MultimodalSensor / Time-seriesVideo
Open X-Embodiment
1M+ real-robot trajectories from 20+ institutions.
1M–10M·Apache-2.0·rosbag
Open
2,922 results
3D / Point Cloud
Objaverse
800K+ annotated 3D models for training 3D and multimodal models.
100K–1M·CC-BY·Point Cloud (LAS/PLY)
ImageMultimodalTabular
Core-VIIRS-Nighttime-Light
Major TOM Core VIIRS Nighttime Light Annual radiance composites from the Visible Infrared Imaging Radiometer Suite (VIIRS) Day/Night…
1M–10M·CC-BY·Parquet
MultimodalTabularText
Real Estate Listings (Synthetic) (Free Sample)
Real Estate Listings (Synthetic) (Free Sample)
1K–10K·CC-BY-NC·CSV
MultimodalSensor / Time-seriesTabular
Valencia Urban Mobility Open-Data Archive
Valencia Urban Mobility Open-Data Archive
1K–10K·CC-BY
Image
A multi-hazard, multi-sensor, and multi-task vision-language dataset for global-scale disaster asses
A multi-hazard, multi-sensor, and multi-task vision-language dataset for global-scale disaster assessment and response.
>1B·CC-BY-NC-SA
MultimodalTabularText
CityShiftBench Scale-122
CityShiftBench Scale-122
100K–1M·Custom / Research-only·CSV
ImageMultimodalText
Norwegian Traffic Accident Scene Images with Embeddings
Norwegian Traffic Accident Scene Images with Embeddings
1K–10K·Custom / Research-only·Images (folder)
MultimodalTabularText
transit-guest-images
Transit guest images Prebuilt guest artifacts for Transit, downloaded on demand at VM creation time. catalog.json and its…
<1K·Custom / Research-only·JSON
Multimodal
CarbonFluxBench
CarbonFluxBench: A Global Benchmark for Upscaling of Carbon Fluxes Using Zero-Shot Learning CarbonFluxBench comprises over 1.3 million daily…
1M–10M·MIT
Image
Earth Surface Water Dataset (Sentinel-2)
Earth Surface Water Dataset (Sentinel-2)
<1K·CC-BY·Images (folder)
Text
PANGAEA-2 VHR: PureForest, xView2, mpv4ger, SN3 Roads
PANGAEA-2 VHR: PureForest, xView2, mpv4ger, SN3 Roads
100K–1M·Custom / Research-only·Parquet
ImageMultimodalText
MGRS-200k
MGRS-200k Dataset MGRS-200k is the first multi-granularity remote sensing (RS) image-text dataset, introduced in the paper FarSLIP: Discovering…
100K–1M·WebDataset