# Open-source tool for detecting issues in robot-learning datasets

> Source: <https://discuss.huggingface.co/t/open-source-tool-for-detecting-issues-in-robot-learning-datasets/178602#post_6>
> Published: 2026-09-06 13:12:24+00:00

**Open-source tool for robot dataset issues**

I’ve been working on **Calibra**, an open-source toolkit for analyzing robot-learning datasets.

While working with demo data, I noticed a common issue: we often waste compute training on data that shouldn’t be there in the first place.

Datasets can include:

So I built Calibra to help catch these issues before training.

Would love to hear thoughts from anyone working with **LeRobot or imitation learning.**

GitHub: [GitHub - omertt27/Calibra: Dataset observability and coreset selection for robotics imitation learning · GitHub](https://github.com/omertt27/Calibra)

I would appreciate any feedbacks.

 
 
This is a useful problem. One thing I’d want to see is a clear split between corrupted data and unusual but valid demonstrations, since rare movements can still be valuable in real-world data. A few visual examples for each failure type would make the tool much easier to trust.

 
 
[Sattyam](https://discuss.huggingface.co/u/Sattyam)
5
 
The corrupted-versus-unusual-but-valid distinction raised upthread is the whole problem, and it has a measurable form: what is Calibra’s false-positive rate on a dataset you know is clean?

Concretely. Take a LeRobot dataset you are confident in, run the detector, and publish the per-check flag rate. Jittery motion and calibration drift will both flag on some genuinely good demonstrations, because a human demonstrator is jittery and a real rig does drift. That rate is not a bug to tune to zero; it is the number a user needs in order to decide what to do with 40 flags on their own data.

The reason this matters more here than in most tooling: a detector at the dataset layer sits upstream of every policy trained on it. If it silently removes 8 percent of valid demonstrations on a particular task family, the downstream success rate moves and nobody attributes it to the cleaner.

Same discipline we hold ourselves to one layer down, at policy evaluation: every detection rate gets published beside its firing rate on known-benign input, and if the firings concentrate on a subset rather than spreading, that is a signal about the detector, not the data.

 
 
[omert27](https://discuss.huggingface.co/u/omert27)
6
 
Thanks for the feedback Sattyam, i will implement these benchmarks.