Is your feature request related to a problem? Please describe.
it would be nice for people to be able to run localai on their surplus of raspberry pi's so my proposal is for a arm64 docker image especially for people tinkering with a turing pi cluster even though it would be slow as hell maybe even run it using vulkan since the pi gpu driver is now vulkan conformant . maybe even a jetson nano specific build

Describe the solution you'd like
an arm64 docker build with optional vulkan support in time maybe if feasible

Describe alternatives you've considered
gpus and cpu's are expensive and guzzle power

Additional context
x gpu's and cpu's are expensive and a raspberry pi is almost universally part of someones tinker environment