In addition to the notebook docker image described in the main README, we provide the gptdev/shell Docker image which is further described below. For more information on how to use Docker, please see the Docker documentation.

This image includes a user preinstalled Python 3.10 and GPT setup. Start the session with

docker run --rm -it -v$(pwd):/gpt-code gptdev/shell

which mounts the current working directory in the image's working directory. Alternatively, you can use a non-persistent version of the GPT applications, benchmarks, and tests folders by running

docker run --rm -it gptdev/shell