deep learning - Need cpu-mode implementation of Deformable Convnets using MxNET -
i'm using deformable convnets object detection , using prebuilt mxnet binaries. have demos , training working on gpu-mode in both ubuntu , windows.
the deformable related operators part of official mxnet code. however, implemented gpu-mode only.
i need demo (not training) work on cpu-only-mode (using mx.cpu() api) getting error implemented gpu.
the file this one.
can give pointers or suggestions on how implement cpu mode? possible copy deformable gpu-mode code, make few changes, , use cpu mode? if so, please specific steps.
github issues have been opened in respective repos well.
Comments
Post a Comment