GR00T N1
NVIDIA
WorldBench editorial pick
Open foundation model for generalist humanoid robots, focused on whole-body and manipulation behavior from multimodal robot data.
VLAActionsObservation
featured · 338 downloadsView record
Choose robot data with evidence.
Every robot behavior begins with a data decision. Trace the source, judge the fit, and inspect the evidence before the next run.
Worth a closer look.
Selected for source evidence, usability, adoption, freshness, and distinct value.
3 selected from 355
Datasets + models
Open
Xperience-10M
Hugging Face · ropedia-ai/xperience-10m
A large egocentric multimodal dataset with synchronized video streams, audio, depth, poses, mocap, IMU, and hierarchical language annotations.
View record
SmolVLA
Hugging Face LeRobot
Compact open vision-language-action policy designed for practical robot fine-tuning and deployment through the LeRobot ecosystem.
View record
From source record to training decision.
Each step narrows uncertainty and leaves a result the team can review.
- 01
Trace the source
Open the official repository, paper, or project page.
- 02
Check the fit
Compare access, license, modalities, artifacts, and adoption.
- 03
Inspect locally
Audit the exact dataset revision with openbot-data.
- 04
Record the decision
Use, defer, repair, or reject—with the reason preserved.
openbot-data
Run the audit where the data lives.
Inspect a robot-video or LeRobot dataset locally. Nothing is uploaded, and every result stays reviewable before the next run.
Install openbot-datapip install openbot-data
openbot-data audit ./lerobot_dataset \
--format lerobot \
--integrity full \
--out ./audit.json \
--fail-on error
openbot-data readiness ./lerobot_dataset \
--profile lerobot-act \
--integrity full \
--out ./readiness.json- Local
- Dataset stays on your machine
- Repeatable
- Deterministic findings and diffs
- Reviewable
- Evidence before training decisions