https://github.com/charmve/puppygo

vision language model and large language model powered embodied robot
https://github.com/charmve/puppygo

chatgpt deepmind embodied-agent palm2 rt-2 voxposer

Last synced: 4 months ago
JSON representation

vision language model and large language model powered embodied robot

• Host: GitHub
• URL: https://github.com/charmve/puppygo
• Owner: Charmve
• Created: 2023-09-05T02:05:57.000Z (over 1 year ago)
• Default Branch: main
• Last Pushed: 2023-09-18T03:45:23.000Z (over 1 year ago)
• Last Synced: 2024-04-15T09:05:25.589Z (about 1 year ago)
• Topics: chatgpt, deepmind, embodied-agent, palm2, rt-2, voxposer
• Homepage:
• Size: 11.7 MB
• Stars: 5
• Watchers: 2
• Forks: 0
• Open Issues: 0
• Metadata Files:
  • Readme: README.md

Awesome Lists containing this project

README

        
# PuppyGo

Vision language model and large language model powered embodied agent.



  


    

    

  



## Here’s what I did:

- Language Models as Zero-Shot Planners: Extracting Actionable Knowledge for Embodied Agents

- extracts affordances and constraints from large language models and vision-language models to compose 3D value maps, which are used by motion planners to zero-shot synthesize trajectories for everyday manipulation tasks.

- combine with e2e large model trainning framework, like UniAD;

## This Package Is Sponsorware 💰💰💰

[![](https://img.shields.io/static/v1?label=Sponsor&message=%E2%9D%A4&logo=GitHub&color=%23fe8e86)](https://github.com/sponsors/Charmve?frequency=one-time&sponsor=Charmve) https://github.com/sponsors/Charmve?frequency=one-time&sponsor=Charmve

This repo was only available to my sponsors on GitHub Sponsors until I reached 15 sponsors.

Learn more about **Sponsorware** at [github.com/sponsorware/docs](https://github.com/sponsorware/docs) 💰.

![image](https://github.com/Charmve/PuppyGo/assets/29084184/3577e267-53a0-4904-909f-6733995e892e)




## Execution under Disturbances

Because the language model output stays the same throughout the task, we can cache its output and re-evaluate the generated code using closed-loop visual feedback, which enables fast replanning using MPC. This enables VoxPoser to be robust to online disturbances.

      



  


    

      

    

    


      "Sort the paper trash into the blue tray."

    

  


  


    

      

    

    


      "Close the top drawer."