AdityaNG/cranberry

Teaching LMMs to interact with computers

CRANBERRY-GPT

Teaching LMMs to interact with computers

Install it from PyPI

pip install cranberry_gpt

Citation

If you find our work useful cite us!

@misc{NG2024CRANBERRYGPT,
  author       = {Aditya Nalgunda Ganesh},
  title        = {CRANBERRY-GPT: Teaching LMMs to interact with computers},
  year         = {2024},
  month        = {September},
  note         = {Release 1.0.0, 24th September 2024},
  url          = {https://github.com/AdityaNG/cranberry/}
}

Usage

TODO

# TODO

Setup for Development

# Download Repo
git clone https://github.com/AdityaNG/cranberry
cd cranberry
git pull

make virtualenv

TODO Points

• Draw Architecture diagram
• Build the Dockercontainer w/ OpenInterpreter
• Implement LLM Proxy to intercept all the chat data
• Implement MCTS
• Logging for all the agent's data
• Setup RL experiments (need to list out more experiments)
  • Create a file on the desktop, open it in the GUI and save the screenshot
  • Open the Gmail and Summarise the top 5 unread emails
• Generate dataset of runs with the RL experiments
• Fine-tune LMMs on the newly created
• Documentation: mkdocs gh-deploy
• Test Cases

Development

Read the CONTRIBUTING.md file.

References

• TODO