arnoegw

Please try with `pip install tensorflow-gnn==0.2.0` only (and nothing from git). If it still fails, please provide a complete repro - `WeightedSumConvolution` appears to not be from the TF-GNN library.

Hi @smith-co ! There was no further activity on this issue, and I can't detect an actionable defect report or feature request for the TF-GNN library, so I'm going to...

Are you trying to install multiple versions of tensorflow-gnn at once? Please install just one; we recommend the latest stable (at this time, 0.2.1).

Hi @TimSchmeier , thanks for reaching out! TF-GNN does support building GNNs suitable for link prediction/detection/classification. For a demonstration of an advanced model on a toy dataset, please see our...

HI @ly3106, thank you for your contribution! The api_docs are generated (every so often) from the docstrings embedded in the Python code. A direct edit to api_docs will be lost...

The basic advice for TF-GNN is to save the input graphs for training and validation in a TFRecord file, with each record being one GraphTensor serialized as a tf.Example proto....

Hi @mxmessi and @cjoverbay, thanks for reaching out! Did your investigations turn up anything we should do in the TF-GNN code to avoid this problem? Or has it just gone...

I suspect this issue (reported for tensorflow-gnn 0.1.0) is no longer current, so I'd like to close it. If you still see it happen, please reopen with detailed repro instructions....

TF-GNN has a number of good [Colab tutorials](https://github.com/tensorflow/gnn/blob/main/tensorflow_gnn/docs/guide/overview.md#colab-tutorials) for a while now. The reported file `examples/citeseer/graph_tensor_example.py` indeed is broken as @yarri-oss and @janpfeifer described, so I have deleted it with...

Thank you again for your contribution. Without the CLA, it was not possible for us to merge it into the repo. Meanwhile, it appears the problem has been fixed, based...