liveBook · Manning

TabularDataset

As explained in Machine Learning for Tabular Data: XGBoost, Deep Learning, and AI

A managed dataset resource in Vertex AI that organizes data for machine-learning development and training.

TabularDataset is a managed Vertex AI dataset resource that gives a machine-learning training job a named, reusable handle for tabular data. Instead of passing a local CSV directly to the job, upload the processed file to Cloud Storage, create a Vertex AI dataset, and retain its dataset ID for pipeline configuration. section 11.2.7 section 11.5

A managed dataset supports the training workflow by:

  • Referencing the source data through a resource path assembled from the project, region, and dataset ID. section 11.3.6
  • Making the dataset available through Cloud Storage for the training script. section 11.2.7
  • Dividing the data into training, validation, and test subsets according to the pipeline configuration, then splitting those subsets into CSV files. section 11.2.7
  • Serving as the dataset argument when a CustomTrainingJob runs the training script. section 11.3.6

Using a Kuala Lumpur real-estate dataset

After uploading the cleaned Kuala Lumpur real-estate CSV to a Cloud Storage location such as gs://first-project-ml-tabular-bucket/processed_dataset/kl_real_estate_output.csv, create a tabular regression or classification dataset in Vertex AI and record its ID. The pipeline can then construct and use it as follows: dataset_path = 'projects/' + project_id + '/locations/' + region + '/datasets/' + dataset_id; ds = aiplatform.TabularDataset(dataset_path); model = job.run(dataset=ds, training_fraction_split=training_fraction, validation_fraction_split=validation_fraction, test_fraction_split=test_fraction). section 11.2.6 section 11.2.7 section 11.3.6

dataset_path = \
'projects/'+config['project_id']+\
'/locations/'+config['region']+\
'/datasets/'+config['dataset_id']
ds = aiplatform.TabularDataset(dataset_path)

Inspect the Python listing that builds the managed dataset resource path and creates the TabularDataset object used by the pipeline. section 11.3.6

The resource does not replace pipeline preparation: the cleaned file must already be uploaded, and the training container must use Cloud Storage locations because it cannot access the surrounding file system. section 11.5

How Machine Learning for Tabular Data: XGBoost, Deep Learning, and AI covers it

How deep it goes
Explains creation, resource-path construction, automatic Cloud Storage preparation, dataset splitting, and training-job integration.
Examples
Uses a cleaned Kuala Lumpur real-estate CSV, a Cloud Storage URI, a dataset ID, and Python SDK training-job code.
What it assumes
Assumes a cleaned CSV, Cloud Storage bucket, Google Cloud project settings, and a created Vertex AI dataset with a recorded ID.
What you can do after
Operate/configure a Vertex AI pipeline that references tabular data, runs training with explicit splits, and optionally deploys the model.
How it connects
Connects Cloud Storage, Vertex AI SDK, CustomTrainingJob, training-validation-test splits, ML pipelines, and Vertex AI endpoints.
Cover of Machine Learning for Tabular Data: XGBoost, Deep Learning, and AI
liveBookmanning.com
Read Machine Learning for Tabular Data: XGBoost, Deep Learning, and AI on liveBook Get the book at manning.com

Related concepts

This explainer was generated by Manning's knowledge graph from the text of Machine Learning for Tabular Data: XGBoost, Deep Learning, and AI. The section links point to the parts of the book it draws on.