Seedream 5.0 Pro
ByteDance Seed, 2026Seedream 5.0 Pro supports precise interactive editing: beyond generation, it is built for design and productivity.
Xinyu Huang (黄新宇)
I am a researcher at ByteDance Seed, working on multimodal generation.
I obtained my Ph.D. from Fudan University, advised by Prof. Rui Feng. I was also a visiting student at MMLab, Nanyang Technological University, advised by Prof. Ziwei Liu.
From 2020 to 2024, I worked on visual perception and created the Recognize Anything Model (RAM) family, a series of open-source image perception models that exceed OpenAI's CLIP by more than 20 points in fine-grained perception. This work was done at OPPO Research Institute and IDEA, together with Youcai Zhang, Prof. Yandong Guo, and Prof. Lei Zhang.
From 2024 to 2025, I co-led the development (pre-training, SFT, and RL) of large multimodal models at TikTok AI Innovation Center.
Research & Work
ByteDance Seed
Paint-Anything
Paint-Anything: Unified Any-Color Control for Image Generation and Editing
NeurIPS 2026Paint-Anything enables any-color control with hex values for both image generation and editing.
Ph.D. at Fudan University
Multi-Turn Grounding-Based RL (MGPO)
High-Resolution Visual Reasoning via Multi-Turn Grounding-Based Reinforcement Learning
ACL 2026MGPO lets LMMs iteratively focus on key image regions through automatic grounding, achieving superior performance on high-resolution visual tasks without any grounding annotations.
Recognize Anything Plus Model (RAM++)
Open-Set Image Tagging with Multi-Grained Text Supervision
ACM MM 2025RAM++ is the next generation of RAM, which recognizes any category with high accuracy, covering both predefined common categories and diverse open-set categories.
Recognize Anything Model (RAM)
Recognize Anything: A Strong Image Tagging Model
CVPR 2024 Workshop on Multimodal Foundation Modelsproject page / arXiv / demo / code
RAM is an image tagging model that recognizes any common category with high accuracy.
Tag2Text Vision-Language Model
Tag2Text: Guiding Vision-Language Model via Image Tagging
ICLR 2024project page / arXiv / demo / code
Tag2Text is a vision-language model guided by tagging, which supports tagging and comprehensive captioning simultaneously.
Simple and Robust Loss Design for Multi-Label Learning with Missing Labels
arXiv 2021Multi-label learning with missing labels (MLML) is a challenging problem. We propose two simple yet effective methods via robust loss design.
Open-Source Projects
Recognize Anything Family
The Recognize Anything Model (RAM) family, demonstrating superior image recognition ability across common and open-set categories.
RAM-Grounded-SAM
The RAM family married with Grounded-SAM, which can automatically recognize, detect, and segment anything in an image.
Academic Service
Reviewer for CVPR, ICCV, ECCV, ICLR, NeurIPS, IJCV, WACV, and ACM MM.
