Log inSign up
Log inSign up
Runjia Li
62 posts
Runjia Li profile banner
@RunjiaLi

Runjia Li

@RunjiaLi
DPhil in Computer Vision, University of Oxford || Clarendon Scholar @Oxford_VGG @OxfordTVG || Experience with @Snap @AIatMeta
Noxus
runjiali-rl.github.io
Joined October 2022
218 Following
232 Followers
RepliesRepliesRepostsRepostsMediaMedia

Runjia Li’s posts

Log in or sign up for X

See what’s happening and join the conversation

Continue with phone
or
Log in with username or email
Terms·Privacy·Cookies·Accessibility·US TIDA·Ads Info·© 2026 X Corp.
  • @RunjiaLi
    Runjia Li
    @RunjiaLi
    Nov 11
    🎉 OnlyAct is accepted at #NeurIPS2026! Take the motion from one video, play it in any world 🎬 We turn a pretrained video encoder into an action encoder with zero action labels. 🌐 Page: runjiali-rl.github.io/onlyact 💻 Code: github.com/runjiali-rl/on… #WorldModels #Robotics 🧵👇
    00:00
    19
  • @RunjiaLi
    Runjia Li
    @RunjiaLi
    Oct 26
    World models that see only through the players' eyes forget what happens out of view. OverSeer adds a few fixed cameras and generates every view together, so the world stays consistent over time and across players. 📄 arxiv.org/abs/2610.28554 🌐 runjiali-rl.github.io/OverSeer/ 🧵👇
    00:00
    11
  • @RunjiaLi
    Runjia Li
    @RunjiaLi
    May 19
    Impressive!!
    @jianyuan_wang
    Jianyuan
    @jianyuan_wang
    May 19
    Introducing VGGT-Ω: scaling feed-forward reconstruction across static and dynamic scenes, and studying whether the learned geometric representations transfer beyond reconstruction.
    00:00
  • @RunjiaLi
    Runjia Li
    @RunjiaLi
    Mar 16
    🎉EgoEdit @Snapchat has been accepted to CVPR 2026! 🏆👻 We are bringing high-quality, real-time editing to egocentric videos. Our massive 100k video dataset and benchmark are ALREADY PUBLIC! 🔓🚀 🏠 Project Page: snap-research.github.io/EgoEdit/ 🤗 Dataset: huggingface.co/datasets/ligua…
    00:00
    @_akhaliq
    AK
    @_akhaliq
    Dec 9, 2025
    00:40
    EgoEdit Dataset, Real-Time Streaming Model, and Benchmark for Egocentric Video Editing
    5
  • @RunjiaLi
    Runjia Li
    @RunjiaLi
    Jun 26, 2025
    🎉 VMem is officially accepted to ICCV 2025! Excited to chat with everyone in Hawaii about making video generation consistent and interactive with our Surfel-Indexed View Memory 🏝️🎥 Also, huge thanks to my insanely helpful coauthors!
    @_akhaliq
    AK
    @_akhaliq
    Jun 24, 2025
    VMem Consistent Interactive Video Scene Generation with Surfel-Indexed View Memory
    00:00
    5