This model delivers decent photorealistic capabilities, truly shining in close-up images. You can expect highly realistic and detailed results when generating visuals that zoom in on subjects. It's a solid option for producing clear and well-defined close-up shots. Built on the SDXL model, it also offers various capabilities for animated and CGI-based generation, though its primary focus remains on realism.
This model excels at intimate close-up shots across diverse subjects like people, races, species, and even machines. It's highly versatile with prompting, allowing for both SFW and decent NSFW outputs.
<Copy the node workflow from the demo images to get my exact workflow and defined values.>
How to use?
Prompt: Simple explanation of the image, try to specify your prompts simply.
Steps: 10 - 20
CFG Scale: 1.5 - 3
Personal settings. Portrait: (Steps: 10 + CFG Scale: 1.8), Details: (Steps: 20 + CFG Scale: 3)
Sampler: DPMPP_SDE +Karras Hires fix with another sampler for fixing irregularities
Face Detailer recommended if image is not a closeup portrait.
Vae baked in
Negative prompts example template:
Start with no negative prompts, and add whatever is not needed in negative or add the below for consistent
bad proportions, low resolution, bad, ugly, bad hands, bad teeth, terrible, painting, 3d, render, comic, anime, manga, unrealistic, flat, watermark, signature, worst quality, low quality
Image size:
Use any SDXL portrait or landscape resolutions. Or use the SDXL Empty Latent Image (rgthree)
~Note this is my first time working with image generation models, kindly share your thoughts and go nuts with the generation~