The model is already available on Mage.Space (main sponsor)
It's important! Read it!
The model is still in the training phase. This is not the final version and may contain artifacts and perform poorly in some cases.
RealVisXL V3.0 Status: - Training Images: 2000 - Training Steps: 400k - Approximate percentage of completion: ~22%
The model is aimed at photorealism. Can produce sfw and nsfw images of decent quality.
Recommended Negative Prompt:
(worst quality, low quality, illustration, 3d, 2d, painting, cartoons, sketch), open mouth
or another negative prompt
Recommended Generation Parameters:
Sampling Steps: 15-30
Sampling Method: DPM++ SDE Karras / or another
Hires Fix Parameters:
Upscaler: 4x-UltraSharp upscaler / or another
Denoising strength: 0.1-0.35
Upscale by: 1.1-2.0
This model is:
The training took place locally on a single video card and took about 140 hours. Version 2.0 is not the final version, the model will be updated.
Huge thanks to the creators of these great models that were used in the merge.