Diffusion Policy

Architecture and scope

The release includes configurations and checkpoints behind its experiments. Evaluation depends on the observation encoder, action horizon and task dataset rather than the Diffusion Policy name alone.

Integration and evaluation

This is a policy-learning method, useful as a reproducible baseline and as an action-generation component. It should not inherit language-understanding claims from later VLA models that use diffusion heads.

Evidence boundary

A method-level comparison must specify the trained policy and benchmark task.

Model identity

Developer

Cheng Chi and coauthors

Primary specReviewed 2026-09-28

Sources 1

real-stanford/diffusion_policy official repository

real-stanford · Publication date not disclosed · Source accessed 2026-09-28

Official README inspected live. Repository code terms do not automatically cover model weights, data, or third-party assets.

Exact version / scope

Diffusion Policy reference implementation

Primary specReviewed 2026-09-28

Sources 1

real-stanford/diffusion_policy official repository

real-stanford · Publication date not disclosed · Source accessed 2026-09-28

Official README inspected live. Repository code terms do not automatically cover model weights, data, or third-party assets.

Category

Diffusion action policy

Primary specReviewed 2026-09-28

Sources 1

real-stanford/diffusion_policy official repository

real-stanford · Publication date not disclosed · Source accessed 2026-09-28

Official README inspected live. Repository code terms do not automatically cover model weights, data, or third-party assets.

Access model

Public training code, configurations, logs and checkpoints

Primary specReviewed 2026-09-28

Sources 1

real-stanford/diffusion_policy official repository

real-stanford · Publication date not disclosed · Source accessed 2026-09-28

Official README inspected live. Repository code terms do not automatically cover model weights, data, or third-party assets.

License

Not publicly disclosed

Inputs, outputs & embodiment

Inputs

Not publicly disclosed

Outputs

Not publicly disclosed

Embodiment

Not publicly disclosed

Implementation & training

Compute

Not publicly disclosed

Training-data disclosures

Not publicly disclosed

Integrations

Not publicly disclosed

Reported benchmark results

Relationships & deployments

Related reading is an editorial crosslink. Sourced connections describe relationships reported in the cited material. A link to a versioned profile does not establish compatibility with that version unless the connection note explicitly identifies it.

Record history & verified changes

A research review records when we checked a source. It does not mark a product launch or a new deployment.

Initial reviewed record. No subsequent field change has been recorded.