For this model, our goal is to make it easily accessible for generating high-quality images through minimal prompts, ensuring stunning results without the need for additional embellishments. The objective is to create the image in a single generation by leveraging the strengths of each model and blending them into a new, user-friendly model. The focus is on producing realistic images akin to photographs or real people. Our ongoing development efforts will continue to enhance these capabilities.
Description
Recommended Parameters for V1.1: (Update wrong information in V1.0)
Sampler: DPM++ 2M karras, Euler Ancestral
Steps: 40-60
Clip Skip: 2
For optimal results, consider creating an image with dimensions of 1,536 x 1,024. However, if your machine experiences performance issues or is unable to handle this size efficiently, you have the option to use a smaller size, such as 1,024 x 768, and later enhance it using ESRGAN X2+ or ESRGAN X4+ (Image to Image).