“Yandex” presented a prototype of its own neural network for generating images based on user text queries, the company said. You can try it out in the 0.1.4 beta version of the Masterpiece app, which is available on Android and iOS.
Shedevroom recognizes Russian and English and is best responsive to specific requests.
The neural network also takes into account special wishes (eg “photorealism” or “high detail”), can imitate famous painters and work in certain artistic styles.
The Masterroom neural network creates images using the cascade diffusion method: first, it forms images according to the request, after which it gradually increases their resolution and saturates them with details.
The first version of the generative model was trained on 240 million image examples with text descriptions. Currently, training continues on a dataset of 500 million examples, and Yandex promises to improve the quality in future updates.
To generate images based on your own requests, you need to leave a request and wait for your confirmation. Subsequently, all registered users will receive this opportunity.
According to Aleksey Gusakov, head of Yandex’s research and artificial intelligence department, by the time of the final release, the neural network will be able to generate banners, illustrations, and create images for online storefronts.
Sberbank is also working on creating a neural network to generate images based on a text description. The day before, the company released the third update to its Kandinsky model, which is trained on a billion text-image pairs and creates images in seconds.
I am a professional journalist and content creator with extensive experience writing for news websites. I currently work as an author at Gadget Onus, where I specialize in covering hot news topics. My written pieces have been published on some of the biggest media outlets around the world, including The Guardian and BBC News.