8 papers
DraftAttention2: Fast Video Diffusion with Low-Resolution-Guided Mixed-Precision Attention
Rui Ding, Haopeng Li, Weize Ma +10
Video generation has broad applications in content creation and entertainment. Diffusion transformers have advanced the quality of generated videos, but attention over spatiotempor…
Chimera: Designing and Chinchilla-Scaling Hybrid Visual Diffusion Transformers
Chongjian Ge, Hanwen Jiang, Tianyu Wang +9
Visual generation increasingly requires high-resolution images, long videos, and multimodal context, making the quadratic cost of full attention prohibitive. We introduce Chimera,…
FastCar: Cache Attentive Replay for Fast Auto-Regressive Video Generation on the Edge
Xuan Shen, Weize Ma, Yufa Zhou +11
Auto-regressive (AR) models, initially successful in language generation, have recently shown promise in visual generation tasks due to their superior sampling efficiency. Unlike i…
DraftAttention: Fast Video Diffusion via Low-Resolution Attention Guidance
Xuan Shen, Chenxia Han, Yufa Zhou +7
Diffusion transformer-based video generation models (DiTs) have recently attracted widespread attention for their excellent generation quality. However, their computational cost re…
QuartDepth: Post-Training Quantization for Real-Time Depth Estimation on the Edge
Xuan Shen, Weize Ma, Jing Liu +9
Monocular Depth Estimation (MDE) has emerged as a pivotal task in computer vision, supporting numerous real-world applications. However, deploying accurate depth estimation models…
METAL: A Multi-Agent Framework for Chart Generation with Test-Time Scaling
Bingxuan Li, Yiwei Wang, Jiuxiang Gu +2
Chart generation aims to generate code to produce charts satisfying the desired visual properties, e.g., texts, layout, color, and type. It has great potential to empower the autom…
Efficient Reasoning with Hidden Thinking
Xuan Shen, Yizhou Wang, Yufa Zhou +4
Chain-of-Thought (CoT) reasoning has become a powerful framework for improving complex problem-solving capabilities in Multimodal Large Language Models (MLLMs). However, the verbos…
MegaSynth: Scaling Up 3D Scene Reconstruction with Synthesized Data
Hanwen Jiang, Zexiang Xu, Desai Xie +11
We propose scaling up 3D scene reconstruction by training with synthesized data. At the core of our work is MegaSynth, a procedurally generated 3D dataset comprising 700K scenes -…