Evaluation

What is dataset curation?

Dataset curation selects, cleans, and versions the rows that train or evaluate a model.

Updated · Part of What is active learning?

How it works

  • Gather candidate rows.
  • Drop exact and near duplicates, fix labels, and hold out a test slice.
  • Pin the version the run used.

What it is not

It is not an unversioned scrape, and it is not only storing weights.

dataset curation: this, and the thing it is confused with

dataset curation: this, and the thing it is confused with
ThisNot this
OutputA versioned set of rowsA dump
IncludesLabels and mediaWeights
ToolsFilters, embeddings, peopleA spreadsheet with paths

Where Pixeltable fits

Pixeltable curation is queries and columns on the media table: dedup, labels, and a version you can name.

Questions

How does dataset curation work?
Gather candidate rows. Drop exact and near duplicates, fix labels, and hold out a test slice. Pin the version the run used.
What is dataset curation often confused with?
It is not an unversioned scrape, and it is not only storing weights.