Fine-tune using LoRA

Hi, 

I would like to know if it is possible to fine-tune the model for the specific downstream task using LoRA? 

I noticed that there is a file related to LoRA: `fourm/models/lora_utils.py` but could not find how it is utilized. It would be highly appreciated if you can provide a tutorial of how we can use LoRA for fine-tuning? Thank you!