GPT playground

Just me playing with large language models, langchain, etc.

gptme

An interactive CLI to let you interact with LLMs in a Chat-style interface.

With features like:

• Supports OpenAI and any model that runs in llama
  • Thanks to llama-cpp-server!
• Tools
  • Access to the local machine
  • Execute shell/Python code on the local machine.
    • Command output (stdout & stderr + error code) will be feeded back to the agent, making it able to self-correct errors etc.
• Can handle long context sizes through summarization.
  • (not very well developed)

Usage

Install deps:

poetry install

To use locally, you need to start llama-cpp-server:

poetry run python -m llama_cpp.server --model ~/ML/Manticore-13B.ggmlv3.q4_1.bin

Then you can interact with it using:

gptme --llm llama

TODO

Ideas for things to try:

• An agent that looks up the latest CI job, and if it failed, tries to figure out how to fix it.
• An agent that looks up recent GitHub issues, tries to generate an answer.
• An agent that looks up recent GitHub PRs, tries to generate an action (comment, ask for human review, merge, close).