1 paper · 1 filter
Benjamin Milde, Chris Biemann
The Sparsespeech model is an unsupervised acoustic model that can generate discrete pseudo-labels for untranscribed speech. We extend the Sparsespeech model to allow for sampling o…