GuidesDatasets and Models

Tabular Inputs

Prepare CSV, TSV, or Parquet classification and regression data.

Knotest supports tabular classification and regression in CSV, TSV, and Parquet files. The registered Dataset format records how Knotest interprets the target column and feature data.

Preparation checklist

  • Decide whether the target is categorical or numeric.
  • Identify the target column explicitly.
  • Keep feature names and types consistent with the Model input contract.
  • Review missing values and categorical encodings before import.
  • Confirm the Model output matches classification or regression.
  • Avoid leaking the target into the Model feature set.

Review the Dataset profile after import before you generate predictions. A parseable table can still contain scientifically invalid feature or target data.

This example uses a separate synthetic 40-row CSV (not the Helmet vision Dataset). feature is the input, target is the classification target, sample_id identifies rows, and group contains demonstration group labels.

Tabular classification example with sample IDs, feature values, target labels, and groups.

Watch: Review tabular inputsMove from the sample table to column and format settings.Silent demo · Subtitles follow page language · Full screen for detail

CSV Classification settings with target column, sample ID column, and input structure.

Watch: Review tabular settingsInspect the sample table, target column, and sample ID settings.Silent demo · Subtitles follow page language · Full screen for detail

Oriented Bounding Box Inputs

Prepare YOLO four-corner oriented bounding-box annotations and Model output.

Configure an API Profile

Use a synchronous HTTP JSON inference operation as a reusable Inference Source.

On this page

Preparation checklist