HybridRoboticssupermimic
- Repository stars
- 1,918Star
- All-time
- #27,374 all time
- Fork
- 156 Fork
- Repository created
- since 2026
Humanoid whole-body motion tracking with one differentiable world model.
TypeModels & adaptersPurposeMiscellaneousRobotics & Autonomous SystemsLanguagePythonLicenseApache-2.0
Commercial use OKApache-2.0; keep the license notice
Actively maintainedLast push 1 days ago
About
SuperMimic trains humanoid whole-body motion tracking with a differentiable world model.
It targets research and training scenarios needing G1 whole-body tracking, offering generalized and maximal coordinate task representations. Train a world model and policy as described, then play back a trained model.
Summarized by AI from the READMEREADME
As of 2026-11-22, HybridRobotics/supermimic has 1,918 GitHub stars and 156 forks. Written mainly in Python; licensed under Apache-2.0; last pushed on 2026-11-21.Record updated:
README
View on GitHubRaw observations · 12
| Measured at | Repository stars | Source |
|---|---|---|
| 2026-11-22 00:22 UTC | 1,918 | Source |
| 2026-11-21 00:31 UTC | 1,897 | Source |
| 2026-11-20 00:22 UTC | 1,857 | Source |
| 2026-11-19 00:22 UTC | 1,809 | Source |
| 2026-11-18 00:22 UTC | 1,749 | Source |
| 2026-11-17 00:23 UTC | 1,660 | Source |
| 2026-11-16 00:20 UTC | 1,554 | Source |
| 2026-11-15 00:22 UTC | 1,499 | Source |
| 2026-11-14 00:27 UTC | 1,435 | Source |
| 2026-11-13 00:22 UTC | 1,304 | Source |
| 2026-11-12 01:30 UTC | 1,034 | Source |
| 2026-11-12 01:03 UTC | 1,029 | Source |