Skip to content

Conversation

@piotrlaczkowski
Copy link
Collaborator

Describe your changes shortly

This PR introduces Periodic and Parameterized Linear Expansion (PLE) numeric embeddings, along with an Advanced Combined Embedding layer. These new capabilities enhance KDP's performance on tabular tasks by providing sophisticated ways to model numerical features, aiming to close the performance gap between MLPs/Transformers and tree-based models.

The changes include:

  • New PeriodicEmbedding, PLEEmbedding, and AdvancedNumericalEmbedding layers.
  • Integration into kdp/features.py, kdp/layers_factory.py, and kdp/processor.py for flexible usage.
  • Comprehensive unit and integration tests for all new components.
  • Updated documentation with detailed explanations, configuration options, and usage examples.

Checklist

My code follows all rules defined in:

  • pre-commits are passing !
  • unite-tests are passing !
  • code is well documented (docstrings in the corresponding format including examples of usage etc ...)
  • code API is connected to the documentetion
  • all requirements are added using poetry

Open in Cursor Open in Web

Learn more about Cursor Agents

…yers

Co-authored-by: piotr.laczkowski <piotr.laczkowski@gmail.com>
@github-actions
Copy link

PR Preview Action v1.6.2

🚀 View preview at
https://piotrlaczkowski.github.io/keras-data-processor/pr-preview/pr-34/

Built to branch gh-pages at 2025-07-30 21:50 UTC.
Preview will be ready when the GitHub Pages deployment is complete.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

3 participants