Open jeongwhanchoi opened 3 years ago
with return_sequence = False
we're doing many-to-one, e.g. we have EEG sequence and we want to understand whether it contains some potential or not.
with return_sequence = True
we're doing many-to-many classification, e.g. we have EEG sequence and we want to understand type of position at each point of time.
Or do you want shape examples?:)
Actually, it was inspired by NNs from this repo, maybe it will help more.
I found your repo via Kidger's NeuralCDE repo. I wonder how your seq2seq works with CDE.
I would like to know which kind of sequence like below does your
return_sequence
supports? (e.g. many-to-one or many-to-many sequence)