This is a collection of RNN classes built as PyTorch Modules.
The collection includes the following architectures:
- GRU
- DoubleGRU
- LSTM
- Nested LSTM
- Residual GRU
- Residual LSTM
- Residual Nested LSTM
- Residual RNN
Each has been evaluated on a simple memorization task. The goal was to predict the next binary digit in a random sequence given the k digits prior.