This is a Dev Container enabled template for Microsoft LIDA, making it easier for you to instantiate, run, and extend, Jupyter Notebooks running LIDA-driven visualizations in the cloud (using GitHub Codespaces) or on your local machine (using Docker Desktop) with a single click.
To get started using this template,
- First, click
Use this template
in the options provided on the GitHub repository for this project - Next, select
Create a new repository
to instantiate a copy in your profile or organization - Finally, in your copy, click the
Code
option, switch to theCodespaces
tab - andCreate codespace on main
This should launch a GitHub Codespaces version of development container (runtime) in your browser, and let you open - and run - the default basic Python notebook to validate that your environment works.
To start using LIDA, you need to setup the API keys for the relevant Large Language Model of your choice using the related LLMX Package - which provides a unified interface to several LLM providers for chat fine-tuned models including:
- OpenAI
- Azure OpenAI
- Google PaLM (via MakerSuite)
- Google PaLM (via Vertex AI)
- Cohere
- Hugging Face (local)
See examples of setup for usage within Jupyter notebooks and more.
Each LLM provider has its own set of environment variables that need to be setup for usage.
- Docker Desktop | create a
.env
and update it with your keys. We.gitignore
the file so your secrets will not get checked into repo. - GitHub Codespaces | you have to option to manage secrets for your Codespaces via GitHub settings.