FlowHMR open-sources generative mocap from monocular video; LLM agent automates full 3D pipeline
Scobleizer · x · 2026-10-26
The official FlowHMR implementation is now on GitHub (Apache-2.0, 187 stars): a generative model that produces physically plausible motion capture from monocular video.
The authors demoed a fully automated pipeline: an LLM agent takes the prompt "Create a simple 3D animation from this video," builds the scene, calls FlowHMR for mocap, then handles retargeting, motion fixes, camera setup and rendering — with MiniMax-H3 turning the 3D render into a photorealistic video.
- flowhmrbase: 0.46B params, 78.79% tracking success
- Repo will progressively release the full workflow: inference code + checkpoints, pretraining and GRPO post-training code, web demo, synthetic data generation, and a Wild-4K eval clip list
- A Hugging Face demo is available
More from Multimodal
- Dev argues players reject AI slop, not AI-made games — and shows a polished example — ezshine · 2026-10-26
- Real-time fluid sim demo ships 9 looks and 11 physics presets with GPU-baked collision — Michael_Moroz_ · 2026-10-26
- AI-generated frog romance short with a bee punchline delights Reddit — HarshaShastry · 2026-10-26
- Adobe's SliderEdit++ Improves Slider-Based Image Editing by 18.5% with New SlideBench — Arman Zarei · 2026-10-26
- 'La dolce vita': egeberkina shares retro Italian-summer AI image set — egeberkina · 2026-10-26
- AI short film teaser 'Kaalakund' shows near-photoreal moody landscapes — KalpnkAIStudio · 2026-10-26