Image Hub

Currently allows you to create up to 4 images at a time, and lets you set a "seed". Changing seeds allows you to get a slightly different output for the same prompt, by finding one that works, you can work on creating a collection of images that are similar / follow a similar theme (E.g. our cat NFT collection).

Soon this will have a history section, and an image editing section. Along with the ability to mint the images you make as NFTs, or create collections to mint whole NFT collections.

Note: As this model was trained to create our CyberCat NFT collection, it performs best with with the cartoon style. It can be used for other styles, but results are less consistent.

Increasing both the "inference" and "guidance" parameters in the Image Hub will generate image with more contrast and darker lines (more comic look). The top image is with more inference steps and guidance scale

Advanced

The process.

Inference steps:

numInferenceSteps (integer, optional): The number of inference steps to take during image generation. This controls how many steps the model takes during generation. More steps usually result in higher quality and more detailed images, but it takes longer to compute. Fewer steps are faster but can lead to less detailed or lower-quality results.

Initial seed 1-1000:

seeds (integer or array of integers, optional): Random seed(s) for reproducible results. Seeds are numbers that control the randomness in image generation. Using the same seed allows you to reproduce the exact same image each time, while changing the seed generates different variations of the image. Guidance scale:

guidanceScale (number, optional): The scale of guidance to be used during image generation. The guidance scale controls how closely the model follows the input prompt during image generation. A higher guidance scale makes the model generate images that are more aligned with the prompt but can sometimes reduce creativity or lead to unnatural results. A lower guidance scale allows more flexibility and randomness, which can produce more diverse or creative outputs, but might stray further from the prompt. Comic on/off:

useTag (bool, optional): if use the tag that triggers the comic/cartoon style.