The weights of all models trained in the polyGNN companion paper are included in this PR. This means that users of this repository can now make predictions on new polymers using models that I trained for the paper. In total, these models were trained on over 30,000 polymer data points.
Tests have also been included in this PR to make sure that the model outputs stay constant, even as future changes to the codebase occur.
The weights of all models trained in the polyGNN companion paper are included in this PR. This means that users of this repository can now make predictions on new polymers using models that I trained for the paper. In total, these models were trained on over 30,000 polymer data points.
Tests have also been included in this PR to make sure that the model outputs stay constant, even as future changes to the codebase occur.