Text-to-3D Scene Generation Introducing SceneWiz3D A new approach to create high-fidelity 3D scenes from text and 3D object control by Qihang Zhang and colleagues Link and examples below:

Thank AK for sharing our work. As an another plug-in module, CameraCtrl could be inserted into our #AnimateDiff to enable camera controlling. Code and models are avaliable at github.com/hehao13/Camera…

Huggingface demo is available at huggingface.co/spaces/qihang/… You can also find the code and weights in github.com/zqh0253/BerfSc…

🚀Excited to introduce KaleidoDiffusion -- a new method that improves conditional diffusion model generation by incorporating autoregressive latent priors! This allows us generate much more diverse outputs even with high CFG just like a kaleidoscope🔭! (1/n)

Scaling LAW also exists in diffusion transformers that tells how large model and how much data we need, given compute budget. arxiv: arxiv.org/pdf/2410.08184

🚀Excited to introduce our recent work @ AppleMLR -- DART: Denoising AutoRegressive Transformer for Scalable Text-to-Image Generation! A transformer-based model that unifies Autoregressive and Diffusion with a non-Markovian diffusion framework: 🔗 arxiv.org/abs/2410.08159 (1/n)

check out our recent work on T2I generation that unifies autoregressive and diffusion with non-Markovian assumption!

Life update: Excited to share that I will be joining Penn Computer and Information Science Penn Engineering as an Assistant Professor in Fall 2025!🤯 I’m also seeking multiple PhD students passionate about Generative Intelligence and leveraging it to empower AI agents to interact with the Physical World🌟

I was fortunate to be mentored by Jiatao this summer at Apple MLR, and my experience couldn't have been better. Jiatao provided frequent and insightful discussions, covering everything from hands-on experiments to broader research ideas. Definitely consider this opportunity!

🚀Thrilled to share our paper "DART" has been accepted by #ICLR2025! Congrats to my amazing collaborators Yuyang Wang Yizhe Zhang Qihang Zhang Dinghuai Zhang 张鼎怀 Navdeep Jaitly @jsusskin Shuangfei Zhai! Please also check the updated version with more results at arxiv.org/abs/2410.08159

Very cool results! The idea of learning additional motion space shares similar intuitions of our recent work for generating 3D-consistent videos by jointing predicting in XYZ space! arxiv.org/abs/2412.01821 Qihang Zhang Shuangfei Zhai @jsusskin

We are looking for a summer research intern to work on improving TarFlow at Apple. You will be working with myself and a great group of researchers, Jiatao Gu Preetum Nakkiran David Berthelot etc. If interested, send your CV to szhai at apple.com by this week.

We propose Long Context Tuning (LCT) for scene-level video generation to bridge the gap between current single-shot generation and real-world narrative video productions. Homepage: guoyww.github.io/projects/long-… Report: arxiv.org/abs/2503.10589

Glad to share Seaweed-7B, a cost-effective foundation model for video generation. Our tech report highlights the key designs that significantly improve compute efficiency and performance given limited resources, achieving comparable quality against other industry-level models. To

Please drop by and check our highlight poster tomorrow at #CVPR2025! ExHall D Poster #60 Sun 15 Jun 10:30 a.m. CDT — 12:30 p.m. CDT Great work by our Apple intern Qihang Zhang and look forward to more exploration on explicit 3D generation! zqh0253.github.io/wvd/