v0.2.1

DeepSNAP 0.2.1 Release Notes

• Use the collections.abc to be compatible with the PyTorch 1.9.0.
• Minor modification to the hetero_gnn.py.
• Use GitHub Actions for testing.
• Some other minor modifications.

v0.2.0

• Fixed bugs in Graph and HeteroGraph classes.
• Added tensor backend for both the homogeneous and heterogeneous graphs.
• Added SnapX backend for the undirected homogeneous graph.
• New examples and new Colab notebooks.
• Improved HeteroSAGEConv.
• Updated Heterogeneous GNN Functions.
• Updated documentation and code comments.
• New benchmark for transformation.
• Other minor updates and bug fixes.

Transform function improvement

• Documentation for transform functions
• Support transform functions for Batch that return multiple graphs (multiple batches of transformed graph will be returned)