Query for training with lora

[sterzhang]

Hi, how can I finetune the model using lora.
Is it the same as LLaVA-1.5 like adding an extra line below?
--lora_enable True --lora_r 128 --lora_alpha 256 --mm_projector_lr 2e-5 \

[jzhang38]

I believe you are right. But I have not run with lora.