DeepLearningExamples/CUDA-Optimized/FastSpeech/waveglow/loss_function.py

# *****************************************************************************
#  Copyright (c) 2018, NVIDIA CORPORATION.  All rights reserved.
#
#  Redistribution and use in source and binary forms, with or without
#  modification, are permitted provided that the following conditions are met:
#      * Redistributions of source code must retain the above copyright
#        notice, this list of conditions and the following disclaimer.
#      * Redistributions in binary form must reproduce the above copyright
#        notice, this list of conditions and the following disclaimer in the
#        documentation and/or other materials provided with the distribution.
#      * Neither the name of the NVIDIA CORPORATION nor the
#        names of its contributors may be used to endorse or promote products
#        derived from this software without specific prior written permission.
#
#  THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" AND
#  ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED
#  WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE
#  DISCLAIMED. IN NO EVENT SHALL NVIDIA CORPORATION BE LIABLE FOR ANY
#  DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES
#  (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
#  LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND
#  ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
#  (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS
#  SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
#
# *****************************************************************************

import torch

class WaveGlowLoss(torch.nn.Module):
    def __init__(self, sigma=1.0):
        super(WaveGlowLoss, self).__init__()
        self.sigma = sigma

    def forward(self, model_output, clean_audio):
        # clean_audio is unused;
        z, log_s_list, log_det_W_list = model_output
        for i, log_s in enumerate(log_s_list):
            if i == 0:
                log_s_total = torch.sum(log_s)
                log_det_W_total = log_det_W_list[i]
            else:
                log_s_total = log_s_total + torch.sum(log_s)
                log_det_W_total += log_det_W_list[i]

        loss = torch.sum(
            z * z) / (2 * self.sigma * self.sigma) - log_s_total - log_det_W_total  # noqa: E501
        return loss / (z.size(0) * z.size(1) * z.size(2))