ARPO training framework

Replay-based policy-optimization framework for GUI agents, with distributed OSWorld rollouts and screenshot/action histories.

AI-assisted editorial review against a primary source. This is a catalog annotation, not an independent reproduction or endorsement.

What is this resource?

Resource type

Frameworks

Environment

desktop, linux

Topics

reinforcement-learning, long-horizon, automation

Record publication date

Not recorded

Evidence retrieved

2026-10-18T04:27:12Z

Resource ID

github:jia-lab-research/arpo

What licenses are recorded?

Code, training data, weights, and documents can have different terms. A blank declaration means unknown, not unrestricted use.

Code

Apache-2.0 (Reviewed repository license declaration for code; third-party dependencies, data, weights and executable environment images are separate.)

Data

Not recorded

Document

Not recorded

Weights

Not recorded

License review notes

• Code: Reviewed repository license declaration for code; third-party dependencies, data, weights and executable environment images are separate. Source 1

Recorded specifications

repository tree sha

7489ae309e90d8b3b9e823f495b63e5116434060

repository url

https://github.com/JIA-Lab-research/ARPO

Explore related topics

Scope and limitations

Official GitHub API resolves the former dvlab-research/ARPO URL to JIA-Lab-research/ARPO with the same numeric repository ID. Training/evaluation scripts are released; OSWorld environments and the separate model checkpoint are not bundled into this code record. Reported benchmark scores were not reproduced.

Related resources

• documented by: ARPO:End-to-End Policy Optimization for GUI Agents with Experience Replay

Primary evidence and verification

Source retrieval is separate from experiment reproduction. Use the original publication or publisher page when citing claims about the resource.

1. editorial primary review

   Retrieved 2026-10-18T04:27:12Z. Supports: name, kind, summary, environments, topics, scope, notes, relations.

   Content fingerprint

   839b2de0e0c614fe8d2df4782d0f6a6e7d33f5c60d8020c9d4a01c61d533f39e
2. editorial primary review

   Retrieved 2026-10-18T04:40:39Z. Supports: identity.

   Content fingerprint

   bb64c7fc1b5d88fc070e919516a51eefac4acb89e86773464cfa0b3de18f1898
3. editorial primary review

   Retrieved 2026-10-18T04:25:23Z. Supports: identity, release_status, details.repository_tree_sha.

   Content fingerprint

   b3ff7f4e350223827cae0fffd986ee672e0b4466349c262ed2dc9e63a8480aeb
4. editorial primary review

   Retrieved 2026-10-17T15:37:00Z. Supports: identity.

   Content fingerprint

   bb64c7fc1b5d88fc070e919516a51eefac4acb89e86773464cfa0b3de18f1898
5. editorial primary review

   Retrieved 2026-10-18T04:35:19Z. Supports: licenses.code.

   Content fingerprint

   c71d239df91726fc519c6eb72d318ec65820627232b2f796219e87dcf35d0ab4
6. github repository identity

   Retrieved 2026-10-17T15:37:00Z. Supports: aliases.

   Content fingerprint

   bb64c7fc1b5d88fc070e919516a51eefac4acb89e86773464cfa0b3de18f1898
7. github repository identity

   Retrieved 2026-10-18T04:40:39Z. Supports: aliases.

   Content fingerprint

   bb64c7fc1b5d88fc070e919516a51eefac4acb89e86773464cfa0b3de18f1898

Suggest a correction

Open a source-backed correction. Include this record ID and the original source. Read our editorial policy and coverage boundaries.