Retriever: Learning Content-Style Representation as a Token-Level Bipartite Graph
Dacheng Yin*, Xuanchi Ren*, Chong Luo, Yuwang Wang, Zhiwei Xiong and Wenjun Zeng
ICLR 2022

[Paper] [arXiv] [Demo Page]

🔲 Merge and Clean Code

We are cleaning and merging the code and hope to release it very soon.

For the vision part, we provide a sample (uncleaned) code here.

For the audio part, we provide a sample (uncleaned) code here.

In this repo, we propose an unsupervised and modality-agnostic content-style disentanglement framework: Retriever. We demonstrate that our learned representation can benefit zero-shot voice conversion, co-part segmentation, and style transfer.

@inproceedings{yin2022Retriever,
  title   = {Retriever: Learning Content-Style Representation as a Token-Level Bipartite Graph},
  author  = {Yin, Dacheng and Ren, Xuanchi and Luo, Chong and Wang, Yuwang, and Xiong, Zhiwei, and Zeng, Wenjun},
  booktitle = {ICLR},
  year    = {2022}
}