All credits to SG_161222

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 Overall Status: - Training Images: 960 - Training Steps: 192k - Approximate percentage of completion: ~18% RealVisXL V2.0 Status: - Training Images: +200 - Training Steps: +40k - Approximate percentage of completion: ~24%

The model is aimed at photorealism. Can produce sfw and nsfw images of decent quality.

ᅠ

Recommended Negative Prompt: (worst quality, low quality, illustration, 3d, 2d, painting, cartoons, sketch), open mouth or another negative prompt

ᅠ

Recommended Generation Parameters: Sampling Steps: 15-40 Sampling Method: DPM++ SDE Karras

ᅠ

Hires Fix Parameters:

Upscaler: 4x-UltraSharp upscaler

Denoising strength: 0.25-0.5

Upscale by: 1.1-2.0

ᅠ

Optional Parameters:

ENSD: 31337

ᅠ

This model is:

A + B = RealVisXL

A = Trained model on almost 1 thousand images

B = An merge of several models (listed below)

The training took place locally on a single video card and took about 80 hours. Version 1.0 is not the final version, the model will be updated.