aim-uofa / AdelaiDet

AdelaiDet is an open source toolbox for multiple instance-level detection and recognition tasks.
https://git.io/AdelaiDet
Other
3.37k stars 646 forks source link

The detail about the FCPose #551

Open JosonChan1998 opened 2 years ago

JosonChan1998 commented 2 years ago

Thanks for your nice works! But I have some questions about the implement of fcpose.

  1. Does it use 2x downsampling and 8x downsampling feature to compute the mse loss together? https://github.com/aim-uofa/AdelaiDet/blob/c0b2092ce72442b0f40972f7c6dda8bb52c46d16/adet/modeling/fcpose/basis_module.py#L110-L116 2.Does it use 2x downsampling feature to compute the softmax loss? As I notice that you use the upsampler to get 2x downsampling features. https://github.com/aim-uofa/AdelaiDet/blob/c0b2092ce72442b0f40972f7c6dda8bb52c46d16/adet/modeling/fcpose/fcpose_head.py#L213-L216 https://github.com/aim-uofa/AdelaiDet/blob/c0b2092ce72442b0f40972f7c6dda8bb52c46d16/adet/modeling/fcpose/fcpose_head.py#L230-L233

Looking for your reply! Thanks

WeianMao commented 2 years ago

yes, supervise on larger feat can boost the performance