SageMaker Build - Train Pipelines
This folder contains all the SageMaker Pipelines of your project.
buildspec.yml defines how to run a pipeline after each commit to this repository.
ml_pipelines/ contains the SageMaker pipeline definitions.
The expected output of the main pipeline (from training/pipeline.py) is a model registered to SageMaker Model Registry.
source_scripts/ contains the underlying scripts run by the steps of your SageMaker Pipelines. For example, if your SageMaker Pipeline runs a Processing Job as part of a Processing Step, the code being run inside the Processing Job should be defined in this folder.
Typically source_scripts/ can contain individual scripts preprocessing.py, training.py, postprocessing.py, evaluate.py, depending on the nature of the steps run as part of the SageMaker Pipeline.
We provide here an example with the Abalone dataset, to train an XGBoost model and evaluate it on a test set before sending it for manual approval to SageMaker Model Registry inside the SageMaker ModelPackageGroup defined when creating the SageMaker Project.
Additionally, if you use custom containers, the Dockerfile definitions should be found in that folder.
notebooks/ contains experimentation notebooks.