The Nvidia website has an online text-to-image generation tool called Stable Diffusion XL, which can be used for free online. Simply open the webpage to use it, no login required. Stable Diffusion XL (SDXL) can generate images based on word descriptions.
Stable Diffusion XL
Website: https://catalog.ngc.nvidia.com/orgs/nvidia/teams/playground/models/sdxl
Guests can use it 10 times, while logged-in users can use it unlimited times for free. Registering an account only requires email verification, no phone number or credit card needed. Simply access the website to open it. If you encounter a green triangle spinning or a failed human verification prompt, try refreshing the page or logging in again.
How to use:
Enter a text description, reverse description, set some parameters, and submit to generate an image. The speed is not very fast, taking around a few tens of seconds to generate once. It is best to use English descriptions as Chinese support is not very good.
The image above shows a close-up of a rose generated as an example, with its petals shining under the sunlight. By using a reverse description of white, a white rose will not be generated. The generated images do not have watermarks.
Image demonstration:
The following images were generated by Stable Diffusion XL:
The generated images are quite detailed and do not have watermarks like those generated by AI.
For more information, please visit: https://catalog.ngc.nvidia.com/orgs/nvidia/teams/playground/models/sdxl/overview
Interested students can test it out with detailed parameter settings.