Does the SFT pipeline support LoRA / PEFT for training on lower-VRAM setups?

[minji00116-ctrl]

Does the SFT pipeline support LoRA / PEFT for training on lower-VRAM setups?

[super-anova]

Currently it doesn't support LoRA/PEFT as is. We can take that into consideration if there is strong demand for it from the community. So stay tuned for potential future capabilities!

[HuangChiEn]

+1 for this feature enhancement

Does the SFT pipeline support LoRA / PEFT for training on lower-VRAM setups?

[rdesc]

+1 !