6 papers
RAE-AR: Taming Autoregressive Models with Representation Autoencoders
Hu Yu, Hang Xu, Jie Huang +4
The latent space of generative modeling is long dominated by the VAE encoder. The latents from the pretrained representation encoders (e.g., DINO, SigLIP, MAE) are previously consi…
MaskFocus: Focusing Policy Optimization on Critical Steps for Masked Image Generation
Guohui Zhang, Hu Yu, Xiaoxiao Ma +3
Reinforcement learning (RL) has demonstrated significant potential for post-training language models and autoregressive visual generative models, but adapting RL to masked generati…
Towards Better & Faster Autoregressive Image Generation: From the Perspective of Entropy
Xiaoxiao Ma, Feng Zhao, Pengyang Ling +6
In this work, we first revisit the sampling issues in current autoregressive (AR) image generation models and identify that image tokens, unlike text tokens, exhibit lower informat…
Group Critical-token Policy Optimization for Autoregressive Image Generation
Guohui Zhang, Hu Yu, Xiaoxiao Ma +6
Recent studies have extended Reinforcement Learning with Verifiable Rewards (RLVR) to autoregressive (AR) visual generation and achieved promising progress. However, existing metho…
FreePCA: Integrating Consistency Information across Long-short Frames in Training-free Long Video Generation via Principal Component Analysis
Jiangtong Tan, Hu Yu, Jie Huang +2
Long video generation involves generating extended videos using models trained on short videos, suffering from distribution shifts due to varying frame counts. It necessitates the…
Frequency Autoregressive Image Generation with Continuous Tokens
Hu Yu, Hao Luo, Hangjie Yuan +3
Autoregressive (AR) models for image generation typically adopt a two-stage paradigm of vector quantization and raster-scan ``next-token prediction", inspired by its great success…