Sber added an analogue of generative filling from Photoshop to the Kandinsky neural network
Miscellaneous / / November 22, 2023
Sber presented a major update to the Kandinsky generative neural network. With the update to Kandinsky 3.0, illustration generation has significantly improved, and the new Kandinsky Video is the first Russian neural network to create videos.
Let's start with Kandinsky 3.0. The main innovation compared to version 2.2 was improved query recognition: generation now more accurately matches industrial data without sacrificing quality. In addition, the quality of understanding of queries on the topic of the domestic cultural code - heroes of Soviet and Russian films and cartoons - has increased exponentially. The difference was clearly shown in queries with Cheburashka and the brownie Kuzya:
They also showed a comparison with older versions of Kandinsky and other popular models - Midjourney (signed as MJv.52), Stable Diffusion XL (SDXL) and DALL-E 3. Here are the generation results for the query “beautiful girl”:
This is the “man with a beard”:
And this is “Barbie and Ken Go Shopping”:
Another innovation was the Inpainting and Outpainting modes - the ability to “fit” a new object into an existing picture or complete it (reminiscent generative fill in Photoshop). Here is an example of finishing:
And these are examples of adding an object.
You can try the neural network on the platform Fusion Brain, in the official Telegram bot or "In contact with».
Kandinsky Video allows you to create small animations based on a text request - up to 8 seconds with a frequency of about 30 frames per second and a resolution of up to 512 pixels on the larger side. The height and width can be set by the user.
Video Creation is available in beta mode on Fusion Brain, A Telegram bot For now it only offers to sign up for the waiting list.
More news about AI🤖🦾🌐
- The Google Photos app has learned to tidy up its gallery using AI
- ChatGPT voice queries are now available to users with free accounts
- OnePlus has launched a free neural network that creates music and videos