fix(lora): keep fused LoRA weights in the base weight's dtype - #2527

fix(lora): keep fused LoRA weights in the base weight's dtype#2527

inureyes merged 1 commit into

lablup:mainlablup/mlxcel:mainfrom

daemyung-lablup:fix/lora-fuse-base-dtypedaemyung-lablup/mlxcel:fix/lora-fuse-base-dtypeCopy head branch name to clipboard