3D / Point Cloud
Objaverse
800K+ annotated 3D models for training 3D and multimodal models.
100K–1M·CC-BY·Point Cloud (LAS/PLY)
Warehouse Scene 1 ( oavfos7g) SiRLab publishes LiDAR-first synthetic datasets with dense per-point semantics, synchronized RGB, and ego-motion, built to plug into modern perception pipelines. Best for: point-cloud semantic segmentation, filtering/analytics, fusion prototyping. Commercialization: you may commercialize models/outputs trained on the Full dataset (dataset redistribution is prohibited; see License below). Full access / quote / integration: see Commercial access &… See the full description on the dataset page: oavfos7g-preview.
Source: Hugging Face Hub (sirlab-ai/robotics-warehouse-daylight__oavfos7g-preview). Metadata imported from the dataset’s Hub tags.