High-quality image generation AI 'Stable Diffusion 3' announced, enabling high-precision depiction of specified characters and multiple subjects, which are difficult for image generation AI.
Stability AI announced its image generation AI, ' Stable Diffusion 3 ,' on Friday, February 23, 2024. Stable Diffusion 3 enables operations that were difficult with conventional image generation AI, such as 'naturally depicting specified text within the generated image' and 'high-resolution depiction of multiple subjects.'
Stable Diffusion 3 — Stability AI Japan
Have you checked out #stablediffusion3 that we introduced the other day? https://t.co/0I6DQuUzuQ
— Stability AI Japan (@StabilityAI_JP) February 25, 2024
In the thread below, we will introduce some images actually generated with #stablediffusion3 .
The image below was generated using Stable Diffusion 3 with the prompt, 'Cinematic photo of a red apple on a table in a classroom, on the blackboard are the words 'go big or go home' written in chalk.' Although there are differences in capitalization, the words 'GO BIG OR go HOME' are written on the blackboard as instructed. Another point is that they are depicted as 'letters written in chalk' as instructed.
The image generated by the prompt 'a painting of an astronaut riding a pig wearing a tutu holding a pink umbrella, on the ground next to the pig is a robin bird wearing a top hat, in the corner are the words' stable diffusion ' (a painting of an astronaut holding a pink umbrella riding a pig wearing a tutu, on the ground next to the pig there is a robin wearing a top hat, in the corner the phrase' stable diffusion ' is written)' is shown below. In addition to being able to depict 'astronaut', 'pig', and 'robin' as instructed, the string 'STABLE DIFFUSION' is written in the lower left of the image. However, although the prompt instructed 'stable diffusion' in lowercase, it is worrisome that it is depicted in uppercase in the generated image.
'Studio photograph closeup of a chameleon over a black background' looks like this. The area around the chameleon's face is clearly depicted, while the body is blurred, giving it the feel of a macro lens.
Stable Diffusion 3 is said to have multiple models with 8 million to 8 billion parameters. At the time of writing, it is in the preview stage, and you can register for the waitlist at the link below.
SD 3 Waitlist — Stability AI
https://stability.ai/stablediffusion3
Related Posts:
