Closed littletomatodonkey closed 3 years ago
Hi, thank you for being interested in our work. The training code is released. We use 1x1 and 3x3 conv for channel trans, and use nearest interpolation and avg pooling for size trans.
ok, thanks for your remind~
Hi, thanks for your great job! I wrote a kr version using paddle, could you please help see is there any problems? thank you!
https://github.com/littletomatodonkey/code_scipts/blob/main/knowledge_review/knowledge_review.py
I used
conv_1x1
for all the channel transform andadaptative avg pool
for the size transform.