Module Quiz » Week 2 » ML Pipelines On Google Cloud

Q: When designing a custom TFX component from a Python function, which of the following steps are needed?

Correct! Initializing your custom component is the last step in building a custom component from a Python function. <br> Correct! Decorating an existing Python function with a @component decorator is the first step in building a custom component from a Python function. <br> Correct! Using Python type annotations to define artifact and runtime parameter types is the second step in building a custom component from a Python function.

Q: Your ML Engineering team has existing ML pipeline code written in Java. Which TFX custom component option can you re-use this code in your TFX pipeline to speed up development? Select one.

Correct! You can create custom components from pre-built containers backed by command line programs

Q: How do TFX pipeline development steps relate to the levels of ML process automation? Select all that apply:

Correct! The TFX InteractiveContext enables pipeline developers to manually execute each pipeline component interactively in a Jupyter notebook. <br> Correct! Google Cloud tools like Cloud Build enable CICD automated builds, testing, and deployments of continuous training TFX pipelines. <br> Correct! Google Cloud AI Platform Pipelines enable automated continuous pipeline training on Kubeflow Pipelines.