One of the most popular neural networks for generating images based on text queries recently received a major update. The main advantage of the Stable Diffusion XL is the ability to deploy it locally on a device and create images without the need for an internet connection. And this can be done not only on powerful computers but even on iPhones starting from the 11th generation.
The new generation of the Stable Diffusion neural network, known as Stable Diffusion XL 1.0 (SDXL), allows for the generation of more detailed images with a base resolution of 1024 × 1024. The color reproduction, contrast, and shadow processing have also been improved. One interesting new feature is the ability to generate readable text and calligraphy. Previously, users had to rely on something that only remotely resembled text. Since Stable Diffusion is launched locally, it is practically impossible to censor the results of its work. However, with this update, the developers have made an effort to limit the generation of malicious content by cleaning the data set used to train the model.
The source, Arstechnica, reports that Stable Diffusion XL is a significant upgrade from the previous version. The improvements in image quality, readability of text, and the ability to generate calligraphy make it a valuable tool for various applications, from design to content creation.
Overall, the Stable Diffusion XL neural network update brings enhanced image generation capabilities with improved resolution, color reproduction, contrast, and the ability to generate readable text and calligraphy. Its ability to run locally without an internet connection allows for greater accessibility and privacy. The developers’ efforts to limit the generation of malicious content further enhance the network’s reliability.