I'm the author of Private LLM. Looks like it's just become possible[1] to run quantized LLM inference using the ANE with iOS 18. I think there are some major ef

I'm the author of Private LLM. Looks like it's just become possible[1] to run quantized LLM inference using the ANE with iOS 18. I think there are some major efficiency gains on the table now. [1]: https://github.com/apple/coremltools/pull/2232 https://github.com/apple/coremltools/pull/2232