Users of the Kandinsky 2.1 generative model developed by Sber created more than 70 million unique images in three months. The most popular queries include “Russia”, “love”, “anime”, “cat” and “space”. This was announced by Alexander Vedyakhin, the first deputy chairman of the board of directors of Sberbank.
He emphasized that such an explosive interest in neural networks among residents of our country is absolutely natural, since Russia occupies a leading position in the development of technologies in the field of artificial intelligence.
“We created and develop Kandinsky so that people can discover new ideas for creativity in themselves and use our model absolutely free to implement them,” said Vedyakhin.
Recall that Sber introduced the Kandinsky 2.1 productive model in early April. It reached 1 million users within four days of its release, making it the fastest growing neural network in the world.
The Kandinsky 2.1 generative model was developed and trained by Sber AI researchers with the joint support of scientists from AIRI Institute of Artificial Intelligence on the combined dataset of Sber AI and SberDevices.
The model can generate images in just a few seconds based on a natural language text query. The user can choose from 20 different styles, including “classicism”, “mosaic”, “Khokhloma”, “Aivazovsky” and others.
Kandinsky 2.1 understands Russian and English requests and makes it possible to create a picture in endless canvas mode, draw parts of the image and mix several drawings.
Previously a representative of Sberbank said on the future of artificial intelligence in the creative industries.