You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
This paper introduces a contrastive alignment loss and ID loss to minimize the contamination of ID information on the original model while ensuring fidelity. So the generated results are very similar to SDXL Lightning but still preserve the ID consistency. If so, I think the vanilla ControlNet for SDXL Lightning should work well with PuLID without tuning or only a few iterations tuning. Is this correct?
The text was updated successfully, but these errors were encountered:
Hi! Thank you for the great work!
This paper introduces a contrastive alignment loss and ID loss to minimize the contamination of ID information on the original model while ensuring fidelity. So the generated results are very similar to SDXL Lightning but still preserve the ID consistency. If so, I think the vanilla ControlNet for SDXL Lightning should work well with PuLID without tuning or only a few iterations tuning. Is this correct?
The text was updated successfully, but these errors were encountered: