This was inspired by a similar model for SD 1.5.
Does anyone else think V2 is somehow worse? Let me know.
Reduced image count, and improved tag dataset. Does anyone else think this is somehow worse?
Jewel_V2.safetensors