Closed iancharest closed 5 years ago
Tomas,
i did a quick fix to allow a list of images for getting captions. goes much faster to get many annotations.
returns a list of caption dicts.
I have tested it on the shared1000 and it works. not sure if other stimuli come from an annotation json that is not train2017 or val2017?
let me know what you think.
ian
oh, somehow there are still some indentation bugs in my code. will close this and fix.
Tomas,
i did a quick fix to allow a list of images for getting captions. goes much faster to get many annotations.
returns a list of caption dicts.
I have tested it on the shared1000 and it works. not sure if other stimuli come from an annotation json that is not train2017 or val2017?
let me know what you think.
ian