5 papers · 1 filter
Think Locally, Refine Globally for Memory-Efficient 3D Reconstruction
Jingke Zhou, Chenhang Ma, Zhizhou Zhong +6
We propose LoG-VGGT, a memory-efficient framework for long-sequence 3D reconstruction that balances local temporal modeling with global camera consistency. Instead of relying on fu…
Forcing-KV: Hybrid KV Cache Compression for Efficient Autoregressive Video Diffusion Models
Yicheng Ji, Zhizhou Zhong, Jun Zhang +7
Autoregressive (AR) video diffusion models adopt a streaming generation framework, enabling long-horizon video generation with real-time responsiveness, as exemplified by the Self…
ParallelVLM: Lossless Video-LLM Acceleration with Visual Alignment Aware Parallel Speculative Decoding
Quan Kong, Yuhao Shen, Yicheng Ji +2
Although current Video-LLMs achieve impressive performance in video understanding tasks, their autoregressive decoding efficiency remains constrained by the massive number of video…
AnyTalker: Scaling Multi-Person Talking Video Generation with Interactivity Refinement
Zhizhou Zhong, Yicheng Ji, Zhe Kong +12
Recently, multi-person video generation has started to gain prominence. While a few preliminary works have explored audio-driven multi-person talking video generation, they often f…
SpecVLM: Enhancing Speculative Decoding of Video LLMs via Verifier-Guided Token Pruning
Yicheng Ji, Jun Zhang, Heming Xia +4
Video large language models (Vid-LLMs) have shown strong capabilities in understanding video content. However, their reliance on dense video token representations introduces substa…