Datasets:
Tasks:
Question Answering
Sub-tasks:
extractive-qa
Languages:
English
Size:
100K<n<1M
ArXiv:
License:
Dattu Sharma
commited on
Commit
·
8c64faa
1
Parent(s):
4676410
Remove unncessary 'pylint disable' message in ReadMe (#3955)
Browse filesCommit from https://github.com/huggingface/datasets/commit/b7d99ad266703f114272bccb7f6ea1fe357756ba
README.md
CHANGED
|
@@ -43,7 +43,6 @@ pretty_name: SearchQA
|
|
| 43 |
|
| 44 |
### Dataset Summary
|
| 45 |
|
| 46 |
-
# pylint: disable=line-too-long
|
| 47 |
We publicly release a new large-scale dataset, called SearchQA, for machine comprehension, or question-answering. Unlike recently released datasets, such as DeepMind
|
| 48 |
CNN/DailyMail and SQuAD, the proposed SearchQA was constructed to reflect a full pipeline of general question-answering. That is, we start not from an existing article
|
| 49 |
and generate a question-answer pair, but start from an existing question-answer pair, crawled from J! Archive, and augment it with text snippets retrieved by Google.
|
|
|
|
| 43 |
|
| 44 |
### Dataset Summary
|
| 45 |
|
|
|
|
| 46 |
We publicly release a new large-scale dataset, called SearchQA, for machine comprehension, or question-answering. Unlike recently released datasets, such as DeepMind
|
| 47 |
CNN/DailyMail and SQuAD, the proposed SearchQA was constructed to reflect a full pipeline of general question-answering. That is, we start not from an existing article
|
| 48 |
and generate a question-answer pair, but start from an existing question-answer pair, crawled from J! Archive, and augment it with text snippets retrieved by Google.
|