shredx.modules.rnn

Recurrent neural network encoders for sequence modeling.

Implements GRU, LSTM, and MLP encoders compatible with an encoder–decoder architecture. Also implements MOE-GRU and MOE-LSTM encoders.

Classes

GRUEncoder(input_size, hidden_size, ...[, ...])	GRU encoder for sequence-to-sequence modeling.
LSTMEncoder(input_size, hidden_size, ...[, ...])	LSTM encoder for sequence-to-sequence modeling.
MOEGRUEncoder(input_size, hidden_size, ...)	Mixture of Experts GRU with SINDy layer forecasting.
MOELSTMEncoder(input_size, hidden_size, ...)	Mixture of Experts LSTM with SINDy layer forecasting.
SINDyLossGRUEncoder(input_size, hidden_size, ...)	GRU encoder with SINDy loss regularization.
SINDyLossLSTMEncoder(input_size, ...[, device])	LSTM encoder with SINDy loss regularization.