Xiawu Zheng

Flow Caching for Autoregressive Video Generation

Mix-Ecom: Towards Mixed-Type E-Commerce Dialogues with Complex Domain Rules

HS-SFT: Hybrid Sparse Supervised Fine-tuning for Offline LLM KV Cache Eviction

Training-Free Multimodal Large Language Model Orchestration

Event-Anchored Frame Selection for Efficient Long-Video Understanding

RoPK: A Head-Level Key Cache Channel Pruning Method for Efficient Long-Context LLM Inference

FastVGGT: Fast Visual Geometry Transformer

Breaking the Bias: Quantifying the Attention of Industrial Anomaly Detection

FOLD: Fast Correct Speculative Decoding

Modality-Aware Quantization: Balancing Visual and Textual Fidelity in Multimodal Compression

Pseudo-Label Supervision in Unsupervised Industrial Anomaly Detection

SplitQuant: Efficient Low-Bit Quantization for Diffusion Transformers via In-Channel Dimension Splitting

Determining Layer-wise Sparsity for Large Language Models Through a Theoretical Perspective

BAME: Block-Aware Mask Evolution for Efficient N:M Sparse Training

Video-RAG: Visually-aligned Retrieval-Augmented Long Video Comprehension

VITA-1.5: Towards GPT-4o Level Real-Time Vision and Speech Interaction

Multimodal Quantitative Language for Generative Recommendation

Learning Interleaved Image-Text Comprehension in Vision-Language Large Models

Zooming from Context to Cue: Hierarchical Preference Optimization for Multi-Image MLLMs

polybasic Speculative Decoding Through a Theoretical Perspective

Feast Your Eyes: Mixture-of-Resolution Adaptation for Multimodal Large Language Models

Discovering Important Experts for Mixture-of-Experts Models Pruning Through a Theoretical Perspective

Benchmarking Abstract and Reasoning Abilities Through A Theoretical Perspective

Data Interpreter: An LLM Agent For Data Science

Dynamic Low-Rank Sparse Adaptation for Large Language Models

BAME: Block-Aware Mask Evolution for Efficient N:M Sparse Training

Low Rank Quantization Adaptation for Large Language Model

Automated Fine-Grained Mixture-of-Experts Quantization

Mitigating Gradient Interference for Efficient Sparse Fine-Tuning of Large Language Models

Extreme composite compression of large language models through joint optimization

Debiased Imbalanced Pseudo-Labeling for Generalized Category Discovery

Towards Efficient Automatic Self-Pruning of Large Language Models

Polybasic Speculative Decoding Under a Theoretical Perspective