# real-stanford/diffusion_policy

[RSS 2023] Diffusion Policy Visuomotor Policy Learning via Action Diffusion

Repository: https://github.com/real-stanford/diffusion_policy
Canonical: https://ross.abutalabs.com/products/diffusion_policy
Homepage: https://diffusion-policy.cs.columbia.edu/
Language: Python
License: MIT
License Family: permissive
Topics: robotics
Last push: 2024-12-24T19:48:03+00:00

## Health v2 (maintenance only)
Score: 30/100 (v2, computed 2026-09-03T02:20:16.233290+00:00)
- activity 0, release rhythm 35, longevity 91
- inputs: {"age_days": 1275, "days_push": 617, "days_rel": null, "gap_med": null, "n_releases_24m": 0}
- flags: no_releases
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 4491, forks 830 (observed 2026-08-28T04:08:51.307817+00:00)

## What it is
Official PyTorch implementation of Diffusion Policy, a visuomotor robot policy learning method that represents robot behavior as a conditional denoising diffusion process. It includes training/evaluation code, experiment configs, checkpoints, and Colab notebooks for state-based and vision-based manipulation tasks.

## Use cases
- train diffusion-based robot manipulation policies from demonstrations
- reproduce visuomotor policy learning experiments from the RSS 2023 paper
- benchmark imitation learning methods on robot manipulation tasks
- run pretrained policies in simulation environments
- experiment with diffusion models for multimodal action distribution learning

## When to choose
- you need state-of-the-art visuomotor policy learning for robot manipulation
- you want to reproduce or extend published diffusion policy experiments
- your task has multimodal action distributions that Gaussian policies handle poorly

## When to avoid
- you need a lightweight production robot control stack rather than research code
- you lack GPU resources for diffusion training and inference
- your domain is unrelated to robot learning or manipulation

## Facets
- artifact type: library
- maturity: stable
- function: machine-learning, deep-learning, robotics, simulation
- domain: robotics, machine-learning, deep-learning, artificial-intelligence
- platform: python
- tags: diffusion-models, imitation-learning, visuomotor-policy, robot-manipulation, research-code, linux, gpu

## Member repositories
- real-stanford/diffusion_policy (main) score 30

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:08:51.307817+00:00.
- Health v2: computed from the inputs above; adoption is never an input.
- Inferred fields (summary, facets, guidance): AI-extracted, prompt v1, taxonomy v1, on 2026-08-29T18:20:29.052339+00:00, confidence not recorded.
  - readme: https://github.com/real-stanford/diffusion_policy (fetched 2026-08-28T04:08:51.307817+00:00, sha 7d0b97f2c038)
  - homepage: https://diffusion-policy.cs.columbia.edu/ (fetched 2026-08-29T09:06:42.168953+00:00, sha edadad05d2bd)
- Data as of 2026-08-30T08:39:29.467469+00:00.