How to handle multi-graph tasks, that is, the training data is (B, N, C)

[hellomuyi]

Hello, does PyG support graph-level tasks, such as graph classification, which means that the shape of the input training data should be (batch_size, num_nodes, num_features). As is common practice in DNNs, the batch_size is num_graphs instead of num_nodes. If so, can you give me a link to an example?

[rusty1s]

Yes, PyG supports graph-level tasks, but the batching is a little bit different from traditional neural networks, see here and here. Here is a practical example of a graph-level classification task.