Closed RishabhAttri closed 3 years ago
It's the groundtruth of test frames.
Why is the shape of groundtruth 16 times the frame size?
The feature extractor (I3D or C3D) takes 16 frames as an input and extract a snippet feature.
Thanks a lot!
The feature extractor (I3D or C3D) takes 16 frames as an input and extract a snippet feature.
What does gt-ucf.npy represent?