Sida Peng retweeted
The InfiniSplat team have released v3 as SplatAnything. Just 22 days between the v2 and v3 papers 🤯
V2 could already turn overlapping photos into a Gaussian splat without training each scene from scratch. In v3, the team uses the spacing between splats to help decide how big they should be. Too small and you see gaps, while too much overlap softens the detail.
V3 is designed to get that balance right whether you’re scanning an apple on a desk or a whole city block.
Thrilled to release InfiniSplat v3 (SplatAnything)!
We teamed up with @TencentHunyuan to scale up InfiniSplat v2’s training data & introduced Gaussian Spacing Normalization for high-quality rendering across all scales, from object to aerial.
Code is now open-source👇
Sida Peng retweeted
A tabletop. A room. A city.
SplatAnything reconstructs them with the same feed-forward 3D model. The key is normalizing Gaussian spacing, rather than scene size alone.
Here's the idea in 3 minutes.
#3DGS #ComputerVision
Thrilled to release InfiniSplat v3 (SplatAnything)!
We teamed up with @TencentHunyuan to scale up InfiniSplat v2’s training data & introduced Gaussian Spacing Normalization for high-quality rendering across all scales, from object to aerial.
Code is now open-source👇
Code: github.com/zju3dv/InfiniSpla…
Project page: zju3dv.github.io/SplatAnythi…
Paper: arxiv.org/abs/2611.08264
Demo: huggingface.co/spaces/ritian…
Results are actually much better than earlier versions.
Replying to @Almorgand
Project page: zju3dv.github.io/SplatAnythi…
Paper: arxiv.org/abs/2611.08264
Code: github.com/zju3dv/SplatAnyth…
Model: huggingface.co/ritianyu/Spla…
Demo: huggingface.co/spaces/ritian…
SplatAnything
Gaussian Spacing Normalization for Feed-forward 3DGS across Arbitrary Scenes
paper: huggingface.co/papers/2611.0…
Sida Peng retweeted
Excited to share SplatAnything!
From everyday objects and human subjects to expansive aerial scenes, SplatAnything transforms simple captures into 3D Gaussian scenes in few seconds. Give it a Try!
🌐 Project Page: zju3dv.github.io/SplatAnythi…
🤗 Huggingface: huggingface.co/spaces/ritian…
Sida Peng retweeted
SplatAnything
Gaussian Spacing Normalization for feed-forward 3DGS across arbitrary scenes.
From objects and people to aerial views, this method reconstructs 3D Gaussian scenes from simple captures with state-of-the-art quality.
Sida Peng retweeted
开源AI游戏工作室OhMyGame大升级!🔥🔥
升级到Beta 5版本后 ->>
1. OhMyGame 现在可以直接用Claude Code、Codex CLI 来制作游戏了,Bring your own agent!
2. 新增加的各种游戏平台插件,让用户可以方便地把制作的游戏发布到Taptap 等游戏平台上
3. 新增加了对互动游戏的进一步支持。 OhMyGame 是第一个直接把网页作为画布节点的游戏工具。每一个节点都是可以自由控制的网页。非常适合创作各类型的交互游戏和叙事游戏,比如影游和各类型的交互故事。
告诉我你对后续开发的建议!
#aigame #vibegame #gamedev #indiegame
Thanks to a tip from the Discord community, we found InfiniSplat v2 is great at panorama-to-3DGS conversion.
The feature is now live on the HF Space: huggingface.co/spaces/Edison…
And the full workflow is documented on GitHub: github.com/zju3dv/InfiniSpla…
Sida Peng retweeted
10 days after open-sourcing 🎮 OhMyGame: 936 stars, 11 releases, 131 merged PRs. 🔥🚀
Small team, no ads.
🤔 Biggest surprise: X sent more visitors to our repo than GitHub, Google, our site and V2EX combined.
InfiniSplat v2: our follow-up to InfiniSplat is out.
🤗 Multi-view images in, clean high-fidelity Gaussians out — unbounded free-viewpoint rendering.
👉 Try it now: huggingface.co/spaces/Edison…
💻 Code: github.com/zju3dv/InfiniSpla…
Sida Peng retweeted
Turning a set of photos into a Gaussian Splat—without optimizing every scene from scratch.
InfiniSplat v2 works with posed or unposed images, meaning camera data is optional. It learns to fade out unreliable points, helping reduce floaters and artifacts, while producing more coherent results as the camera moves beyond the original views.
The team has made the code and weights public so others can test and build on their amazing work 😄
Sida Peng retweeted
Single video in, 4D human out.
4DAnyone turns a casual monocular video into a 4DGS model.
No camera rig, no calibration, no tripod.
Project: 4danyone.github.io/
On indoor scenes, our InfiniSplat delivers consistent and significant gains over the well-known SHARP.
Feel free to try it: huggingface.co/spaces/PLUS-W…
InfiniSplat (SIGGRAPH Asia 2026) is trending on @HuggingFace!
🚀 Turn a single image into a navigable 3DGS scene in <1 second.
🏠 Stunning indoor view synthesis quality.
👉 Try it now: huggingface.co/spaces/PLUS-W…
🌐 Project: zju3dv.github.io/InfiniSplat…
InfiniSplat (SIGGRAPH Asia 2026) is trending on @HuggingFace!
🚀 Turn a single image into a navigable 3DGS scene in <1 second.
🏠 Stunning indoor view synthesis quality.
👉 Try it now: huggingface.co/spaces/PLUS-W…
🌐 Project: zju3dv.github.io/InfiniSplat…
Sida Peng retweeted
Current robot policies overfit specific language templates, handling 'pick and place' but freezing on 'drag it to me ' or 'push it closer to me.' They also lack control over execution: which hand, what approach angle, where to grasp, which path to follow.
🤖 FineVLA make robots steerable : changing instruction alters execution; same task, different phrasing, distinct actions — all faithfully done.
🏠 Homepage: finevla.xlang.ai
📄 Paper: huggingface.co/papers/2605.2…
💻Codebase: github.com/xlang-ai/FineVLA
🧵[1/6]











