Tutorials

We have included a few tutorials to demonstrate the usage of quanda. To install the library with tutorial dependencies, run:

pip install -e '.[tutorials]'

The tutorials currently included in quanda are:

• Explainers: shows how different explainers can be used with quanda. This tutorial goes through all the explainers that are included in quanda and walks through the steps of initializing the Explainer object, generating explanations and plotting them.

• Metrics: shows how to use the metrics in quanda to evaluate the performance of a method. This tutorial goes through all the metrics that are included in quanda and walks through the steps of initializing the metric and evaluating the performance of a TDA method.

• Benchmarks: shows how to use the benchmarking tools in quanda to evaluate a data attribution method. This tutorial includes 3 different examples of benchmarks.

• Linear Datamodeling Score (LDS): caveats and best practices for the LDS benchmark, including how to precompute and reuse counterfactual subset logits across explainers.