Skip to content

Comment on How to run DeepSeek R1 locally

Comments

DeepSeek is available via HuggingFace and can be run in/from tools like LM Studio locally with enough Ram. 14B models is pretty impressive.

If you want the optionality of using the full model, there are private hosted models that can be connected in cheaply for those use cases into a "one place for all the models" locally.

AboutSource Built by g1lg1l

Hackerly is an independent reader for Hacker News, built on the public HN API. Not affiliated with Y Combinator.