Stable Diffusion 3 Medium is Stability’s most advanced image AI

Stability AI released the Stable Diffusion 3 Medium open model to generate lifelike images with optimized performance on personal computers

Stability AI announced Stable Diffusion 3 Medium, the first open version of its next-generation AI model for image generation. According to the company, it is the “most advanced open text-to-image model to date” with two billion parameters.

• Image Playground | All about artificial intelligence to create Apple images
• Kwai owner launches artificial intelligence that creates realistic Sora-style videos

The biggest highlights are the realism and skill with typography. The first images revealed by SD3 Medium draw attention to the amount of details that AI normally has difficulty reproducing correctly, such as faces and hands, as well as providing high-quality results.

Diffusion Transformer’s architecture allows you to create images with text without typos or letter formatting issues, while the model can understand increasingly complex instructions to help customize the result.

Optimized performance

The two-billion-parameter model is optimized for efficiency on personal computers and enterprise-grade GPUs: Stable Diffusion 3 has models that can reach eight billion parameters, for example, so Stability AI aims to use the average version size as standard for text. -image conversion.

The company communicates that the consumption of VRAM has been reduced, which improves the tool’s performance on more limited video cards. Furthermore, the developer had the collaboration of NVIDIA and OMG to optimize Stable Diffusion 3 on different devices such as RTX series graphics cards and AMD APUs.

Now available

Stability of artificial intelligence announced the Stable Diffusion 3 family of models in February of this year, but it has not yet been possible to test it on different instruments. SD3 Medium is available for use via APIs from the Stability Platform or through the Stable Assistant and Stable Artisan paid services.

The company is also responsible for other generative artificial intelligence models, such as Stable audiocapable of creating songs up to three minutes long from text instructions.

Trends on Canaltech:

• The JBL box survives after 25 days under the mud in RS
• The patient uses a clear plastic implant that exposes his brain
• Tatá Werneck, Dani Calabresa and others | Inside Out 2 voice actors
• 8 AI to create animations
• The 50 funniest Google Assistant jokes
• Brazil x United States | Like watching the last Seleção friendly before the Copa América

Source: Terra

Rose

Rose James is a Gossipify movie and series reviewer known for her in-depth analysis and unique perspective on the latest releases. With a background in film studies, she provides engaging and informative reviews, and keeps readers up to date with industry trends and emerging talents.