Skip to content
This repository has been archived by the owner on Mar 6, 2023. It is now read-only.

Tensorflow CUDA GPU support #6

Open
gmernov opened this issue Mar 31, 2022 · 1 comment
Open

Tensorflow CUDA GPU support #6

gmernov opened this issue Mar 31, 2022 · 1 comment

Comments

@gmernov
Copy link

gmernov commented Mar 31, 2022

Can i ask you to compile tensorflow aarch64 with CUDA 10.2 and Python 3.8 support ? I've installed your version, but it seems like it doesn't use GPU for calculations:

tf.config.list_physical_devices('GPU')
[]

I'm using NVIDIA's Jetson Nano device.
I'm trying to build it in a cross compile environment. But i spend almost month and tons of tries to, but still have no result.

@KumaTea
Copy link
Owner

KumaTea commented Sep 12, 2022

Hi! The binaries provided here (both I compiled and Linaro's builds) are built on CPUs only.

To make use of Nvidia GPUs on an ARM board, I'm afraid that you'll have to compile it natively on your SBC.

Could you provide some errors printed while compiling?

Sign up for free to subscribe to this conversation on GitHub. Already have an account? Sign in.
Labels
None yet
Projects
None yet
Development

No branches or pull requests

2 participants