benedekrozemberczki/CapsGNN
A PyTorch implementation of Capsule Graph Neural Network for graph-level embeddings using routing mechanisms and attention.
Not currently ranked — collecting fresh signals.
star history
CapsGNN adapts Capsule Neural Networks to Graph Neural Networks for graph classification tasks. It extracts node features in capsule form and leverages routing mechanisms to aggregate information across multiple aspects at the graph level. An attention module incorporated into the architecture handles graphs of varying sizes by focusing on critical substructures, producing multiple embeddings per graph to capture diverse properties.
Frequently asked
- What is benedekrozemberczki/CapsGNN?
- A PyTorch implementation of Capsule Graph Neural Network for graph-level embeddings using routing mechanisms and attention.
- Is CapsGNN open source?
- Yes — benedekrozemberczki/CapsGNN is open source, released under the GPL-3.0 license.
- What language is CapsGNN written in?
- benedekrozemberczki/CapsGNN is primarily written in Python.
- How popular is CapsGNN?
- benedekrozemberczki/CapsGNN has 1.3k stars on GitHub.
- Where can I find CapsGNN?
- benedekrozemberczki/CapsGNN is on GitHub at https://github.com/benedekrozemberczki/CapsGNN.