3D / Point Cloud
Objaverse
800K+ annotated 3D models for training 3D and multimodal models.
100K–1M·CC-BY·Point Cloud (LAS/PLY)
3D / Point Cloud
3EED
3EED: Ground Everything Everywhere in 3D — Dataset Card A cross-platform, multi-modal 3D visual grounding dataset spanning vehicle, drone, and…
3EED: Ground Everything Everywhere in 3D — Dataset Card A cross-platform, multi-modal 3D visual grounding dataset spanning vehicle, drone, and quadruped platforms, with synchronized RGB, LiDAR, and language annotations. This page documents how to obtain and organize the dataset from HuggingFace and how to connect it with the training/evaluation code in the 3EED repository. Project Page: Code (Baselines & Evaluation): Paper:… See the full description on the dataset page:
Source: Hugging Face Hub (RRRong/3EED). Metadata imported from the dataset’s Hub tags.