Datasets:

---

HuggingFaceVLA

/

community_dataset_v3

like 23

Follow

Hugging Face Vision Language Action Models Research 103

Teleoperated imitation learning dataset for the SO-101 robotic arm. Demonstrates picking a piece of chocolate and placing it into a blue bowl, optimized for evaluation on SmolVLA and ACT policy architectures.

• Robot Hardware: SO-101 Arm (Single-arm setup)
• Tasks: Pick chocolate and place inside the blue bowl
• Format: LeRobotDataset v3 (Parquet states + sharded MP4 camera feeds)