Can I run the code with only CPU?

[YantianZha]

Hello,

When I tried to do the sh make.sh, I got the error: g++: error: Hypothesis.o: No such file or directory hough_voting_layer make.sh: 67: make.sh: nvcc: not found g++: error: roi_pooling_op.cu.o: No such file or directory roi_pooling_layer make.sh: 77: make.sh: nvcc: not found g++: error: triplet_loss_op.cu.o: No such file or directory build triplet loss make.sh: 87: make.sh: nvcc: not found g++: error: lifted_structured_loss_op.cu.o: No such file or directory build lifted structured loss make.sh: 97: make.sh: nvcc: not found g++: error: computing_flow_op.cu.o: No such file or directory build computing flow layer make.sh: 107: make.sh: nvcc: not found g++: error: backprojecting_op.cu.o: No such file or directory build backprojecting layer make.sh: 117: make.sh: nvcc: not found g++: error: projecting_op.cu.o: No such file or directory build projecting layer make.sh: 127: make.sh: nvcc: not found g++: error: computing_label_op.cu.o: No such file or directory build computing label layer

Those errors indicate that I need to have CUDA installed. That said, my machine does not have a dedicated GPU. Because tensorflow has a CPU-only version, I was wondering if I can also run this code without CUDA.

Thanks!

[niwanli]

I meet the same problem with you, have you found a solution? Thanks!

[Luedeke]

Hello @niwanli @YantianZha ,

Got the same Errors ->

1. installed Cuda 9.0 (Tensorflow requires 9.0!) https://developer.nvidia.com/cuda-90-download-archive?target_os=Linux&target_arch=x86_64&target_distro=Ubuntu&target_version=1604&target_type=deblocal 1.1 Add with nano ~/.bashrc the following line: export PATH=/usr/local/cuda-9.0/bin${PATH:+:${PATH}}

2. then install cuddn https://developer.nvidia.com/rdp/cudnn-download2.3.2. Installing from a Debian File Navigate to your directory containing cuDNN Debian file.

   Install the runtime library, for example: sudo dpkg -i libcudnn7_7.0.3.11-1+cuda9.0_amd64.deb

   Install the developer library, for example: sudo dpkg -i libcudnn7-dev_7.0.3.11-1+cuda9.0_amd64.deb

   Install the code samples and the cuDNN Library User Guide, for example: sudo dpkg -i libcudnn7-doc_7.0.3.11-1+cuda9.0_amd64.deb

now it should work. And i would recommend you to work with virtualevn.

[yuzw213]

hi，I used cada 9.0 and downloaded and installed tensorflow-gpu 1.8.0, but I still reported the same error when I executed sh make.sh. What's the reason?

[yuzw213]

hi， @Luedeke I tried your method and still reported such a mistake..

[Luedeke]

@yuzw213 I'm sorry, but then I can't help you either.