main/library/algorithm/refinegan.py

import os
import sys
import math
import torch

import numpy as np
import torch.nn.functional as F
import torch.utils.checkpoint as checkpoint

from torch.nn.utils.parametrizations import weight_norm
from torch.nn.utils.parametrize import remove_parametrizations

sys.path.append(os.getcwd())

from .commons import get_padding

class ResBlock(torch.nn.Module):
    def __init__(self, *, in_channels, out_channels, kernel_size = 7, dilation = (1, 3, 5), leaky_relu_slope = 0.2):
        super(ResBlock, self).__init__()
        self.leaky_relu_slope = leaky_relu_slope
        self.in_channels = in_channels
        self.out_channels = out_channels

        self.convs1 = torch.nn.ModuleList([weight_norm(torch.nn.Conv1d(in_channels=in_channels if idx == 0 else out_channels, out_channels=out_channels, kernel_size=kernel_size, stride=1, dilation=d, padding=get_padding(kernel_size, d))) for idx, d in enumerate(dilation)])
        self.convs1.apply(self.init_weights)
        
        self.convs2 = torch.nn.ModuleList([weight_norm(torch.nn.Conv1d(in_channels=out_channels, out_channels=out_channels, kernel_size=kernel_size, stride=1, dilation=d, padding=get_padding(kernel_size, d))) for _, d in enumerate(dilation)])
        self.convs2.apply(self.init_weights)

    def forward(self, x):
        for idx, (c1, c2) in enumerate(zip(self.convs1, self.convs2)):
            xt = c2(F.leaky_relu(c1(F.leaky_relu(x, self.leaky_relu_slope)), self.leaky_relu_slope))
            x = (xt + x) if idx != 0 or self.in_channels == self.out_channels else xt

        return x
    
    def remove_parametrizations(self):
        for c1, c2 in zip(self.convs1, self.convs2):
            remove_parametrizations(c1)
            remove_parametrizations(c2)

    def init_weights(self, m):
        if type(m) == torch.nn.Conv1d:
            m.weight.data.normal_(0, 0.01)
            m.bias.data.fill_(0.0)

class AdaIN(torch.nn.Module):
    def __init__(self, *, channels, leaky_relu_slope = 0.2):
        super().__init__()
        self.weight = torch.nn.Parameter(torch.ones(channels))
        self.activation = torch.nn.LeakyReLU(leaky_relu_slope)

    def forward(self, x):
        return self.activation(x + (torch.randn_like(x) * self.weight[None, :, None]))
    
class ParallelResBlock(torch.nn.Module):
    def __init__(self, *, in_channels, out_channels, kernel_sizes = (3, 7, 11), dilation = (1, 3, 5), leaky_relu_slope = 0.2):
        super().__init__()
        self.in_channels = in_channels
        self.out_channels = out_channels
        self.input_conv = torch.nn.Conv1d(in_channels=in_channels, out_channels=out_channels, kernel_size=7, stride=1, padding=3)
        self.blocks = torch.nn.ModuleList([torch.nn.Sequential(AdaIN(channels=out_channels), ResBlock(in_channels=out_channels, out_channels=out_channels, kernel_size=kernel_size, dilation=dilation, leaky_relu_slope=leaky_relu_slope), AdaIN(channels=out_channels)) for kernel_size in kernel_sizes])

    def forward(self, x):
        return torch.mean(torch.stack([block(self.input_conv(x)) for block in self.blocks]), dim=0)
    
    def remove_parametrizations(self):
        for block in self.blocks:
            block[1].remove_parametrizations()

class SineGenerator(torch.nn.Module):
    def __init__(self, samp_rate, harmonic_num=0, sine_amp=0.1, noise_std=0.003, voiced_threshold=0):
        super(SineGenerator, self).__init__()
        self.sine_amp = sine_amp
        self.noise_std = noise_std
        self.harmonic_num = harmonic_num
        self.dim = self.harmonic_num + 1
        self.sampling_rate = samp_rate
        self.voiced_threshold = voiced_threshold
        self.merge = torch.nn.Sequential(torch.nn.Linear(self.dim, 1, bias=False), torch.nn.Tanh())   

    def _f02uv(self, f0):
        return torch.ones_like(f0) * (f0 > self.voiced_threshold)
    
    def _f02sine(self, f0_values):
        rad_values = (f0_values / self.sampling_rate) % 1
        rand_ini = torch.rand(f0_values.shape[0], f0_values.shape[2], device=f0_values.device)

        rand_ini[:, 0] = 0
        rad_values[:, 0, :] = rad_values[:, 0, :] + rand_ini

        tmp_over_one = torch.cumsum(rad_values, 1) % 1
        tmp_over_one_idx = (tmp_over_one[:, 1:, :] - tmp_over_one[:, :-1, :]) < 0

        cumsum_shift = torch.zeros_like(rad_values)
        cumsum_shift[:, 1:, :] = tmp_over_one_idx * -1.0

        return torch.sin(torch.cumsum(rad_values + cumsum_shift, dim=1) * 2 * np.pi)
    
    def forward(self, f0):
        with torch.no_grad():
            f0_buf = torch.zeros(f0.shape[0], f0.shape[1], self.dim, device=f0.device)
            f0_buf[:, :, 0] = f0[:, :, 0]

            for idx in np.arange(self.harmonic_num):
                f0_buf[:, :, idx + 1] = f0_buf[:, :, 0] * (idx + 2)

            sine_waves = self._f02sine(f0_buf) * self.sine_amp
            uv = self._f02uv(f0)

            sine_waves = sine_waves * uv + (uv * self.noise_std + (1 - uv) * self.sine_amp / 3) * torch.randn_like(sine_waves)
            sine_waves = sine_waves - sine_waves.mean(dim=1, keepdim=True)

        return self.merge(sine_waves)
    
class RefineGANGenerator(torch.nn.Module):
    def __init__(self, *, sample_rate = 44100, upsample_rates = (8, 8, 2, 2), leaky_relu_slope = 0.2, num_mels = 128, gin_channels = 256, checkpointing = False, upsample_initial_channel = 512):
        super().__init__()
        self.upsample_rates = upsample_rates
        self.checkpointing = checkpointing
        self.leaky_relu_slope = leaky_relu_slope
        self.upp = np.prod(upsample_rates)
        self.m_source = SineGenerator(sample_rate)
        self.pre_conv = weight_norm(torch.nn.Conv1d(in_channels=1, out_channels=upsample_initial_channel // 2, kernel_size=7, stride=1, padding=3, bias=False))
        channels = upsample_initial_channel
        self.downsample_blocks = torch.nn.ModuleList([])

        stride_f0s = [math.prod(upsample_rates[i + 1 :]) if i + 1 < len(upsample_rates) else 1 for i in range(len(upsample_rates))]

        for i, _ in enumerate(upsample_rates):
            stride = stride_f0s[i]
            kernel = 1 if stride == 1 else stride * 2 - stride % 2

            self.downsample_blocks.append(torch.nn.Conv1d(in_channels=1, out_channels=channels // 2 ** (i + 2), kernel_size=kernel, stride=stride, padding=0 if stride == 1 else (kernel - stride) // 2))

        self.mel_conv = weight_norm(torch.nn.Conv1d(in_channels=num_mels, out_channels=channels // 2, kernel_size=7, stride=1, padding=3))
        if gin_channels != 0: self.cond = torch.nn.Conv1d(256, channels // 2, 1)

        self.upsample_blocks = torch.nn.ModuleList([])
        self.upsample_conv_blocks = torch.nn.ModuleList([])
        self.filters = torch.nn.ModuleList([])

        for rate in upsample_rates:
            new_channels = channels // 2
            self.upsample_blocks.append(torch.nn.Upsample(scale_factor=rate, mode="linear"))

            low_pass = torch.nn.Conv1d(channels, channels, kernel_size=15, padding=7,  groups=channels, bias=False)
            low_pass.weight.data.fill_(1.0 / 15)
            
            self.filters.append(low_pass)

            self.upsample_conv_blocks.append(ParallelResBlock(in_channels=channels + channels // 4, out_channels=new_channels, kernel_sizes=(3, 7, 11), dilation=(1, 3, 5), leaky_relu_slope=leaky_relu_slope))
            channels = new_channels

        self.conv_post = weight_norm(torch.nn.Conv1d(in_channels=channels, out_channels=1, kernel_size=7, stride=1, padding=3))

    def forward(self, mel, f0, g = None):
        har_source = self.m_source(f0.transpose(1, 2)).transpose(1, 2)
        x = F.interpolate(self.pre_conv(har_source), size=mel.shape[-1], mode="linear")

        mel = self.mel_conv(mel)
        if g is not None: mel += self.cond(g)

        x = torch.cat([mel, x], dim=1)

        for ups, res, down, flt in zip(self.upsample_blocks, self.upsample_conv_blocks, self.downsample_blocks, self.filters):
            x = checkpoint(res, torch.cat([checkpoint(flt, checkpoint(ups, x, use_reentrant=False), use_reentrant=False), down(har_source)], dim=1), use_reentrant=False) if self.training and self.checkpointing else res(torch.cat([flt(ups(x)), down(har_source)], dim=1))

        return torch.tanh_(self.conv_post(F.leaky_relu_(x, self.leaky_relu_slope)))
    
    def remove_parametrizations(self):
        remove_parametrizations(self.source_conv)
        remove_parametrizations(self.mel_conv)
        remove_parametrizations(self.conv_post)

        for block in self.downsample_blocks:
            block[1].remove_parametrizations()

        for block in self.upsample_conv_blocks:
            block.remove_parametrizations()