GEAR: Joint Training of Tokenizer and Generator for Faster Image Synthesis1. July 2026AI ModelsEnd-to-end training of the tokenizer and generator with dual codebook selection accelerates ImageNet convergence up to 10x compared to LlamaGen-REPA. Share on:
ARM: Autoregressive Model for Unified Image and Text Processing10. June 2026AI ModelsARM combines discrete visual tokens with a 7-billion-parameter model to solve image and text tasks uniformly as token predictions. Share on:
