o i@s6ddlZddlmZddlmZGdddejZdS)N)nn)ISTFTc sReZdZdZd dedededeffdd Zejj d ej d ej fd d Z Z S) ISTFTHeada ISTFT Head module for predicting STFT complex coefficients. Args: dim (int): Hidden dimension of the model. n_fft (int): Size of Fourier transform. hop_length (int): The distance between neighboring sliding window frames, which should align with the resolution of the input features. padding (str, optional): Type of padding. Options are "center" or "same". Defaults to "same". centerdimn_fft hop_lengthpaddingcs8t|d}tj|||_t||||d|_dS)N)rr win_lengthr )super__init__torchrLinearoutristft)selfrrr r out_dim __class__G/home/ubuntu/.local/lib/python3.10/site-packages/soprano/vocos/heads.pyrs zISTFTHead.__init__xreturncCsv||dddd}|jddd\}}t|}tj|dd}t|}t|}||d|}||}|S)ay Forward pass of the ISTFTHead module. Args: x (Tensor): Input tensor of shape (B, L, H), where B is the batch size, L is the sequence length, and H denotes the model dimension. Returns: Tensor: Reconstructed time-domain audio signal of shape (B, T), where T is the length of the output signal. rr )rgY@)maxy?) r transposechunkrexpclipcossinr)rrmagpySaudiorrrforwards     zISTFTHead.forward)r) __name__ __module__ __qualname____doc__intstrrrcompilerdisableTensorr' __classcell__rrrrrs   r)rr spectral_opsrModulerrrrrs