SD3 Legacy
Introducing the SD3 Fantasy model, a highly experimental model designed for a wide range of creative applications. This beta version leverages a CLIP and T5XXL text encoder to generate images
Key Features:
Compatibility: Requires enabling TAMS 2.0 to run this model.
Performance: Initial runs may not yield optimal results, but improvements are on the horizon with future updates.
Configuration Guidelines:
Realistic Images: Use CFG 4-6
Anime and Other Styles: Use CFG 6-8
Clip Encoder: None
Steps: 20-40
Sampler: dpmpp_2m sgm uniform
![girl,realistic,(Best quality, 8k, 32k, Masterpiece, UHD:1.2),beautiful big eyes, round eyes, pretty face, from front, best quality, ((realistic)), high quality, ultra detailed, ((Real picture)), ((realistic skin)), ((realistic face)), sidelighting, naturalm, realistic, [(detailed face:1.2):0.2], shiny eyes, looking at viewer, delicate,high quality, colorful, (photography, photorealistic:1.1), (perfect face:1.4),(lora:zrmswtt:0.95), dynamic angle, movie lighting, portrait, high definition image quality, sexy, 22(yo),beautiful body,(RAW photo, best quality), (realistic, realistic: 1.3), very long hair, silver hair, white hairband, big chest, crystal earrings, cross necklace, military uniform, big scarf, sitting, Leg, winter, Arc de Triomphe](https://image.tensorartassets.com/cdn-cgi/image/anim=false,plain=false,w=300,q=85/posts/images/738273679812785153/3c922bee-b993-44cb-ab65-0ecc467721e4)
