Instruct-Particulate: Inferring Articulated Structures from Static 3D Meshes
MonaJalal_ · x · 2026-07-22
Researchers introduced Instruct-Particulate, a new model for inferring articulated structures from static 3D meshes. It significantly improves generalization to novel object categories and supports kinematic prompting.
To achieve this, the team scaled their training data 40× and redesigned the model to follow kinematic prompts. As a result, diverse, realistic, simulator-compatible articulated 3D assets can now be generated directly from real-world images.
Related event: Instruct-Particulate Infers Articulated Structures from Static 3D Meshes(2 posts)→
More from Multimodal
- H3 long-form video experiment: 7-hour render, int8, peaked at 192GB RAM — SIR_NVAX_A_LOT · 2026-09-11
- 3D ResNet Paper Crosses 3,000 Citations Eight Years After CVPR 2018 — HirokatuKataoka · 2026-09-11
- Non-coder builds full-featured Android ComfyUI client with ChatGPT, submits to Google Play — ComfierUI · 2026-09-11
- RunningHub open-sources H3Lightning, speeding up MiniMax H3 video generation 12x — 智东西 · 2026-09-11
- FastH3-Live hits 22fps: acceleration node benchmarks and the --vram-headroom trick — spartong945 · 2026-09-11
- Midjourney style code share: --sref 2912175708 — tisch_eins · 2026-09-11