Open yuno779 opened 3 months ago
Illustrate the enigmatic Alessandra Ambrosio inspired Mistress in the distinctive, quirky style of Tim Burton, a dark and alluring figure, exudes an aura of mystery and elegance. Her pale, porcelain skin contrasts sharply with her long jet-black, hair, which frames her face in a wild, untamed manner. Her eyes, large and expressive, are a piercing shade of emerald green, capturing the viewer's attention with their intense gaze. A mischievous smile plays upon her lips, hinting at hidden desires and secrets. Behind her, a backdrop of twisted, gnarled trees and eerie, moonlit skies sets the stage for this captivating and haunting portrait. There is a tattoo on her shoulder reading "Crypt Keeper". It isn't half as bad as people say and its good with text. It isn't really finetuned for poses among other things... yet. This much is apparent.
Illustrate the enigmatic Alessandra Ambrosio inspired Mistress in the distinctive, quirky style of Tim Burton, a dark and alluring figure, exudes an aura of mystery and elegance. Her pale, porcelain skin contrasts sharply with her long jet-black, hair, which frames her face in a wild, untamed manner. Her eyes, large and expressive, are a piercing shade of emerald green, capturing the viewer's attention with their intense gaze. A mischievous smile plays upon her lips, hinting at hidden desires and secrets. Behind her, a backdrop of twisted, gnarled trees and eerie, moonlit skies sets the stage for this captivating and haunting portrait. There is a tattoo on her shoulder reading "Crypt Keeper". It isn't half as bad as people say and its good with text. It isn't really finetuned for poses among other things... yet. This much is apparent.
you are missing the point. CLoseup portraits are realy good. Problem is full body shots when people dont just stand like statues. Anything hard like lying , sitting, running etc is a mess
Poses in general. Its not a pony. Still so far any model has been extensively retrained into variants which address these issues - its clearly the training data as there are other things it can not do very well. These are trained right now. That being said i don't see a reason why other models shouldn't be included as well.
Illustrate the enigmatic Alessandra Ambrosio inspired Mistress in the distinctive, quirky style of Tim Burton, a dark and alluring figure, exudes an aura of mystery and elegance. Her pale, porcelain skin contrasts sharply with her long jet-black, hair, which frames her face in a wild, untamed manner. Her eyes, large and expressive, are a piercing shade of emerald green, capturing the viewer's attention with their intense gaze. A mischievous smile plays upon her lips, hinting at hidden desires and secrets. Behind her, a backdrop of twisted, gnarled trees and eerie, moonlit skies sets the stage for this captivating and haunting portrait. There is a tattoo on her shoulder reading "Crypt Keeper". It isn't half as bad as people say and its good with text. It isn't really finetuned for poses among other things... yet. This much is apparent.
you are missing the point. CLoseup portraits are realy good. Problem is full body shots when people dont just stand like statues. Anything hard like lying , sitting, running etc is a mess
This is not true. Maybe for Anime, but there are a LOT of us who just don't care about anime.
Also
Civitai has banned SD3 content
Well, what do you expect? Did you read the TOS that SD dropped? Ridiculous garbage. We all knew it was coming to this though.
Is there an existing issue for this?
What would your feature do ?
Support the inference use of the HunyuanDIT model.
Proposed workflow
Able to perform t2i
Additional information
Since the SD3 effect is so poor, can we support the use of HunyuanDIT? https://www.reddit.com/r/StableDiffusion/comments/1dehbpo/sd3_vs_hunyuandit/#lightbox
https://huggingface.co/Tencent-Hunyuan/HunyuanDiT https://github.com/Tencent/HunyuanDiT https://dit.hunyuan.tencent.com/