OpenNE is an open-source toolkit for network representation learning (NRL), offering a unified interface for various network embedding models and scalable training options. It targets researchers and practitioners in graph analysis, providing implementations of popular models like DeepWalk, node2vec, and GCN, with a focus on incorporating text attributes via TADW for enhanced node classification.

How It Works

OpenNE implements multiple network embedding algorithms, including DeepWalk, LINE, node2vec, GraRep, TADW, GCN, HOPE, GF, SDNE, and Laplacian Eigenmaps. It standardizes input/output interfaces and leverages TensorFlow for GPU-accelerated training. The toolkit emphasizes reproducibility of results from original papers and includes evaluation metrics like Micro-F1 and Macro-F1 for node classification tasks. TADW is highlighted for its ability to integrate node text attributes, improving performance on tasks like node classification.

Quick Start & Requirements

• Install via pip install -r requirements.txt and cd src; python setup.py install.
• Requires Python. Specific CPU details are provided for benchmark datasets (Cora, Wiki, BlogCatalog).
• Official documentation and examples are available within the repository.

Highlighted Details

• Implements 10+ network embedding models, including TADW for text attribute integration.
• Provides GPU-accelerated training via TensorFlow.
• Includes evaluation scripts for node classification tasks, reporting Micro-F1 and Macro-F1.
• Benchmarks show OpenNE implementations often match or exceed original paper results, with TADW outperforming DeepWalk when text features are used.

Maintenance & Community

OpenNE is a sub-project of OpenSKL. No specific community links (Discord, Slack) or active contributor information are detailed in the README.

Licensing & Compatibility

The README mentions MIT license for OpenKE pre-trained embeddings, but the license for OpenNE itself is not explicitly stated. Compatibility for commercial use or closed-source linking is not specified.

Limitations & Caveats

The README does not explicitly state the license for the OpenNE toolkit itself, which could impact commercial use. Community support channels are not detailed.