Adapting Attention Conv & Stem for a stride > 1

leaderj1001 / Stand-Alone-Self-Attention

Implementing Stand-Alone Self-Attention in Vision Models using Pytorch

MIT License

456 stars 83 forks source link

Adapting Attention Conv & Stem for a stride > 1 #17

Open yassouali opened 4 years ago

yassouali commented 4 years ago

The current version only works for attconv & attstem with stride == 1. The change is simple, we change the self.query_conv to have the desired stride and update the height and width after folding.