search

facebookresearch / vilbert-multi-task

Multi Task Vision and Language
MIT License
797 stars 180 forks source link

feature extraction #89

Open enaserianhanzaei opened 3 years ago

enaserianhanzaei commented 3 years ago

Thanks for the very interesting paper.

You mentioned 4 steps for extracting features, but step 2 sounds redundant, isn't it? I mean if I do only step 3, it is the same?