mirror of
https://github.com/ikawrakow/ik_llama.cpp.git
synced 2026-07-01 07:30:20 -05:00
* Allow quantize tool to only copy tensors to allow repackaging models. * Slightly better logic when requantizing. * Change help message to go to `stdout`.
quantize
TODO