spektral
spektral copied to clipboard
Graph Neural Networks with Keras and Tensorflow 2.
currently the dataset containers only support for attributes (i.e, node features, adjacency matrix, edge features & y). Is there any way to add or concatenate the graph level features after...
If you look at the original implementation, BN should be before the relu https://github.com/weihua916/powerful-gnns/blob/9a2ce8ac3e99278307093a464a95caf0fb04b602/models/mlp.py#L48
Dear Devs, thank you for this library. I am currently writing my model using it, but I am a bit confused about the usage. I am working with sparse matrices...
I am unable to `import spektral` package at the beginning of my script. Simply returns `zsh: illegal hardware instruction python`. I have a MacAir with M1 and running in a...
Hi Daniele and all, I'm trying to use DisjointLoader in order to train a GCN to make a node-level prediction. My own dataset consists in a list of graphs with...
Hi! I performed a node classification task using masking on batch mode. It runs correctly. However, I also need to get more metrics of the model by using the classification...
Hi Daniele, I was wondering if it's possible to apply sample weights (as in, the Keras [`sample_weight`](https://github.com/tensorflow/tensorflow/blob/c455215395f312cb7c949d386cf7908d456052c8/tensorflow/python/keras/engine/training.py#L890)) when training with the `DisjointLoader`. I cannot find a way to get it...
Tks about your code. Could you add Unspervised GraphSage and embedding output when model is trained
The idea is to extend graph object by add a timestamp such that it becomes the GraphSnapshot object. Then a DynamicGraph is a sequence of graph snapshots sorted by the...
**Spektral** certainly seems promising and easy to get started. Also, it has some good examples for predictive tasks. However, **Spektral** lacks examples of generative models. _That said, other libraries on...