After the release of the new version of Sber’s prolific model Kandinsky 2.1, users created more than 1.3 million views in just 48 hours, the bank’s press service reported.
The most popular queries were: “cat”, “love” and “space”.
The neural network was developed and trained by Sber AI researchers with the joint support of scientists from AIRI Artificial Intelligence Institute on the combined dataset of Sber AI and SberDevices.
The Kandinsky 2.1 model is trained on a separate aggregated dataset of approximately 1.2 billion text-image pairs, as well as two million pairs of high-quality images.
Sber presented a new version of the prolific model on April 4. Kandinsky 2.1 can generate images from natural language text descriptions in seconds. User can create images in more than ten styles.