fanyang3/fanyang3.github.io

When will the code be open source?

Opened this issue · 0 comments

Thank you for the amazing ideas of OIF-PCR. However, I'm still confused about jointly optimizing positional encodings and correspondences. Since there is a cross-self-attention module before inlier learning and position embedding, what is the position embedding of the self-attention module before the reference point is determined? When will the code be open source?