Why does DeformableDETR take samples with shape [batch_size x 3 x H x W] instead of [batch_size x T x 3 x H x W] #47
Description
I thought it takes multiple frames as the input.
The Vid_multi dataset certainly returns multiple images
{{ message }}