CI Tips - jimmytwei/oneAPI-samples GitHub Wiki

Occasionally we encounter issues with CI when we start adding tools or other things. Here are a couple of issues we have found and the resolution that worked

Problem: Installation of runipy causes CI issues. This has caused issues on two samples currently, normally we would just install and run “runipy sample.ipynb” which began causing issues for running on a wrong conda environment in some samples.

• Solution: Issue is resolved by running the JN in a Jupyter Kernel after it was created using nbconvert – this new method can let you run your jupyter notebook on a specific registered jupyter kernel, for specifics take a look at the sample.json from IntelTensorFlow_InferenceOptimization and sample.json from IntelPython_XGBoost_Performance