FSoft-AI4Code / AgileCoder

Incorporating Agile methodology into agents to create complex real-world softwares
MIT License
374 stars 52 forks source link

Minimum requirements to run locally #10

Closed mbenencase closed 3 months ago

mbenencase commented 3 months ago

Hi, following my other issue, I have two questions:

  1. How to run locally? In the README.md I saw that we must configure some environment variables.
  2. What are the minimum hardware requirements? RAM memory, GPU VRAM, and so on.

Thanks in advance!

minhngh commented 3 months ago

Hi mbenencase,

Thank you for your questions. Below is our answer:

  1. To run on local machines, we install the library and then set some environment variables, as described in Readme.md

    • Install the library:
      pip install agilecoder
      pip install git+https://github.com/tree-sitter/tree-sitter-python.git@master
    • Set environment variables: Because our method uses LLMs like Azure GPTs as backend models, we have to set relevant variables:
      API_KEY=your_api_key
      RESOURCE_ENDPOINT=your_resource_endpoint
      API_TYPE=azure
      API_VERSION=your_api_version
      API_ENGINE=your_api_engine
  2. There are no special requirements to run our library. Our method employs Azure GPTs, so it doesn't require GPUs and consume much RAM or CPU.