Google Cloud Pipeline Components¶

Google Cloud Pipeline Components (GCPC) provides predefined KFP components that can be run on Google Cloud Vertex AI Pipelines and other KFP-conformant pipeline execution backends. You can compose the components together into pipelines using the Kubeflow Pipelines SDK.

Documentation¶

User documentation¶

Please see the Google Cloud Pipeline Components user guide.

API documentation¶

Please see the Google Cloud Pipeline Components API reference documentation.

Release details¶

For details about previous and upcoming releases, please see the release notes.

Examples¶

• Train an image classification model using Vertex AI AutoML.

• Train a classification model using tabular data and Vertex AI AutoML.

• Train a linear regression model using tabular data and Vertex AI AutoML.

• Train a text classification model using Vertex AI AutoML.

• Use the Google Cloud pipeline components to upload and deploy a model.

Installation¶

Requirements¶

• Python >= 3.7

• A Google Cloud project with the Vertex API enabled.

• An authenticated GCP account

Install latest release¶

Use the following command to install Google Cloud Pipeline Components from PyPI.

pip install -U google-cloud-pipeline-components

Install from source¶

Use the following commands to install Google Cloud Pipeline Components from GitHub.

git clone https://github.com/kubeflow/pipelines.git
pip install pipelines/components/google-cloud/.

Build the package from source and install¶

Use the following commands build Google Cloud Pipeline Components from the source code and install the package.

source_root=$(pwd)

git clone https://github.com/kubeflow/pipelines.git
cd pipelines/components/google-cloud
python setup.py bdist_wheel clean

WHEEL_FILE=$(find "$source_root/pipelines/components/google-cloud/dist/" -name "google_cloud_pipeline_components*.whl")
pip3 install --upgrade $WHEEL_FILE