1 paper
Jan Vanek, Josef Michalek, Jan Zelinka +1
Recently, recurrent neural networks have become state-of-the-art in acoustic modeling for automatic speech recognition. The long short-term memory (LSTM) units are the most popular…