Generated | |||
Prompt | Cold Init.-1.3B | TWIST-1.3B | TWIST-7B |
Prompt | Generated | Ground truth (resynthesized from speech tokens using a vocoder) |
Prompt | Generated | Ground truth (resynthesized from speech tokens using a vocoder) |
@misc{hassid2023textually, title={Textually Pretrained Speech Language Models}, author={Michael Hassid, Tal Remez, Tu Anh Nguyen, Itai Gat, Alexis Conneau, Felix Kreuk, Jade Copet, Alexandre Defossez, Gabriel Synnaeve, Emmanuel Dupoux, Roy Schwartz, Yossi Adi}, year={2023}, note = {{arXiv}:2305.13009} url = {https://arxiv.org/abs/2305.13009}, }