Fetching the paper…

SpeechT5: Unified-Modal Encoder-Decoder Pre-Training for Spoken Language Processing · Around