Visuali

The quickest way to get started with our illustration generator is to use pre-made models. If you want to depict a style or an object that is not available in the premade models you can train your own fine tune by uploading 5-15 illustrations. there are no limits to fine-tuning you can use it for illustrations icons or any assets you need. Read more about fine-tuning. Illustrations are exportable in PNG and SVG formats. Fine-tuning allows you to train the stable-diffusion AI model, on a particular object or style, and create a new model that generates images of those objects or styles. The fine-tuning will be only as good as the data you provide. Around 5-15 images are recommended for fine-tuning. Images can be of any unique object or style. Images should contain only the subject itself, without background noise or other objects. Images must not include any gradients or shadows if you want to export it as SVG later. PNG export still works fine with gradients and shadows.

Introducing Mobile Diffusion, the innovative image generator that uses the latest AI technology to bring your imagination to life. With this app, you can create stunning images based on your own text prompt. No need for an internet connection, it works offline right on your device. Mobile Diffusion uses the Stable Diffusion v2.1 model to power its AI-based image generation. Thanks to CoreML optimization, it’s up to 2x faster than other image generation apps. It requires just a one-time download of the 4.5 GB model to work offline, and then you can use it anytime, anywhere. With the ability to specify both positive and negative prompts, you can fine-tune your image output to suit your needs. Sharing your generated images is easy, and the app is completely free to use. This app was made for research and development purposes only. The goal was to demonstrate the ability to run a diffusion model on a mobile device with acceptable performance.