Hi there, I'm Jinheng Xie. 👋

A third-year PhD student at Show Lab, National University of Singapore, working with Prof. Mike Shou. Prior to my PhD, I dedicated three years to exploring label-efficient learning for scene understanding, focusing on weakly-supervised object localization and semantic segmentation. In my first year of PhD journey, I delved into visual prompt learning and effective controllable image synthesis. Currently, I’m concentrating on unifying multimodal understanding and generation within a native unified multimodal model. I have trained two unified multimodal models, Show-o and Show-o2, with trainable parameters up to 7 billion and utilizing billion-scale datasets.