A question regarding whether the LoRA has been successfully applied to the inference process？

[zhangqi420]

I checked that load_lora only loads the LoRA parameters into self.loras in the Flux class, but this self.loras is not involved in the forward inference process.