wskwon

vor 3 Jahrendiscuss

Thanks! Please try it out and share any feedback you might have.

vor 3 Jahrendiscuss

Thanks for the explanation! I believe the two ideas are basically orthogonal. FlashAttention reduces memory read/writes, while PagedAttention reduces memory waste.

wskwon

vor 3 Jahrendiscuss

Yes, vLLM focuses on maximizing throughput when the VRAM is fully utilized. Nevertheless, I believe users can still benefit from vLLM even if they don't utilize the memory to its full capacity, because vLLM also includes other optimizations orthogonal to the PagedAttention (e.g., optimized CUDA kernels).

wskwon

vor 3 Jahrendiscuss

We used matplotlib for the performance charts, and used a free website to convert google slides to the animation gifs.

wskwon

vor 3 Jahrendiscuss

Not really. vLLM optimizes the throughput of your LLM, but does not reduce the minimum required amount of resource to run your model.

wskwon

vor 3 Jahrendiscuss

You can think of LMSYS Vicuna: https://chat.lmsys.org as our hosted demo, as it actually uses vLLM as the backend.

wskwon

vor 3 Jahrendiscuss

vLLM has been adopted by LMSYS for serving Vicuna and Chatbot Arena.

Submissions

vLLM: Easy, Fast, and Cheap LLM Serving with PagedAttention

comments