Update README.md

README.md

@@ -10,13 +10,13 @@ CLEVRER has QA pairs for each 5000 training videos.
 ```json
 {'video_filename': int, 'scene_index': str (same as filename), 'questions': list [{'question_type': , 'question_subtype': , 'question_text': , 'answer_text': , 'program'(question attributes): }]} 
 ```
-We select 'descriptive' type, 'count' subtype questions, they are object counting. 
+We select 'descriptive' type, 'count' subtype questions, they are object counting task questions. In the 'program' list, it shows how complex the question is (longer means more complex), so we filter out those longer than 9 to reduce difficulty.
 
 CLEVRER contains both positive questions and negative (non-exist) questions, so no need to construct negative samples.
 
-some questions are 'event' specific, counting moving/stationary objects when a certain event happens. i.e., 'How many objects are stationary when the yellow object enters the scene?'
+Some questions are 'event' specific, counting moving/stationary objects when a certain event happens. i.e., 'How many objects are stationary when the yellow object enters the scene?'
 
-downloading videos from: http://clevrer.csail.mit.edu/
+Downloading videos from: http://clevrer.csail.mit.edu/
 
 
 ### VisualGenome