A question about the BiO-Net++ about bix-nas HOT 1 CLOSED

Hi,

We first reduced the computational complexity in BiO-Net by optimizing the feature fusion scheme and feature channel numbers at each level. Such optimizations are adopted in the design of both BiO-Net++ and BiX-Net yielding only 0.43 M trainable parameters. Our Phase1 search algorithm provides a further 8.6% MACs reduction, and our BiX-Net eventually achieves a 20% MACs reduction compared to BiO-Net++.

If you check the code, you can find that we constrained the channel number for each scale and avoid using concatenation (use weighted addition instead) when fusing the skips.

Thank you for your question.

Regards,

from bix-nas.