1 paper
Andrea Fasoli, Chia-Yu Chen, Mauricio Serrano +9
We investigate the impact of aggressive low-precision representations of weights and activations in two families of large LSTM-based architectures for Automatic Speech Recognition…