epic: llama.cpp should support `LLVM` for ARM-based CPUs

janhq / cortex.cpp

Run and customize Local LLMs.

https://cortex.so

Apache License 2.0

1.92k stars 105 forks source link

Open dan-homebrew opened 13 hours ago

dan-homebrew commented 13 hours ago

[ ] What should be the hardware heuristic to opt for LLVM?
[ ] Should Installer support this?
[ ] This may require us to have cortex engines be able to define a "default" llama.cpp version (if two or more exist)

dan-homebrew commented 13 hours ago

@eckartal Can you link the Reddit post here (and any other requests for this?)

eckartal commented 8 hours ago

@eckartal Can you link the Reddit post here (and any other requests for this?)