/ml-explainability

Primary LanguageJupyter NotebookApache License 2.0Apache-2.0

Using LIME and ANCHOR to explain CMLE models/predictions

Before you begin

  1. Create a new Cloud Platform project.

  2. Enable billing for your project.

  3. Enable APIs for

  • BigQuery API
  • AI Platform
  • Compute Engine API

Run in AI Platform Notebooks

  1. Create a Tensorflow 1.x Notebook following the instructions in the AI Platform Documentation

  2. Open JupyterLab

  3. Use the 'git clone' button to clone this repository from git, or the 'upload files' button to upload from a local directory.

  4. Open a Terminal window in JupyterLab

  5. pip install anchor-exp lime