Multi-Model Chorus Recognition

Code and Dataset of ICANN2021 paper: Multi-Model Chorus Recognition for Improving Song Search

Dependency

• python 3.6+

• PyTorch 1.0+

• Transformers 2.6.0

• others

  • python_speech_features
  • Pandas

CHORD

You can download the CHOrus Recognition Dataset here and unzip to data/ file.

Pre-trained Language Model

We used BERT-wwm-ext, Chinese pre-trained language model
Related introduction and download link please refer to <u>Chinese-BERT-wwm</u>.

Train and Test

For training, you can run commands like this:

python train.py

For evaluation, the command may like this:

python test.py

Citation

If you use this code useful, please star our repo or consider citing:

@article{Wang2021MultiModalCR,
  title={Multi-Modal Chorus Recognition for Improving Song Search},
  author={Jiaan Wang and Zhixu Li and Binbin Gu and Tingyi Zhang and Qingsheng Liu and Zhigang Chen},
  journal={ArXiv},
  year={2021},
  volume={abs/2106.16153}
}