I am trying to implement my own search space (mixture of Conv-based like ResNet, UNet combined with Transformer and RNN-based like the one right here
). Say, every is integrated within class
Model with hyperparameters-to-be-inputted in the __init__(self, n_layers, n_neurons, rnn_type,....)But the thing is that I don’t really know if I might mis-typed some part (such as * made into + or the skip connections are misplaced). In case the errors don’t pop up, how can I best test the implemented code?