File Limit Request: nm-vllm - 400 MiB

[mgoin]

Project URL

https://pypi.org/project/nm-vllm/

Does this project already exist?

• Yes

New Limit

400 MB

Update issue title

• I have updated the title.

Which indexes

PyPI

About the project

vLLM is a fast and easy-to-use library for LLM inference and serving that already has a file limit increase to 400 MB in issue # 3792. nm-vllm is an enterprise-supported fork of vLLM that requires similar file size because of the amount of compiled kernels.

Reasons for the request

Pre-compiling these kernels means that users can deploy quickly and deterministically, rather than needing to setup a compilation environment where they need to deploy. As we extend our optimized inference for more hardware platforms, the binary size will grow so we would like to follow the standard vLLM sets.

Code of Conduct

• I agree to follow the PSF Code of Conduct