We provides two example B-LoRA models for the example ComfyUI workflows here:
https://github.com/liusida/ComfyUI-B-LoRA
They are trained using one picture and the training code is here:
https://github.com/yardenfren1996/B-LoRA
Training was done in 15 minutes using RTX 4090 for each B-LoRA model.