vllm - 💡(How to fix) Fix [New Model]: Cuda 13 wheels for Blackwell GPUs, Linux-aarm64 and Linux-aamd64 Please [1 comments, 2 participants]

TL;DR

The user needs a pre-built VLLM runtime for the Blackwell GPU environment to simplify adoption and avoid manual building.

Guidance

• Investigate the feasibility of creating a PyPI installable runtime for the Blackwell GPU environment to streamline the installation process.
• Review the VLLM documentation and existing issues to see if there are any plans or workarounds for supporting the Blackwell GPU.
• Consider reaching out to the VLLM community or maintainers to discuss the possibility of adding support for the Blackwell GPU environment.
• Evaluate the effort required to manually build VLLM for the Blackwell GPU and assess whether it's a viable temporary solution.

Notes

The issue lacks specific technical details about the Blackwell GPU environment and the VLLM version being used, which may impact the feasibility of providing a pre-built runtime.

Recommendation

Apply workaround: Manually build VLLM for the Blackwell GPU environment until a pre-built runtime is available, as it seems to be the current viable option.