Proposal: Volunteer Maintainer for Qwen Model Series(may be more)

Open linkedlist771 opened this issue 6 months ago • 2 comments

Hi team, I’ve spent the last few months modeling the Qwen model series on an RTX 3090, consistently achieving ~10 % limit-error on memory consumption and latency. Given the steady stream of new releases, I’d like to volunteer as a maintainer for the Qwen checkpoints/configs, keeping them up-to-date and performance-verified. If this sounds helpful, I’m happy to open a PR with my current workflow, evaluation scripts, and results so the community can review and reproduce them. Please let me know if you’re interested or have any requirements before I proceed.

Jun 06 '25 03:06 linkedlist771

@linkedlist771 That would be awesome! Please open a PR and we can get that merged soon for the community to use. Thanks!

Aug 28 '25 17:08 cli99

Ok, soon or later I may a PR. I am wondering the DeepSeek's analysis is available or not in current version.

Aug 29 '25 06:08 linkedlist771