CLARU
JackLiu0406apache-2.0

DA3-XVLA Training & Eval Pipeline

A custom Vision Language Action architecture combining Florence-2 VLM with frozen Depth Anything 3 backbone, Perceiver resampler, and DiT-style flow-matching action decoder, with full training and evaluation pipelines for RoboPRO/Roboreal and RoboTwin 2.0 tasks.

Downloads39

Technical Profile

Modalities
rgbdepth
Environment
simulation
Task Types
manipulationpick_and_place
Data Format
HDF5
License
apache-2.0
Part of the DA3-XVLA Training & Eval Pipeline family

Access

View on HuggingFace

Need custom rgb data?

Claru builds purpose-built datasets for simulation applications with dense human annotations and quality assurance.

Request a Sample Pack

Related Datasets

OmniAction

A large-scale multimodal dataset for proactive robot manipulation with 141,162 episodes covering contextual instruction following through spoken dialogue, environmental sounds, and visual cues. The dataset includes 5,096 distinct speaker timbres, 2,482 non-verbal sound events, and 640 environmental backgrounds across six categories of contextual instructions.

rgbaudiolanguage
104K downloadsApr 2026cc-by-nc-4.0

OmniAction

A large-scale multimodal dataset for proactive robot manipulation comprising 141,162 episodes with cross-modal contextual instructions derived from spoken dialogue, environmental sounds, and visual cues rather than explicit commands.

rgbaudiolanguage
98K downloadsMar 2026cc-by-nc-4.0

Xperience-10M

A large-scale egocentric multimodal dataset of human experience containing 10 million interactions and 10,000 hours of synchronized first-person recordings with six video streams, audio, stereo depth, camera pose, hand mocap, full-body mocap, IMU, and hierarchical language annotations for embodied AI, robotics, and world modeling research.

rgbaudiodepthproprioception+3
89K downloadsApr 2026other

ABC-130k

The largest open-source robot teleoperation dataset containing bimanual manipulation trajectories collected on two-arm YAM stations with 130,822 episodes across 3,555 hours of data.

rgbproprioception
87K downloadsJun 2026apache-2.0

OmniAction

A large-scale multimodal dataset for proactive robot manipulation with 141,162 episodes covering contextual instruction following through spoken dialogue, environmental sounds, and visual cues.

rgbaudiolanguage
86K downloadsApr 2026cc-by-nc-4.0

Hy-Embodied-0.5-VLA-Data

A large-scale bimanual manipulation dataset with 2,163 hours of high-fidelity demonstrations collected via custom fingertip UMI device with optical motion-capture, spanning 70+ manipulation tasks for training Vision-Language-Action foundation models.

rgbproprioceptionlanguage
76K downloadsJun 2026cc-by-4.0