U %dO'@sddlmZmZddlZdgZejejdddZGdddejjZ Gd d d ejjZ Gd d d ejjZ Gd ddejjZ dS))OptionalTupleN Conformer)lengthsreturncCsF|jd}tt|}tj||j|jd||| dk}|S)Nr)devicedtype) shapeinttorchmaxitemZarangerrexpandZ unsqueeze)rZ batch_size max_lengthZ padding_maskr?/tmp/pip-unpacked-wheel-lbdmvq91/torchaudio/models/conformer.py_lengths_to_padding_mask s rc sHeZdZdZd eeeeeeddfdd Zej ej dd d Z Z S) _ConvolutionModuleaConformer convolution module. Args: input_dim (int): input dimension. num_channels (int): number of depthwise convolution layer input channels. depthwise_kernel_size (int): kernel size of depthwise convolution layer. dropout (float, optional): dropout probability. (Default: 0.0) bias (bool, optional): indicates whether to add bias term to each convolution layer. (Default: ``False``) use_group_norm (bool, optional): use GroupNorm rather than BatchNorm. (Default: ``False``) FN) input_dim num_channelsdepthwise_kernel_sizedropoutbiasuse_group_normrcst|dddks"tdtj||_tjtjj|d|ddd|dtjj ddtjj|||d|dd||d|rtjj d|dn tj |tj tjj||ddd|d tj ||_dS) Nr rz