Fangxun Shu

SAIL-RL: Guiding MLLMs in When and How to Think via Dual-Reward Supervision

Towards Enhanced Image Generation via Multi-Modal Chain of Thought in Unified Generative Models

LLaVA-MoD: Making LLaVA Tiny via MoE-Knowledge Distillation

Streaming Video Question-Answering with In-context Video KV-Cache Retrieval

Autoregressive Pretraining with Mamba in Vision

SAG: Style-Aligned Generation via Language Model Collaboration

HyperLLaVA: Dynamic Visual and Language Expert Tuning for Multimodal Large Language Models