Sign up to save tools and stay up to date with the latest in AI
bg
bg
1

What comes after Stable Diffusion? Stable Cascade could be Stability AI’s future text-to-image generative AI model

Feb 14, 2024 - venturebeat.com
Stability AI, the company behind the Stable Diffusion text-to-image generative AI technology, is previewing a new image generation model called Stable Cascade. The new model is designed to be more flexible and efficient than the current generation of Stable Diffusion models. Stable Cascade uses a different architecture to generate images and builds on the Würstchen architecture. It uses a pipeline of three distinct smaller models, which provides major advantages in training efficiency and customization.

In evaluations, Stable Cascade outperformed other leading AI art models including SDXL in terms of image quality and prompt alignment. It also has faster inference times despite having 1.4 billion more parameters than SDXL. The model also supports other capabilities including image variations and image-to-image translations. Stable Cascade is currently in research preview and available for non-commercial usage with a code available on GitHub.

Key takeaways:

  • Stability AI is previewing a new image generation model called Stable Cascade, which is designed to be more flexible and efficient than the current generation of Stable Diffusion models.
  • Stable Cascade uses a three-stage architecture, which provides major advantages in training efficiency and customization. It also has the potential for Direct Preference Optimization (DPO) to further improve image quality.
  • In Stability AI’s evaluations, Stable Cascade outperformed other leading AI art models including SDXL in terms of both image quality and prompt alignment. It also has faster inference times despite having 1.4 billion more parameters than SDXL.
  • Stable Cascade supports other capabilities including image variations and image-to-image translations. It is currently in research preview and available for non-commercial usage with a code available on GitHub.
View Full Article

Comments (0)

Be the first to comment!