Kandinsky 2.1 can mix several drawings, modify them according to a text description, create images similar to the given one, fill in the missing parts of the picture and create images in infinite canvas mode (inside paint/exterior paint). Requests are accepted in 101 languages. You can also set the image style while creating.
The new model was based on the previous version, but trained on an additional 170 million high-resolution text-image pairs. He also trained on a set of two million pairs of high-quality images that were assembled separately. Other improvements have also been made. As a result, the model contains 3.3 billion parameters instead of 2 billion in Kandinsky 2.0.
You can try the neural network using the “Start Art” command from the demo page of the model, on Sber smart devices and in the Salyut mobile app. There is also a Telegram bot.
Source: Ferra
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.
