Skip to content

stavshamir/impllmentors

Repository files navigation

impllmentors

impllmentors is a POC project demonstrating the emerging approach of structured and flow-controlled LLM agents (as opposed to freely roaming agents as in AutoGPT and BabyAGI).

It is humble, and can only implement simple python modules. However simple, it also:

  • Has well-defined structure and steps, avoids running in circles.
  • Asks for human feedback between steps.
  • Writes unit tests.
  • Runs the unit tests, and if they fail, tries to fix the code/

How To Use

  1. Clone the repository: git clone https://github.com/stavshamir/impllmentors.git
  2. Run pip install -f requirements.txt
  3. Set an OPENAI_API_KEY environment variable
  4. Run python runner.py

Other Things To Consider

  • The generated code and tests are written to the directory from which you run the script.
  • You can view the written files and edit them between steps to help the AI.
  • After the unit tests are written, the AI will not attempt to fix them in case they fail.
  • This uses gpt-3.5-turbo, as I (and many others) still don't have access to the gpt-4 api. This means that hallucinations still happen quite often.

Demo

(Sorry for the low quality) demo

Contribution

Feel free to suggest improvements and provide PRs!