Just remembered I already asked you about the CNNs, my bad! Oh wow, didn't realize that these were multi-layer pre-trained models. Also, started going through

Just remembered I already asked you about the CNNs, my bad! Oh wow, didn't realize that these were multi-layer pre-trained models. Also, started going through the QRNNs, they mention they've updated the AWD-LSTM Language model to use QRNNs, which is what your paper uses!