[ Frontier data scoring for physical AI ]
Train your model
on your best data.
Skyu scores every training data sample by its real impact on your model, so you can train on the best subset that improves your model.
Train your model
Find the demonstrations that hurt your policy.
Skyu scores every episode in a dataset against your own trained policy, flags the ones worth a closer look, and crafts a report. It runs on-prem, offline, and air-gapped down to the edge.
$ pip install skyu-0.2.1-py3-none-any.whl
$ skyu score --dataset ./pusht --checkpoint ./ckpt/030000
scored 1,204 episodes
flagged 96 for review
→ report.html
report.html
1,204
scored
96
flagged
[ HOW IT WORKS ]
01
Point
Give Skyu a LeRobot, MCAP, ROS-bag (or any other format) dataset and a policy checkpoint. Runs on-prem, offline, and air-gapped.
02
Score
Skyu scores every episode in your dataset and ranks them, so the demonstrations most likely to be hurting your policy rise to the top.
03
Review
The report flags episodes worth a closer look, with a reason and a short clip of each episode.