Kai Han

Beyond Masks: Efficient, Flexible Diffusion Language Models via Deletion-Insertion Processes

PPE: Positional Preservation Embedding for Token Compression in Multimodal Large Language Models

Diff-SSR: Diffusion Model with Structure-Modulated for Image Super-Resolution

OmniEval: A Benchmark for Evaluating Omni-modal Models with Visual, Auditory, and Textual Inputs

Circle-RoPE: Cone-like Decoupled Rotary Positional Embedding for Vision-Language Models

An Empirical Study of Extremely Low-Bit Quantization for Large Language Models

Forest-of-Thought: Scaling Test-Time Compute for Enhancing LLM Reasoning

Mixture of Lookup Experts

SAN-Diff: Structure-aware noise for super-resolution diffusion model

LLM Data Selection and Utilization via Dynamic Bi-level Optimization

SpeCache: Speculative Key-Value Caching for Efficient Generation of LLMs

Empirical Study on Enhancing Efficiency in Masked Image Modeling Pre-training

Prompt-guided Visual Perception for Efficient Training-free Video LLM

Weak-to-Strong Enhanced Vision Model