News for hackers. Stuff that matters.

A comment on LLaMA-7B in Pure C++ with full Apple Silicon support.

Re:LLaMA-7B in Pure C++ with full Apple Silicon support by KETpXDDzR on Monday March 13, 2023 @02:38PM (#35140362)
I can confirm that the 7B and 13B models run on a 8GB NVIDIA 1080. The original Meta implementation always crashed with "out of memory". Too bad Llama doesn't work as well as (Chat-)GPT for instructional prompts.
[ Parent | on HN ]

There are no comments on this comment yet.