Hi friends! Attending ICLR 2025 in Singapore? Interested in visiting the NUS campus NUS? 🎉 Join us for the Multimodal Gathering Workshop @ NUS! 🙌 A half‑day meetup for researchers and students to exchange ideas, showcase work, and explore the latest in multimodal AI.

Our previous work, FAR, proposes a next-frame prediction paradigm based on long short-term context modeling with asymmetric patchification. Paper: arxiv.org/abs/2503.19325 Code: github.com/showlab/FAR Glad to see this idea adopted and extended in FramePack, demonstrating

Insightful keynote from Violet Peng at SSNLP about creativity and control — To fulfill a meaningful goal in a constrained setting, one naturally has to be innovative! can’t agree more

LiveCC just dropped on Hugging Face Learning Video LLM with Streaming Speech Transcription at Scale video LLM capable of real-time commentary, trained with a novel video-ASR streaming method, SOTA on both streaming and offline benchmarks.

🎉 We’re excited to host two challenges at LOVE: Multimodal Video Agent Workshop at CVPR 2025, advancing the frontier of video-language understanding! #CVPR2025 #CVPR2025 📌 Track 1A: [VDC] Video Detailed Captioning Challenge Generate rich and structured captions that cover multiple

<Impossible Videos> is accepted by ICML 2025. Congrats to the team! 🎉🥳

Happening now: Keynote by Mike Shou #CVPR2025

Customizing Your LLMs in seconds using prompts🥳! Excited to share our latest work with HPC-AI Lab, VITA Group, Konstantin Schürholt, Yang You, Michael Bronstein, Damian Borth : Drag-and-Drop LLMs(DnD). 2 features: tuning-free, comparable or even better than full-shot tuning.(🧵1/8)

🚀A new way to use diffusion models for style transfer! Style Matching Score (SMS) is accepted to #ICCV2025🌺 We reframe image stylization as a style distribution matching problem. -Paper: arxiv.org/abs/2503.07601 -Code: github.com/showlab/SMS -Project: yuxinn-j.github.io/projects/SMS.h…

Movies are more than just video clips, they are stories! 🎬 We’re hosting the 1st SLoMO Workshop at #ICCV2025 to discuss Story-Level Movie Understanding & Audio Descriptions! Website: slomo-workshop.github.io Competition: huggingface.co/spaces/SLoMO-W…

Glad to see GPT-5’s dynamic router balances fast answers and deep reasoning based on query complexity and user intent. In our recent work “Think-or-Not”: arxiv.org/pdf/2505.16854 We study this adaptive reasoning and introduce an easy-to-follow “thought-dropout” mechanism that

wow very interesting finding!

Introducing MCPMark, a collaboration with Eval Sys and LobeHub! We created a challenging benchmark to stress-test MCP use in comprehensive contexts. - 127 high-quality data samples created by experts. - GPT-5 takes the current lead and achieves a Pass@1 of 46.96% while the

Attended Gemini event earlier this week and always enjoyed the talks from legendary google pioneers👍 One fun fact learned is, today, many ML people work on NLP, many NLPer work on vision-language, many vision people work on robot 😆

Academia invents the sparks e.g. Attention, GANs, DMs. Industry grows them into societal impacts s.a. AlphaFold, LLMs, GenAI. If you want to explore 1,000 ideas, go to academia -- if you want to scale-up a few promising ideas, go to industry.

Big thanks to Zhenheng and my advisor Mike Shou for their support🙏Our unified multimodal model “Show-o2” got accepted to NeurIPS 2025! See you in San Diego 👋 #NeurIPS2025 Code is open-sourced & maintained here: github.com/showlab/Show-o

As an SAC for #NeurIPS2025, I don't agree with PCs approach to reject papers based on ranking. I ranked papers as requested and explicitly stated that I support the acceptance of all papers. I wasn't given an explanation of why papers at the end of the rank were rejected.