V2 is a prior epoch in case V1 is overtrained for some base models.

Trained to add more details to hands.

Both sample images are the same prompt/seed. One just has the lora applied.