Skip to content

Requesting a Feature / Change

We welcome feature requests for new neural primitives, loss functions, metrics, integration helpers, or workflow automation scripts!

💡 Feature Request Guidelines

When suggesting a new feature:

  1. Academic Provenance: If proposing a new loss function or neural block, provide the paper reference (Title, Authors, DOI/arXiv URL).
  2. Use Case: Explain why this feature is valuable to academic researchers or production practitioners.
  3. Proposed API: Sketch out a Python snippet demonstrating how the new class or method would be invoked.