frontpage.
newsnewestaskshowjobs

Made with ♥ by @iamnishanth

Open Source @Github

Open in hackernews

Optimising DeepSeek-R1-Distill-Qwen-7B for use in production

https://fin.ai/research/think-fast-reasoning-at-3ms-a-token/
1•destraynor•2h ago