Yu-Cheng Chou

Try our Reference-to-Video model! Grok it!🚀

Congratulations, @jieneng_chen!

@Aryakris_20 Done! Thanks!

@johnson111788 Hi, could you please enable your DMs!

@JackAMAustin You’re totally correct!Still long way to go

@johnson111788 Flying through a consistent generated world is a genuinely wild thing to be possible right now

CVPR 2026🎥We built a model that lets you fly through a generated video world. Not just generating frames — but maintaining a consistent 3D world under complex camera motion. Code, ckpt, and even the data pipeline are all open-sourced ↓ #AI #worldmodel #videogen #cvpr #drone

@mervenoyann Thank you so much! I’ll contact you if needed!

@johnson111788 Hello! I saw you're open sourcing your weights, if you want to build a demo we'd love to provide ZeroGPU H200 for it 🤗

@carbon787777 Thanks! Got really nice mentors🥹

@johnson111788 nice view, congrats on the CVPR acceptance.

@wildpinesai @Scobleizer Exactly!

@johnson111788 @Scobleizer basically a game engine that dreams instead of renders. this is how you get embodied AI training environments without hand-building every world

@somebobcat8327 Glad you like it🤩

@johnson111788 Wow. Beautiful

@DaveSMT @Scobleizer Thanks!

@johnson111788 @Scobleizer That’s cool 😎

Deep gratitude to our advisors @never1andd, @YuilleAlan, and @cihangxie for their guidance and support throughout this work. Thank you! 🙏

🔗 Code: github.com/johnson111788/… 🔗 Project: johnson111788.github.io/open-safari/ If you’re working on video / world models — would love to connect.

Importantly, we open-source the entire curation pipeline: • trajectory reconstruction (SfM / hloc) • geometric verification • motion consistency checks • automatic repair & filtering → You can build your own OpenSafari.

We also built OpenSafari, a dataset designed to break existing models: • in-the-wild FPV drone videos • large-scale camera motion • strong parallax & elevation changes Every trajectory is geometrically verified.

The difference is most visible under aggressive motion: • sharp turns • large parallax • long trajectories Baselines collapse, ours stays consistent.

Our idea: build a world memory. At each frame, the model retrieves 3D-consistent information conditioned on the camera pose. → This enables stable generation under 6-DoF motion → Even for long trajectories in complex outdoor scenes

But in reality, most video generation models today: ❌ Only work on narrow domains (e.g., real estate scenes) ❌ Break as soon as the camera moves ❌ Fail to follow the camera trajectory The core issue? → No persistent world representation

@CVPR Can we put the title in the main text to save space?

Before you hit submit: Check if your paper title is included. It must be there to comply with the #CVPR2026 rebuttal template. 🔍

Join us at #ICCV2025 for the 1st Embodied Spatial Reasoning Workshop! We're thrilled to host amazing speakers from industry and academia, featuring Sifei Liu, @xiaolonw, @xf1280, and @kate_saenko_, to discuss frontiers of spatial reasoning, embodied agents, and robotics! 🔗 tinyurl.com/yn7b6mu6

Deep gratitude to our advisors @cihangxie and @never1andd for their guidance and support throughout this work. Thank you! 🙏

Well-Formed, Ill-Grounded. We re-check GPT-OSS (InternVL3.5-aligned) vs Qwen on MMMU. Verdict: visual alignment ≫ formatting. Qwen wins with longer, grounded chains; OSS looks tidy but often doesn’t see the image. Blog + data ↓ #LLM #GPT #Qwen #InternVL irradiated-lace-40e.notion.site/Well-Formed-Il…