Improving Joint Speech-Text Representations Without Alignment

Open in new window