w2v-cif-bert

Code for paper: Efficiently Fusing Pretrained Acoustic and Linguistic Encoders for Low-resource Speech Recognition

We only provide key files of our model, w2v-cif-bert, which can be reimplement based on fairseq. If you have any questions on the reimplementation, please consult yicheng2016@ia.ac.cn.

update

• 2021.5.14 Following others' requirement of the baselines used in our paper, we reveal the implementation of w2v-seq2seq and w2v-nar (relative scripts are in baselines/*). NOTE: These codes are based on the out-of-date commit（23d8502bdde88a3e58e0910e2ee49834f8478b39 upstream/master）of Fairseq without testing in the new one.

Please cite as:

@article{yi2021efficiently,
  title={Efficiently Fusing Pretrained Acoustic and Linguistic Encoders for Low-Resource Speech Recognition},
  author={Yi, Cheng and Zhou, Shiyu and Xu, Bo},
  journal={IEEE Signal Processing Letters},
  volume={28},
  pages={788--792},
  year={2021},
  publisher={IEEE}
}