GraphVite is an open-source graph embedding system, designed for high-speed and large-scale training. The goal of this system is to provide a general and high-performance framework for a broad family of embedding methods, which facilitates research and deployment of graph learning algorithms.
Through the development of GraphVite, it is always targeted to solve extremely large graphs on CPU-GPU hybrid architectures. Tailored to the nature of graph embeddings, GraphVite is distinguished from other libraries by fast training and linear scalability. It is even able to work under very limited GPU memory, where other systems might fail.
With GraphVite, it is easy to reproduce cutting-edge models of many applications, deploy them on large real-world datasets, and develop new models for graph representation learning.
GraphVite is released under the Apache License 2.0.
If you have any question about the library, please open an issue in the GitHub repo.