Does Diffusion Policy run on Mobile ALOHA?
works with caveats
Capabilities
| Capability | State | Notes | Evidence |
|---|---|---|---|
| Train / fine-tune | works | ev-mobile-aloha-paper, ev-act-plus-plus | |
| Real-robot inference | works | ev-mobile-aloha-paper | |
| Simulation inference | no verified evidence | — | |
| Teleop + record | no verified evidence | — |
Caveats
- With 50 demonstrations Diffusion Policy was less precise than ACT on Wipe Wine (65% success); the authors attribute this to too few demonstrations.
- Documented with the Mobile ALOHA authors' codebase (MarkFzp/act-plus-plus, DDIM scheduler, co-training with static ALOHA data), not LeRobot's diffusion policy.
Evidence 2 sources; the cell shows the highest tier
- T1 · upstream documentedpapersource 2024-01-04verified 2026-10-24 (1d ago)
Mobile ALOHA paper (Sec. 6.2) trains Diffusion Policy on Mobile ALOHA: 100% on Push Chairs with co-training, 65% on Wipe Wine.
“Overall, Diffusion Policy performs similarly to ACT on Push Chairs, both obtaining 100% with co-training. For Wipe Wine, we observe worse performance with diffusion at 65% success.”
https://arxiv.org/abs/2401.02117 - T1 · upstream documentedsource-codesource 2024-01-11verified 2026-10-24 (1d ago)
Authors' repo implementing ACT, Diffusion Policy and VINN with co-training for Mobile ALOHA.
“This repo contains the implementation of ACT, Diffusion Policy and VINN, together with 2 simulated environments:”
https://github.com/MarkFzp/act-plus-plus/blob/26bab0789d05b7496bacef04f5c6b2541a4403b5/README.md
Embed this cell
Badge for READMEs and model cards. It always shows the current state.
[](https://caniembody.pages.dev/compat/mobile-aloha/diffusion)<iframe src="https://caniembody.pages.dev/embed/mobile-aloha/diffusion" width="380" height="132" style="border:0" title="Diffusion Policy on Mobile ALOHA | CanIEmbody"></iframe>History
- 2026-10-25 fecccd1 Initial release: robot policy compatibility matrix · Danilo Znamerovszkij