This field marks the number of tasks for model
The number of classes to predict per task (only used when mode is ‘classification’)
The length of the initial atom feature vectors.
Number of attention heads in each GAT layer.
The dropout probability within each GAT layer.
The size for hidden representations in the output MLP predictor.
The dropout probability for the output MLP predictor.
A hyperparameter in LeakyReLU, which is the slope for negative values.