🏢 Jülich Supercomputing Centre
Scaling Image Tokenizers with Grouped Spherical Quantization
·7140 words·34 mins·
loading
·
loading
AI Generated
🤗 Daily Papers
Computer Vision
Image Generation
🏢 Jülich Supercomputing Centre
GSQ-GAN, a novel image tokenizer, achieves superior reconstruction quality with 16x downsampling using grouped spherical quantization, enabling efficient scaling for high-fidelity image generation.