wskwon

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

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

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).

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

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

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

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