Textually Pretrained Speech Language Models
Fuente:
arXiv
Gespeichert in:
| Hauptverfasser: | , , , , , , , , , , , |
|---|---|
| Format: | Preprint |
| Veröffentlicht: |
2023
|
| Schlagworte: | |
| Online-Zugang: | |
| Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
| _version_ | 1866916109515489280 |
|---|---|
| author | Hassid, Michael Remez, Tal Nguyen, Tu Anh Gat, Itai Conneau, Alexis Kreuk, Felix Copet, Jade Defossez, Alexandre Synnaeve, Gabriel Dupoux, Emmanuel Schwartz, Roy Adi, Yossi |
| author_facet | Hassid, Michael Remez, Tal Nguyen, Tu Anh Gat, Itai Conneau, Alexis Kreuk, Felix Copet, Jade Defossez, Alexandre Synnaeve, Gabriel Dupoux, Emmanuel Schwartz, Roy Adi, Yossi |
| contents | Speech language models (SpeechLMs) process and generate acoustic data only, without textual supervision. In this work, we propose TWIST, a method for training SpeechLMs using a warm-start from a pretrained textual language models. We show using both automatic and human evaluations that TWIST outperforms a cold-start SpeechLM across the board. We empirically analyze the effect of different model design choices such as the speech tokenizer, the pretrained textual model, and the dataset size. We find that model and dataset scale both play an important role in constructing better-performing SpeechLMs. Based on our observations, we present the largest (to the best of our knowledge) SpeechLM both in terms of number of parameters and training data. We additionally introduce two spoken versions of the StoryCloze textual benchmark to further improve model evaluation and advance future research in the field. We make speech samples, code and models publicly available: https://pages.cs.huji.ac.il/adiyoss-lab/twist/ . |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2305_13009 |
| institution | arXiv |
| publishDate | 2023 |
| record_format | arxiv |
| spellingShingle | Textually Pretrained Speech Language Models Hassid, Michael Remez, Tal Nguyen, Tu Anh Gat, Itai Conneau, Alexis Kreuk, Felix Copet, Jade Defossez, Alexandre Synnaeve, Gabriel Dupoux, Emmanuel Schwartz, Roy Adi, Yossi Computation and Language Machine Learning Sound Audio and Speech Processing Speech language models (SpeechLMs) process and generate acoustic data only, without textual supervision. In this work, we propose TWIST, a method for training SpeechLMs using a warm-start from a pretrained textual language models. We show using both automatic and human evaluations that TWIST outperforms a cold-start SpeechLM across the board. We empirically analyze the effect of different model design choices such as the speech tokenizer, the pretrained textual model, and the dataset size. We find that model and dataset scale both play an important role in constructing better-performing SpeechLMs. Based on our observations, we present the largest (to the best of our knowledge) SpeechLM both in terms of number of parameters and training data. We additionally introduce two spoken versions of the StoryCloze textual benchmark to further improve model evaluation and advance future research in the field. We make speech samples, code and models publicly available: https://pages.cs.huji.ac.il/adiyoss-lab/twist/ . |
| title | Textually Pretrained Speech Language Models |
| topic | Computation and Language Machine Learning Sound Audio and Speech Processing |
| url | https://arxiv.org/abs/2305.13009 |