Welcome to edge2torch Discussions!

[Thomas-Rauter]

Welcome to edge2torch Discussions

This is a place for open-ended questions, ideas, and usage discussions around edge2torch.

edge2torch builds sparse PyTorch neural networks from edge lists of named nodes, with optional feature- and node-level attribution.

Use Discussions for:

• asking how to model a specific graph-defined architecture;
• discussing possible applications, such as knowledge-primed neural networks, pathway- or ontology-constrained models, and interpretable structured neural networks;
• sharing examples or workflows built with edge2torch;
• proposing early ideas before turning them into feature requests;
• asking general usage questions that are not clearly bugs.

Please use GitHub Issues instead for:

• reproducible bugs;
• concrete feature requests;
• documentation problems;
• installation or packaging errors.

When asking a usage question, it helps to include:

• the backend you are using: feedforward, recurrent, or graphnn;
• a small example edge list;
• the behavior you expected;
• the relevant code snippet, if available.

Thanks for trying edge2torch. Feedback is especially useful while the public API is still stabilizing.