modules/wenet_extractor/transformer/cmvn.py

# This module is from [WeNet](https://github.com/wenet-e2e/wenet).

# ## Citations

# ```bibtex
# @inproceedings{yao2021wenet,
#   title={WeNet: Production oriented Streaming and Non-streaming End-to-End Speech Recognition Toolkit},
#   author={Yao, Zhuoyuan and Wu, Di and Wang, Xiong and Zhang, Binbin and Yu, Fan and Yang, Chao and Peng, Zhendong and Chen, Xiaoyu and Xie, Lei and Lei, Xin},
#   booktitle={Proc. Interspeech},
#   year={2021},
#   address={Brno, Czech Republic },
#   organization={IEEE}
# }

# @article{zhang2022wenet,
#   title={WeNet 2.0: More Productive End-to-End Speech Recognition Toolkit},
#   author={Zhang, Binbin and Wu, Di and Peng, Zhendong and Song, Xingchen and Yao, Zhuoyuan and Lv, Hang and Xie, Lei and Yang, Chao and Pan, Fuping and Niu, Jianwei},
#   journal={arXiv preprint arXiv:2203.15455},
#   year={2022}
# }
#

import torch


class GlobalCMVN(torch.nn.Module):
    def __init__(self, mean: torch.Tensor, istd: torch.Tensor, norm_var: bool = True):
        """
        Args:
            mean (torch.Tensor): mean stats
            istd (torch.Tensor): inverse std, std which is 1.0 / std
        """
        super().__init__()
        assert mean.shape == istd.shape
        self.norm_var = norm_var
        # The buffer can be accessed from this module using self.mean
        self.register_buffer("mean", mean)
        self.register_buffer("istd", istd)

    def forward(self, x: torch.Tensor):
        """
        Args:
            x (torch.Tensor): (batch, max_len, feat_dim)

        Returns:
            (torch.Tensor): normalized feature
        """
        x = x - self.mean
        if self.norm_var:
            x = x * self.istd
        return x