How to change code to run with Intel GPU from CUDA

[advanchip]

I have downloaded an application source code from github and its prerequisites are:

* Python (3.6)

* PyTorch (0.4.1)

* torchvision (0.2.0)

* NumPy (1.14.3)

* SciPy (1.0.0)

* H5py (2.7.1)

* imageio (2.4.1)

* pandas (0.22.0)

* tqdm (4.19.8)

* opencv-python (3.4.2)

1. How can I verify and install its prerequisites in Intel DevCloud for the Edge Jupyter Notebook environment?
2. It requires to run “bash run.sh”.  I can run this command in Anaconda3 Windows 10 environment, but it didn’t work in Intel DevCloud for the Edge Jupyter Notebook environment. Please see attached error and let me know how to run this command in Intel DevCloud for the Edge Jupyter Notebook. 
3. The code was set by default to run with NVIDIA GPU in CUDA, please see attached error and tell me how to change code to run with Intel GPU. 
4. How to submit jobs to the compute nodes with an Intel GPU?

Thanks.

[ChrisB_Intel · Accepted solution]

Advanchip,

Not knowing exactly what you are trying to run, I may not be able to answer your questions completely. If you need some additional information, please provide me with the information around what you are running. If you don't want it public feel free to send me a private message.  

We do not give users on the DevCloud the ability to install, upgrade, or change software. So you will not have permissions to do some of the things you are trying. With regards to which version are running, you can use the command !pip3 show <package name>.

Not sure if you have checked out any of the samples we have a available, but you may find one that has what you are looking for.  

Thanks

Chris

Thanks

Chris