Iamanorange

You missed the quotation mark for `TF_INCLUDE` in Makefile. Attention: It is \`PATH\`(back quote), not \'PATH\'(single quote) or \"PATH\"(double quote).

And it should be `-DNDEBUG` not `-DNDBUG`.

Paste error log.

May be related to some including problem. Double check your environment. Official tensorflow 1.12 was build against CUDA 9. If you are using a custom build of tensorflow, try the...

I'm not sure if the custom ops can be compiled on Windows. But anaconda should be OK. If you are tring to compile on Windows, at least you should change...

Try to install a newer CUDA toolkit. CUDA 9.0 works on my system.

Edit Makefile: from CFLAGS = -std=c++11 -I$(TF_INC) **-I"$(CUDA_HOME)/include"** -DGOOGLE_CUDA=1 to CFLAGS = -std=c++11 -I$(TF_INC) **-I"$(CUDA_HOME)/.."** -DGOOGLE_CUDA=1 This error may because of the wrong including directory. The Makefile set $CUDA_HOME as...

https://github.com/sampepose/flownet2-tf/issues/28#issuecomment-406941839

Install CUDA. Build the custom OPs. Run Flownet2.

@barkerje This may be caused by wrong including path. See https://github.com/sampepose/flownet2-tf/issues/45#issuecomment-411713153 to edit Makefile; or https://github.com/sampepose/flownet2-tf/issues/41#issue-337716134 to edit CUDA header files.