Closed pclucas14 closed 5 years ago
When using a RNN with multiple layers, the same hidden state was used for all layers. This does not affect the results in the paper (as I think all models were 1 layer), but still worth addressing and fixing.
When using a RNN with multiple layers, the same hidden state was used for all layers. This does not affect the results in the paper (as I think all models were 1 layer), but still worth addressing and fixing.