87 |
+
)
|
88 |
+
if not self.config.name.startswith("datasets.no_passages"):
|
89 |
+
features.update({
|
90 |
"passages": datasets.Sequence(
|
91 |
{
|
92 |
"passage_id": datasets.Value("int32"),
|
|
|
96 |
),
|
97 |
"positive_passage_indices": datasets.Sequence(datasets.Value("int32")),
|
98 |
"negative_passage_indices": datasets.Sequence(datasets.Value("int32")),
|
99 |
+
})
|
|
|
100 |
elif self.config.name.startswith("passages"):
|
101 |
features = datasets.Features(
|
102 |
{
|