Forrest Iandola
@forresti.bsky.social
AI Research Scientist at Meta | posts are my own
Introducing 𝗦𝗾𝘂𝗲𝗲𝘇𝗲𝗠𝗲: 𝗘𝗳𝗳𝗶𝗰𝗶𝗲𝗻𝘁 𝗚𝗮𝘂𝘀𝘀𝗶𝗮𝗻 𝗔𝘃𝗮𝘁𝗮𝗿𝘀 𝗳𝗼𝗿 𝗩𝗥!
tl;dr: We can run three Gaussian Splatting avatars at 72 frames per second. It runs locally on a Meta Quest 3 VR headset. Previously, this required an NVIDIA GPU.
arXiv: arxiv.org/abs/2412.15171
project page: forresti.github.io/squeezeme
tl;dr: We can run three Gaussian Splatting avatars at 72 frames per second. It runs locally on a Meta Quest 3 VR headset. Previously, this required an NVIDIA GPU.
arXiv: arxiv.org/abs/2412.15171
project page: forresti.github.io/squeezeme
December 21, 2024 at 1:18 AM
Introducing 𝗦𝗾𝘂𝗲𝗲𝘇𝗲𝗠𝗲: 𝗘𝗳𝗳𝗶𝗰𝗶𝗲𝗻𝘁 𝗚𝗮𝘂𝘀𝘀𝗶𝗮𝗻 𝗔𝘃𝗮𝘁𝗮𝗿𝘀 𝗳𝗼𝗿 𝗩𝗥!
tl;dr: We can run three Gaussian Splatting avatars at 72 frames per second. It runs locally on a Meta Quest 3 VR headset. Previously, this required an NVIDIA GPU.
arXiv: arxiv.org/abs/2412.15171
project page: forresti.github.io/squeezeme
tl;dr: We can run three Gaussian Splatting avatars at 72 frames per second. It runs locally on a Meta Quest 3 VR headset. Previously, this required an NVIDIA GPU.
arXiv: arxiv.org/abs/2412.15171
project page: forresti.github.io/squeezeme
𝗘𝗳𝗳𝗶𝗰𝗶𝗲𝗻𝘁 𝗧𝗿𝗮𝗰𝗸 𝗔𝗻𝘆𝘁𝗵𝗶𝗻𝗴 from Meta: interactive video segmentation and tracking on an iPhone!
Paper: arxiv.org/pdf/2411.18933
Project page w/ demo link: yformer.github.io/efficient-tr...
Paper: arxiv.org/pdf/2411.18933
Project page w/ demo link: yformer.github.io/efficient-tr...
December 17, 2024 at 9:00 PM
𝗘𝗳𝗳𝗶𝗰𝗶𝗲𝗻𝘁 𝗧𝗿𝗮𝗰𝗸 𝗔𝗻𝘆𝘁𝗵𝗶𝗻𝗴 from Meta: interactive video segmentation and tracking on an iPhone!
Paper: arxiv.org/pdf/2411.18933
Project page w/ demo link: yformer.github.io/efficient-tr...
Paper: arxiv.org/pdf/2411.18933
Project page w/ demo link: yformer.github.io/efficient-tr...