Update files from the datasets library (from 1.6.0)

Release notes: https://github.com/huggingface/datasets/releases/tag/1.6.0

README.md

@@ -11,7 +11,7 @@ licenses:
 multilinguality:
 - monolingual
 size_categories:
-- n<1K
+- 10K<n<100K
 source_datasets:
 - original
 task_categories:

youtube_caption_corrections.py

@@ -15,7 +15,6 @@
 """Dataset built from <auto-generated, manually corrected> caption pairs of
 YouTube videos with labels capturing the differences between the two."""
 
-from __future__ import absolute_import, division, print_function
 
 import json