I made an ultra short howto how you can install deepseek R1 distills on your own hardware. I do these kind of things in an incus based qemu vm just for isolating these ai models from the rest of the system.
If you have some tips or additions, feel free to write a comment.
cheers