This is the best summary I could come up with:
Ahead of AMD's Advancing AI event coming up quickly in early December, AMD today announced ROCm and PyTorch support has been extended to supporting the Radeon RX 7900 XT graphics card.
It's great seeing them provide official ROCm + PyTorch support now for the Radeon RX 7900 XT, but a bit frustrating that it's taken so long while also continuing to only support ROCm so far on the 7900 level products.
Meanwhile on the NVIDIA side their CUDA GPU support is so robust and can be practically taken for granted.
Even on the Intel side their open-source GPU compute runtime has evolved quite well and is running great across all of their consumer and workstation and DC graphics hardware.
It's just on AMD the frustrating aspect that the official ROCm support continues to be quite limited -- and often times belated -- in comparison to NVIDIA and now even Intel graphics.
AMD's official support for ROCm and PyTorch with the Radeon RX 7900 XT was announced a few minutes ago on their blog.
The original article contains 256 words, the summary contains 173 words. Saved 32%. I'm a bot and I'm open source!