search

SHI-Labs / Neighborhood-Attention-Transformer

Neighborhood Attention Transformer, arxiv 2022 / CVPR 2023. Dilated Neighborhood Attention Transformer, arxiv 2022
MIT License
1.04k stars 85 forks source link

How can I implement a function like joint image filtering? #9

Closed jikerWRN closed 2 years ago

jikerWRN commented 2 years ago

Hi, thanks for your great work and open-source code!

In this code, the input feature first computes weights by qk and aggregates itself in a local window. The qkv is different mappings for the same feature. I want to implement a function like joint image filtering. I have a reference feature and a target feature. I want to obtain weights by the reference feature, and then use the weights to aggregate the target feature. Namly, the qk is different mappings for the reference feature while v is the mapping for the target feature.

Please how can I achieve such functionality using existing code? Thanks!

jikerWRN commented 2 years ago

Ok, I think I know how to do it.