yl3800 / IGV

This repo contains code for Invariant Grounding for Video Question Answering
26 stars 3 forks source link

Are the intervened videos produced before the feature extraction? #1

Closed LemonQC closed 2 years ago

LemonQC commented 2 years ago

Are videos obtained in a pre-process stage.

yl3800 commented 2 years ago

Yes, the video is processed in the shape of [bs, v_len, feature_dim]