In this project, we wrapped public accessible ml/dl models with kubeflow pipeline.
each pipeline should under a named folder with a file named pipeline.py, this is the only reference the kfp.compiler can find you and build a workflow resource dynamically.
Simple launch the following command to build all manifests that we can find with the following pattern find pipelines -name "*pipeline.py"
make gen-manifests