Home SECURITY Google StyleDrop can draw the Mona Lisa in three minutes

Google StyleDrop can draw the Mona Lisa in three minutes

0
Google StyleDrop can draw the Mona Lisa in three minutes

[ad_1]

Google StyleDrop can draw the Mona Lisa in three minutes

A new tool from Google Research lets you create images from text in any style, even if you can’t draw.

Google Research has developed an innovative tool called StyleDrop that uses an image-by-text generator to enable users to create personalized images with specific artistic styles.


StyleDrop is an image-by-text generator that uses the Muse3 generative visual transformer introduced by Google earlier this year. Muse has been trained on 3 billion parameters, providing high quality image generation.

Users can suggest an image and specify an art style, such as “3D melting gold display”, “wood sculpture”, “3D display” or “cartoon drawing”. StyleDrop then generates impressive object images based on the desired style, as well as typography that matches the stylistic features of the images.


StyleDrop works by efficiently learning a new style by adjusting a very small number of trainable parameters (less than 1% of the total number of model parameters) and improving the quality through iterative learning with human or automated feedback. Moreover, StyleDrop is capable of producing impressive results even when the user provides only one image indicating the desired style.

Developers see this program, which has not yet been released to the public, as a valuable tool for art directors and graphic designers who can create photorealistic renderings of given products or themes, including text that reflects the same colors, structure, and style.


The study showed that for the task of styling text-to-image models StyleDrop on Muse convincingly outperforms other methods including DreamBooth, Imagen and Stable Diffusion.

[ad_2]

Source link

www.securitylab.ru

LEAVE A REPLY

Please enter your comment!
Please enter your name here