The dataset is currently empty. Upload or create new data files. Then, you will be able to explore them in the Dataset Viewer.
FusionX Multimodal Sample Dataset This repository contains a sample multimodal dataset including:
RGB camera video (30 Hz, MP4) Depth video (16-bit lossless MKV in millimeters) Tactile glove data (aligned per-frame in Parquet) Per-recording metadata
To eliminate API rate limits and make downloads significantly faster, the raw image sequences have been encoded into standard video containers. Each recording is stored as three highly optimized files:
recording_YYYYMMDD_HHMMSS/
├── frames.parquet # Per-frame aligned metadata + tactile arrays
├── rgb.mp4 # 30 Hz RGB camera video (H.264, lossy)
└── depth.mkv # 16-bit depth video in millimeters (FFV1, lossless)
How to Use
Download the repository via the Hugging Face UI or use huggingface_hub.snapshot_download.
Each recording_* directory represents a single synchronized capture session containing:
Vision data (RGB + depth videos)
Tactile glove data
Frame-level alignment metadata
License
Creative Commons Attribution 4.0 (CC BY 4.0)
You are free to:
Use
Share
Modify
Use commercially
As long as attribution is provided.
- Downloads last month
- 15