Update readme -> inference is local (via hugging face), not cloud. Users need to observe gpu / vram. Mention min machine specs.

[g-h-0-S-t]

Update readme -> inference is local (via hugging face), not cloud. Users need to observe gpu / vram. Mention min machine specs.

[kessler]

Clearly says no cloud on the first paragraph...
re: min spec and memory constraints, I'm on it