Paella: Text to image FASTER than diffusion models | Paella paper explained

Paella creates images from text FASTER than diffusion models (but is arguably a diffusion model too)! It’s not autoregressive like the DALL-E transformers either. Watch this video to find out how.
► Sponsor: Creative Fabrica 👉 https://www.creativefabrica.com/spark/ref/3134459/

Check out our daily #MachineLearning Quiz Questions: https://www.youtube.com/c/AICoffeeBreak/community

📜 Rampas, Dominic, Pablo Pernias, Elea Zhong, and Marc Aubreville. "Fast Text-Conditional Discrete Denoising on Vector-Quantized Latent Spaces." arXiv preprint arXiv:2211.07292 (2022). https://arxiv.org/abs/2211.07292
💻 Paella code: https://github.com/dome272/Paella

Check out Dominic’s (first author) channel: 👉 @outlier4052 …

diffusion diffusion models explained image paper text