Masked Audio Generation using a Single Non-Autoregressive Transformer
Fuente:
arXiv
Saved in:
| Main Authors: | , , , , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2024
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866917604277354496 |
|---|---|
| author | Ziv, Alon Gat, Itai Lan, Gael Le Remez, Tal Kreuk, Felix Défossez, Alexandre Copet, Jade Synnaeve, Gabriel Adi, Yossi |
| author_facet | Ziv, Alon Gat, Itai Lan, Gael Le Remez, Tal Kreuk, Felix Défossez, Alexandre Copet, Jade Synnaeve, Gabriel Adi, Yossi |
| contents | We introduce MAGNeT, a masked generative sequence modeling method that operates directly over several streams of audio tokens. Unlike prior work, MAGNeT is comprised of a single-stage, non-autoregressive transformer. During training, we predict spans of masked tokens obtained from a masking scheduler, while during inference we gradually construct the output sequence using several decoding steps. To further enhance the quality of the generated audio, we introduce a novel rescoring method in which, we leverage an external pre-trained model to rescore and rank predictions from MAGNeT, which will be then used for later decoding steps. Lastly, we explore a hybrid version of MAGNeT, in which we fuse between autoregressive and non-autoregressive models to generate the first few seconds in an autoregressive manner while the rest of the sequence is being decoded in parallel. We demonstrate the efficiency of MAGNeT for the task of text-to-music and text-to-audio generation and conduct an extensive empirical evaluation, considering both objective metrics and human studies. The proposed approach is comparable to the evaluated baselines, while being significantly faster (x7 faster than the autoregressive baseline). Through ablation studies and analysis, we shed light on the importance of each of the components comprising MAGNeT, together with pointing to the trade-offs between autoregressive and non-autoregressive modeling, considering latency, throughput, and generation quality. Samples are available on our demo page https://pages.cs.huji.ac.il/adiyoss-lab/MAGNeT. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2401_04577 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | Masked Audio Generation using a Single Non-Autoregressive Transformer Ziv, Alon Gat, Itai Lan, Gael Le Remez, Tal Kreuk, Felix Défossez, Alexandre Copet, Jade Synnaeve, Gabriel Adi, Yossi Sound Artificial Intelligence Machine Learning Audio and Speech Processing We introduce MAGNeT, a masked generative sequence modeling method that operates directly over several streams of audio tokens. Unlike prior work, MAGNeT is comprised of a single-stage, non-autoregressive transformer. During training, we predict spans of masked tokens obtained from a masking scheduler, while during inference we gradually construct the output sequence using several decoding steps. To further enhance the quality of the generated audio, we introduce a novel rescoring method in which, we leverage an external pre-trained model to rescore and rank predictions from MAGNeT, which will be then used for later decoding steps. Lastly, we explore a hybrid version of MAGNeT, in which we fuse between autoregressive and non-autoregressive models to generate the first few seconds in an autoregressive manner while the rest of the sequence is being decoded in parallel. We demonstrate the efficiency of MAGNeT for the task of text-to-music and text-to-audio generation and conduct an extensive empirical evaluation, considering both objective metrics and human studies. The proposed approach is comparable to the evaluated baselines, while being significantly faster (x7 faster than the autoregressive baseline). Through ablation studies and analysis, we shed light on the importance of each of the components comprising MAGNeT, together with pointing to the trade-offs between autoregressive and non-autoregressive modeling, considering latency, throughput, and generation quality. Samples are available on our demo page https://pages.cs.huji.ac.il/adiyoss-lab/MAGNeT. |
| title | Masked Audio Generation using a Single Non-Autoregressive Transformer |
| topic | Sound Artificial Intelligence Machine Learning Audio and Speech Processing |
| url | https://arxiv.org/abs/2401.04577 |