Did you do full model fine tuning (all layers) or only adapters?
#2
by
SamuelAzran
- opened
Thanks for publishing this!
Thanks! It was trained as a QLoRA targeting all linear layers
emozilla
changed discussion status to
closed