Args:

Reset the positional encodings.

Creates a binary mask for each sequence. Reference: https://discuss.pytorch.org/t/how-to-generate-variable-length-mask/23397/3 Arguments

_summary_ Args: audiopath (_type_): audio_path sampling_rate (_type_): sampling_rate Returns: _type_: _description_

_summary_ Args: prompt_wav (_type_): _description_ prompt_text (_type_): _description_ text (_type_): _de

Initialization. INPUT: - in_features: shape of the input - alpha: trainable parameter alpha is in

Checks the input shape and returns the number of input channels.

This function performs zero-padding on the time axis such that their lengths is unchanged after the convolution. Arguments --

GPT-2 implemented by the HuggingFace library.

_summary_ Args: cond_latents (_type_): speaker embedding text_inputs (_type_): text tokens prompt_tokens

Args: tokens: shape (b, t) Returns: token_embs: shape (b, t, c)

Compute attention context vector. Args: value (torch.Tensor): shape: (b, nh, t2, c//nh). scores (torch.Tensor): shape:

Args: query,key,value: shape (b, t, c) Returns: query,key,value: shape (b, nh, t, c//nh)

_summary_ Args: cond_latents (_type_): _description_ text_inputs (_type_): _description_ prompt_tokens (_

This function computes the number of elements to add for zero-padding. Arguments --------- L_in : int stride: int kernel_size : i