ARPO
Official Implementation of ARPO: End-to-End Policy Optimization for GUI Agents with Experience Replay
About
ARPO (Agentic Replay Policy Optimization) is designed to optimize policy for GUI agents using reinforcement learning. It focuses on completing long-horizon desktop tasks by processing multi-modal inputs, including screenshots and actions. The framework supports distributed rollouts for scalable task execution across parallel environments. It is particularly suited for developers and researchers working on GUI automation and reinforcement learning applications.
Open Source Health
- Stars
- 163
- Forks
- 11
- License
- Apache-2.0
- Last commit
- 1 years ago
Related Categories
Vendor
JIA Lab
JIA Lab
Quick Links
Open Source
More by JIA Lab
Related Products
PolicyPlus
Local Group Policy Editor plus more, for all Windows editions
Top category match
App Privacy Policy Generator
Generate a customized Privacy Policy and Terms of Use document for your mobile apps
Top category match
GPOZaurr
Group Policy Eater is a PowerShell module that aims to gather information about Group Policies but also allows fixing issues that you may find in them.
Top category match
Community-Policy
This repo is for Microsoft Azure customers and Microsoft teams to collaborate in making custom policies.
Top category match
LLMGA
This project is the official implementation of 'LLMGA: Multimodal Large Language Model based Generation Assistant', ECCV2024 Oral
More from this vendor