RuntimeError: Not compiled with GPU support

[dhirajsuvarna]

Hi @ying09 ,

I followed the instructions in the step-by-step installation.txt and was able to go through with no issues.

However, when I try to run the demo\icdar2013_detection.py along with the required options, I get an error RuntimeError: Not compiled with GPU support.

Both the input options and the error is shown in the screenshot below -

Let me know if you need any other information.

[Real-YeJ]

@dhirajsuvarna When you compile TextFuseNet(or detectron2), it's necessary to compile with cuda and make sure your cuda and pytorch match.

[dhirajsuvarna]

@Real-YeJ , I did some searching into this error and it seems that CUDA_HOME env variable is not set.

When you say compile with cuda do I need to do anything other than the steps mentioned in step-by-step installation.txt? I guess this conda install pytorch=1.3.1 torchvision cudatoolkit=10.1 -c pytorch installs the required cuda toolkit.

Is any other separate installation of cuda is required? or do I need to pass any compile time parameters? or any other steps that needs to be taken?

[Real-YeJ]

@dhirajsuvarna don't need other additional steps，just make sure your cuda version is match with pytorch version. more details https://github.com/facebookresearch/detectron2/issues/55

[dhirajsuvarna]

Here are the versions of the the packages installed. Those seem to be in sync.

I manually added CUDA_HOME into my env point it to the folder shown below

sadly still facing the error.

Do you think my CUDA_HOME is pointing to the correct location? or should it be pointing to one of the three folders inside?

[Real-YeJ]

@dhirajsuvarna check the version of your cuda, and I suggest that compile on linux instead of windows.