Update README.md

README.md

@@ -19,9 +19,9 @@ size_categories:
 [DiDeMo](https://openaccess.thecvf.com/content_iccv_2017/html/Hendricks_Localizing_Moments_in_ICCV_2017_paper.html) contains 10K long-form videos from Flickr. For each video, ~4 short sentences are annotated in temporal order. We follow the existing works to concatenate those short sentences and evaluate ‘paragraph-to-video’ retrieval on this benchmark.
 
 We adopt the official split:  
-- Train:  8,395 videos, 8,395 captions (concatenate from 33,005 captions)  
-- Val: 1,065 videos, 1,065 captions (concatenate from 4,290 captions) (We don't have the collection yet.)  
-- Test: 1,004 videos, 1,004 captions (concatenate from 4,021 captions)  
+- Train:  8,395 videos, 8,395 captions (concatenate from 33,005 short captions)  
+- Val: 1,065 videos, 1,065 captions (concatenate from 4,290 short captions) (We don't have the collection yet.)  
+- Test: 1,004 videos, 1,004 captions (concatenate from 4,021 short captions)  
 
 ---