Junru Lin

CS undergrad @UofT.

[Junru Lin @_Linjunru]

Very cool! Congrats @HaojunQiu !

[Haojun Qiu @HaojunQiu]

📢📢📢We introduce Efficient-SID⚡️: training-free single-image diffusion model that generates images by sampling directly from an input image's patch distribution. Our method enables megapixel generation in <1s and scales to gigapixel generation. We also enable stylization, editing, and other applications. The outputs are constrained to follow exactly the patch distribution of the input — something that is very difficult to do with large models!

#CVPR2026 Highlight

🌐 https://t.co/HyPtTbmfvS
📄 https://t.co/caXEAk2wEQ

[1/6]

[Hansheng Chen @HanshengCh]

Excited to announce a new track of accelerating Generative AI:

pi-Flow: Policy-Based Few-Step Generation via Imitation Distillation
https://t.co/6ro55E1XGP

Distill 20B flow models now using just an L2 loss via imitation learning for SOTA diversity and teacher-aligned quality.

[Jia-Bin Huang @jbhuang0604]

Woohoo! Imagine, Verify, Execute (IVE) is accepted to CoRL 2025! 🎉

Congrats to the incredible @umdcs students Seungjae Lee @JayLEE_0301, Daniel Ekpo (@daniekpo7), Haowen Liu!

[Roman Bachmann @roman__bachmann]

We will present FlexTok at #ICML2025 on Tuesday! Drop by to chat with @JRAllardice and me if you're interested in tokenization, flexible ways to encode images, and generative modeling.

📆 Tue, Jul 15, 16:30 PDT
📍 East Exhibition Hall, Poster E-3010
🌐 https://t.co/17oJKymhPl

[Yunqi (Richard) Gu @richard_yunqigu]

Which multimodal LLM should you be using to edit graphics in Blender?

Today, we’re releasing our #CVPR2025 Highlight🌟 work, #BlenderGym 🏋️‍♀️, the first agentic 3D graphics editing benchmark that will tell you exactly how multimodal LLMs compare in their Blender-editing skills.

What'd we find? 🧵👇

[Hansheng Chen @HanshengCh]

Excited to share our work:
Gaussian Mixture Flow Matching Models (GMFlow)
https://t.co/XWAy2VCJlg
GMFlow generalizes diffusion models by predicting Gaussian mixture denoising distributions, enabling precise few-step sampling and high-quality generation.

[Roman Bachmann @roman__bachmann]

Happy to share that we released FlexTok code and models on https://t.co/2pRCUcMbQn.
Try them with our interactive @huggingface demo on https://t.co/WoIB04rKj2

[Ian Huang @IanHuang3D]

🏡Building realistic 3D scenes just got smarter!

Introducing our #CVPR2025 work, 🔥FirePlace, a framework that enables Multimodal LLMs to automatically generate realistic and geometrically valid placements for objects into complex 3D scenes.

How does it work?🧵👇

[Google DeepMind @GoogleDeepMind]

Meet Gemini Robotics: our latest AI models designed for a new generation of helpful robots. 🤖

Based on Gemini 2.0, they bring capabilities such as better reasoning, interactivity, dexterity and generalization into the physical world. 🧵 https://t.co/n230QbZpnd

[Congyue Deng @CongyueD]

In the past, we extended the convolution operator to go from low-level image processing to high-level visual reasoning. Can we also extend physical operators for more high-level physical reasoning?

Introducing the Denoising Hamiltonian Network (DHN): https://t.co/GY76QreRge

[Koichi Namekata @Koichi_N_]

Thrilled to announce that SG-I2V has been accepted at #ICLR2025 ! Huge thanks to the collaborators, reviewers, and ACs. Looking forward to presenting this in Singapore!

[U of T Department of Computer Science @UofTCompSci]

Congratulations to @UofTCompSci undergrads Helen Li, Junru Lin, Leo Tenenbaum and Sarah Walker who have received honourable mentions in the @CRAtweets 2024-2025 Outstanding Undergraduate Researcher Award program!  https://t.co/5jorFZ3Ze2

[Jiaman Li @jiaman01]

🔥 Introducing MVLift: Generate realistic 3D motion without any 3D training data - just using 2D poses from monocular videos! Applicable to human motion, human-object interaction & animal motion. Joint work w/ @jiajunwu_cs & Karen

💡 How? We reformulate 3D motion estimation as generating consistent multi-view 2D pose sequences. Our framework uses 2D motion diffusion to progressively establish multi-view consistency, requiring only single-view 2D pose sequences for training.

Project: https://t.co/R8uTdpEw1w
Video with demonstration: https://t.co/QMu6goyTvi
Paper: https://t.co/sPEYM5WA6y

[Felix Taubner @taubnerfelix]

Introducing 🧢CAP4D🧢

CAP4D turns any number of reference images (single, few, and many) into controllable real-time 4D avatars. 🧵⬇️

Website: https://t.co/l6hRa5jquQ
Paper: https://t.co/fxGGu3X3cz

[Nakayama George @GeorgeNaka40190]

Do large multimodal models understand how to make dresses for your winter holiday party💃?

We introduce AIpparel, a vision-language-garment model capable of generating and editing simulation-ready sewing patterns from text and images. Project page at https://t.co/11j8Kqlbel.
[1/n]

[Yue Wang @yuewang314]

[Hiring!] I am hiring multiple PhDs @CSatUSC @USCViterbi for this cycle. If you're interested in scene representations, neural simulation, generative AI, and robotics, feel free to mention my name in your application (no need to email). For USC masters/undergrads who're interested in our research, feel free to fill in this form https://t.co/PgyUXhwddR.

[Prime (Shengqu) Cai @prime_cai]

Sharing something exciting we've been working on as a Thanksgiving gift: Diffusion Self-Distillation (DSD), which redefines zero-shot customized image generation using FLUX.
DSD is like DreamBooth, but zero-shot/training-free. It works across any input subject and desired context—character consistency, item/asset adaptation, scene relighting, and more.
It even enables the creation of comics/mangas without any effort in fine-tuning or training a personalized model!

📰 Paper: https://t.co/5NzUyw1uP8
🌐 Website: https://t.co/2KUP7EPCVV

Team effort with @ericryanchan, @zhang_yunzhi, @GuibasLeonidas, @jiajunwu_cs, and @GordonWetzstein.

[Sherwin Bahmani @sherwinbahmani]

📢 Excited to share our new work: AC3D: Analyzing and Improving 3D Camera Control in Video Diffusion Transformers

https://t.co/piUgl0MwjV

We analyze what pre-trained video diffusion transformers understand about 3D and demonstrate dynamic scene generation with 3D control.

[Igor Gilitschenski @igilitschenski]

I'm recruiting graduate students for Fall 2025 to work at the intersection of Computer Vision, Deep Learning, and Robotics.

If you are interested in building a controllable organic simulation engine and enabling safe robot learning, consider applying to UofT's CS PhD program 1/n

[Songyou Peng @songyoupeng]

Check out our new paper in feed-forward 3DGS model for large scenes! And the code is also available