Computer Vision

⚙️ StyleVAR: Controllable Image Style Transfer via Visual Autoregressive Modeling featured image

⚙️ StyleVAR: Controllable Image Style Transfer via Visual Autoregressive Modeling

A deep dive into our research on StyleVAR, a framework that formulates style transfer as conditional discrete sequence modeling to balance content structure and artistic texture.

avatar
Liqi Jing
Enhancing VLM Grounding: A Pipeline for Fine-Grained Data Synthesis and Reinforcement Learning featured image

Enhancing VLM Grounding: A Pipeline for Fine-Grained Data Synthesis and Reinforcement Learning

A deep dive into our research on improving Vision-Language Model spatial reasoning through a two-level data distillation pipeline and a customized Chain-of-Thought reward …

avatar
Liqi Jing