NumPy on Giant Board?

Has anyone got NumPy working on the Giant Board? There’s not enough memory to compile it on the device, and some of the pre-compiled versions for ARM/Raspberry Pi haven’t worked for me.

A lot of the project examples I’ve seen use it, so it would be nice to have it to make getting started easier, but can make do without it if it’s not possible to run on such a constrained device.

I can see about building it into the rootfs. I just need a few days to implement it.

1 Like

Thanks for looking into this. Is it possible?

Running some tests right now on the image, but I should be able to push it up here shortly.

There’s a new image with ‘numpy’ in the name of you want to give it a try. Basic tests seem to pass but I’m not sure if everything will work due to memory limitations.

Thank you for NumPy. It’s definitely working. Unfortunately, it turns out I need additional packages like scipy, etc which also can’t be installed on device due to memory constraints. Can you explain how you created this custom image so I can modify and build it myself to suit my needs? If the guide is long or complex, don’t worry about it. I’ll find a workaround. Thanks again!

I have a guide in the docs on building custom images. I recommend using Ubuntu 18.04 or greater when building. WSL on windows won’t work at the moment due to the way the images are made. Looking into alternative build systems that are cross platform and don’t take a lot to get going. If you run into any issues building the images, feel free to ask any questions.

1 Like

Alright, I can handle that.

Leaving a note for myself when I get around to it: This script installs all python packages from this text file.