( Not my Model )
https://civitai.red/models/2914859/sophos-krea?modelVersionId=3297567
Sophos Krea
It took a while for me to arrive at a mix with Krea 2 that I felt was worth sharing. My goal was to produce a model that is easy to use to achieve a certain aesthetic realism style with Krea 2. This model is stylistically opinionated, so be advised that it doesn't do all styles well.
đź“ť Version Notes
Consult the notes below for the differences between the posted versions.
v1.0
This is the initial release. A solid all-around photography model, best used for realistic styles. Can do NSFW. Has a bias towards "aesthetic" or "professional-looking" photography. Detailed, aggressive prompting or amateur photo LoRAs will be needed to get a low-quality amateur look out of this model. You will arguably do better with a different model for that style.
NOTE: This version handles text rendering fine up to a point, but it has a tendency to mangle long text strings. If that happens, try increasing the weight of the text input in the prompt and reroll a few times before giving up.
Suggested Uses
This model is best used for cinematic or aesthetic photos and realistic 3D styles. If you want other styles, like cartoons, paintings, anime, or drawings, you'll be better served by other models or you will need to apply LoRAs—and even then you'll be going against the grain. This model was tuned for photos.
Recommended Workflow
Every example image embeds the ComfyUI workflow I used to generate the image. You can drag and drop any image directly into ComfyUI to load up the workflow.
I like using euler with bong_tangent as the scheduler, 12 steps.
I strongly recommend the Krea 2 Real Vae. It is linked from the Vae section on this model card.
Merge Details
Credit goes where credit is due. This merge wouldn't have been possible without the efforts of the people who created the following constituent models and LoRAs. Please check out their models, like and subscribe, send them some donations, etc.
Sophos Krea v1.0
Checkpoints Used
LoRAs Used