Closed bog-dan-ro closed 2 months ago
Dup of https://github.com/instructlab/instructlab/issues/659? In-progress here https://github.com/instructlab/instructlab/pull/1337. Please close if this is a dup or change the title/description for your enhancement request. Thanks!
It looks like, there is another macos only warning here https://github.com/instructlab/instructlab?tab=readme-ov-file#-serve-the-newly-trained-model but it seems it's a mistake as ilab serve
worked just fine on Linux. (Didn't test the ilab convert
though.
Just checked it: ilab convert
is only implemented for macOS with M-series chips for now.
Aren't you using convert-hf-to-gguf.py
for that job?
Tests for linux implemented, also eval exists. Closing.
Right now some features like
test
are available only on M-series, will be great if they will be available for other platforms (Linux cuda/rocm).