17 citations · 42 across the 27 of their papers we have counts for
8 papers · 1 filter
Where to Refine, When to Stop: Rethinking Redundancy via Latent Discrepancy for Efficient Visual Autoregressive Generation
Changwang Mei, Peisong Wang, Zekun Li +7
Visual Autoregressive (VAR) models deliver high-quality image generation but suffer from significant inference latency at high resolutions. Recent acceleration approaches most rely…
SparVAR: Exploring Sparsity in Visual AutoRegressive Modeling for Training-Free Acceleration
Zekun Li, Ning Wang, Tongxin Bai +4
Visual AutoRegressive (VAR) modeling has garnered significant attention for its innovative next-scale prediction paradigm. However, mainstream VAR paradigms attend to all tokens ac…
FireFlow: Fast Inversion of Rectified Flow for Image Semantic Editing
Yingying Deng, Xiangyu He, Changwang Mei +2
Though Rectified Flows (ReFlows) with distillation offers a promising way for fast sampling, its fast inversion transforms images back to structured noise for recovery and followin…
Soft Threshold Ternary Networks
Weixiang Xu, Xiangyu He, Tianli Zhao +3
Large neural networks are difficult to deploy on mobile devices because of intensive computation and storage. To alleviate it, we study ternarization, a balance between efficiency…
Q-ViT: Fully Differentiable Quantization for Vision Transformer
Zhexin Li, Tong Yang, Peisong Wang +1
In this paper, we propose a fully differentiable quantization method for vision transformer (ViT) named as Q-ViT, in which both of the quantization scales and bit-widths are learna…
Towards Mixed-Precision Quantization of Neural Networks via Constrained Optimization
Weihan Chen, Peisong Wang, Jian Cheng
Quantization is a widely used technique to compress and accelerate deep neural networks. However, conventional quantization methods use the same bit-width for all (or most of) the…
Improving Binary Neural Networks through Fully Utilizing Latent Weights
Weixiang Xu, Qiang Chen, Xiangyu He +2
Binary Neural Networks (BNNs) rely on a real-valued auxiliary variable W to help binary training. However, pioneering binary works only use W to accumulate gradient updates during…
Generative Zero-shot Network Quantization
Xiangyu He, Qinghao Hu, Peisong Wang +1
Convolutional neural networks are able to learn realistic image priors from numerous training samples in low-level image generation and restoration. We show that, for high-level im…