Databricks Mosaic Generative AI Cookbook#
The Databricks Mosaic Generative AI Cookbook is a collection of notebooks demonstrating different use cases for Databricks Mosaic GenAI tools such as the Foundation Model API and the Model Playground.
Featured Notebooks#
How to Run These Notebooks in your Databricks Workspace#
Import individual notebooks#
If you want to run one of these notebooks in your Databricks workspace, click the download icon in the upper right corner of the page you would like to run. Then right-click on .ipynb
and select “Copy link address.” In your Databricks workspace, in the workspace view, select “import” (from the three-dots menu or from the right-click menu) and paste in the url you copied. This will import the notebook to your Databricks workspace.
Clone the repository#
You can also clone the whole repository into your workspace. Simply navigate to “Repos” in your Databricks workspace, click “Add Repo”, and copy in the URL: databricks-genai-cookbook/cookbook. Then click Create Repo. This will clone the repository to your workspace, giving you access to all of the notebooks right in your Databricks workspace.
Official Documentation#
- Query Foundation Models with Databricks SQL
- Fast Classification with Provisioned Throughput
- Using the MLflow Prompt Engineering UI
- Querying the Foundation Model API with the OpenAI Python SDK
- Using the Foundation Model API Outside of Databricks
- Adjusting Foundation Model Behavior with Generation Parameters
- Managing Chat Sessions in Python
- Test models with the Databricks AI Playground
- Serve a Model from Databricks Marketplace
- Streaming Outputs with the Foundation Model API
- Foundation Model API Embeddings for Vector Search