Runpeng Yu
@y30897
Ph.D. Student @NUSingapore. Working on Multimodal Large Language Model, Discrete Diffusion Model, and Neural Lineage.
Life Update: Thrilled to join @HongKongPolyU as Tenure-Track Assistant Professor in Data Science & AI! 🔍 Now hiring PhDs in the hottest topics of AI: 🔥 Generative AI👁️Computer Vision🤖 Agentic AI Just Apply→adamdad.github.io/opening #AIResearch #AI #GenAI #PhD #PolyU #HongKong
Discrete Diffusion in Large Language and Multimodal Models: A Survey just released on Hugging Face Get an overview of research in discrete diffusion LLMs and MLLMs, which achieve performance comparable to autoregressive models with up to 10x faster inference!
Huge thanks to @_akhaliq for sharing our work! 🙏✨ As methods like DUST3R and VGGT push feed-forward 3D reconstruction forward 🚀, we believe post-training techniques on them will become increasingly popular 📈. Test3R is just our first step—more to come! 🔧🔍
Test3R Learning to Reconstruct 3D at Test Time
Excited to introduce our new work! 🚀 We propose VeriThinker, a novel method for CoT compression. Code: github.com/czg1225/VeriTh… Paper: arxiv.org/pdf/2505.17941 Model: huggingface.co/Zigeng/R1-Veri…
Thanks @_akhaliq for sharing our work. Dimple is a discrete diffusion model for visual understanding, offering performance on par with AR and sometimes even faster. Check out our paper and try the model. Paper: arxiv.org/abs/2505.16990 Code: github.com/yu-rp/Dimple
Dimple Discrete Diffusion Multimodal Large Language Model with Parallel Decoding