higher-order forward layer. There was a bug in the concatenate call
notebooks - related to the fact that calling a NODE now returns a t_span and trajectory. This requires writing out the operation within the training step rather than using a nn.Sequential to build the training step.
fixes bugs in
nn.Sequential
to build the training step.