Tag
2 articles
Stability AI has released Stable Audio 3, a family of fast latent diffusion models for audio generation and editing. The open-weight models are optimized for both CPU and GPU environments, with strong performance on benchmark tests.
Google DeepMind introduces Unified Latents (UL), a new machine learning framework that improves latent space regularization using a diffusion prior and decoder, enhancing both efficiency and image quality in generative AI.