distributed-llama icon indicating copy to clipboard operation
distributed-llama copied to clipboard

Will this awesome proj consider supporting GPU acceleration?

Open galenyu opened this issue 1 year ago • 5 comments

A very impressive job!

But it doesn't seem to support the use of GPU. Does the author consider developing code that supports GPU acceleration?

Any suggestions to migrate this project to CUDA/HIP acceleration?

Thanks for any help!

galenyu avatar Apr 29 '24 13:04 galenyu

Hello @galenyu! Yes, GPU acceleration is planned.

b4rtaz avatar Apr 29 '24 20:04 b4rtaz

hello, thanks for your job. when will the gpu acceleration version be released?

460130107 avatar Jul 16 '24 09:07 460130107

Hello. I am currently trying dllama. Also I have a supercomputer, 6 nodes, 96Cores, 768Go RAM, 6 PNY Nvidia RTX 4000 Ada Generation, and I need GPU support.

So, no promise, because I have other related project on the fire, but what is missing, please ? Can you do me a summary of your advancement, ... etc.

Thanks in advance. Best Regards. Benjamin.

lipere123 avatar Aug 14 '24 04:08 lipere123

I can run a Llama 3.1 70B Instruct Q40

lipere123 avatar Aug 14 '24 04:08 lipere123

@b4rtaz im bumping this as im interested to know what the status of this is?

Thanks.

pcfreak30 avatar Feb 19 '25 17:02 pcfreak30