Vulkan support

[pierotofy]

Vulkan has compute capabilities and could be used for speeding up the training with other GPU brands.

https://www.khronos.org/blog/getting-started-with-vulkan-compute-acceleration

It could also be possible to compile torch to use Vulkan, but that's pretty experimental it seems. https://pytorch.org/tutorials/prototype/vulkan_workflow.html

[pfxuan]

Vulkan backend seems very promising:

• https://blog.nomic.ai/posts/gpt4all-gpu-inference-with-vulkan
• https://github.com/DTolm/VkFFT
• https://github.com/shg8/3DGS.cpp/tree/main

[pierotofy]

Also https://github.com/KomputeProject/kompute/

[pfxuan]

Kompute looks really good. Right now, it's an incubating project in LF AI & Data landscape: https://landscape.lfai.foundation/card-mode?category=programming