6 papers
Scaling Muon for Diffusion Transformers
Chenghao Li, Xiao Han, Xinxin Huang +22
The matrix-aware optimizer Muon improves large model training by balancing updates across singular directions, yet its scaling behavior and end-to-end efficiency on large Diffusion…
RedVTP: Training-Free Acceleration of Diffusion Vision-Language Models Inference via Masked Token-Guided Visual Token Pruning
Jingqi Xu, Jingxi Lu, Chenghao Li +3
Vision-Language Models (VLMs) have achieved remarkable progress in multimodal reasoning and generation, yet their high computational demands remain a major challenge. Diffusion Vis…
HIVTP: A Training-Free Method to Improve VLMs Efficiency via Hierarchical Visual Token Pruning Using Middle-Layer-Based Importance Score
Jingqi Xu, Jingxi Lu, Chenghao Li +2
Vision-Language Models (VLMs) have shown strong capabilities on diverse multimodal tasks. However, the large number of visual tokens output by the vision encoder severely hinders i…
FPAN: Mitigating Replication in Diffusion Models through the Fine-Grained Probabilistic Addition of Noise to Token Embeddings
Jingqi Xu, Chenghao Li, Yuke Zhang +1
Diffusion models have demonstrated remarkable potential in generating high-quality images. However, their tendency to replicate training data raises serious privacy concerns, parti…
LoyalDiffusion: A Diffusion Model Guarding Against Data Replication
Chenghao Li, Yuke Zhang, Dake Chen +2
Diffusion models have demonstrated significant potential in image generation. However, their ability to replicate training data presents a privacy risk, particularly when the train…
Mitigate Replication and Copying in Diffusion Models with Generalized Caption and Dual Fusion Enhancement
Chenghao Li, Dake Chen, Yuke Zhang +1
While diffusion models demonstrate a remarkable capability for generating high-quality images, their tendency to `replicate' training data raises privacy concerns. Although recent…