tensorflow-upstream VS linux

I tried to train another model using InceptionResNetV2 and the same issues happens. Also, this happens even using the model.predict() method if using the GPU. Probably this is an issue related to the AMD Radeon RX 6700 XT or some mine misconfiguration. System Inormation: ArchLinux 6.1.32-1-lts - AMD Radeon RX 6700 XT - gfx1031 Opened issues: - https://github.com/RadeonOpenCompute/ROCm/issues/2250 - https://github.com/ROCmSoftwarePlatform/tensorflow-upstream/issues/2125

AMD has had their ROCM Tensorflow port for quite a while now: https://github.com/ROCmSoftwarePlatform/tensorflow-upstream and it works pretty well on my RX 6800; unfortunately it's only compatible with Linux as the ROCm stack is built on top of the Linux kernel I believe

I mean, tensorflow has a fork with ROCm support which is maintained by AMD https://github.com/ROCmSoftwarePlatform/tensorflow-upstream although I'm not entirely sure what you're AI workloads are specifically, I'm just throwing out tensorflow because it's popular. On the enterprise side they also have radeon instinct MI, although I assume you're probably not using enterprise HW but I wanted to throw it out there anyway.

It might work, but will likely require building pytorch/tensorflow from source. I compiled both with ROCm 5.0 today, Pytorch is painless but TF required using this repo maintained by AMD instead

I saw this https://github.com/ROCmSoftwarePlatform/tensorflow-upstream and I thought it could be a way. Do you think it is not stable and ready to be used ?

The repo exists: https://github.com/ROCmSoftwarePlatform/tensorflow-upstream

https://github.com/ROCmSoftwarePlatform/tensorflow-upstream#tensorflow-rocm-port

I tried some case ROCm & Tensorflow-rocm as is on this page( tensorflow-upstream/tensorflow-rocm-release.md at develop-upstream · ROCmSoftwarePlatform/tensorflow-upstream · GitHub ), but I failed to run a simple CNN model with fashion-mnist datasets.

Interesting! I've used Linux Mint for the last 5+ years (and am using the Xanmod kernel, which is on 6.6), but have always been Pop OS curious. This increases my curiosity, I may have to give it a go on my laptop.

https://xanmod.org/

For those on Linux mint you can install the xanmod kernel for improved performance, specifically tailored for gaming on Debian based distros.

Generally, Arch is the better platform for getting various kernel versions, but there's always stuff like Xanmod which is easy to do in any distro. Worst case, could always use linux-tkg script to build your own kernel - with that kind of hardware it shouldn't take long for the kernel to be built.

Still a bit of a noob. I haven't heard of that one. Do you mean this: https://xanmod.org/? Is there any stability or compatibility issues that come with it?

The ArchWiki has some tips that can help tuning your system to squeeze more performance: https://wiki.archlinux.org/title/gaming

This launcher does help too: https://github.com/FeralInteractive/gamemode

On top of that, you can use a patched kernel like: https://xanmod.org/

You can check support status for games here: https://www.protondb.com/

If you run games via Proton with DXVK, you can use the environment variable DXVK_HUD=1 to show a HUD with an FPS meter. Some others prefer mangohud.

Setting your CPU frequency scaling governor and GPU to performance mode also helps.

Xanmod Kernel

There are different drivers available but the best way is to install linux kernel 6.3 which has built in support for rtl wifi. I also have similar adapter and had the problem. I recommend xanmod kernel because it is newer and has the needed driver and it is focussed on desktop performance.

For Debian : https://xanmod.org

Alternatively, try the xanmod kernel https://xanmod.org/