model.py

import math
import random
import torch
from torch import nn
from torch.nn import functional as F
import numpy as np

from models.stylegan2.op import FusedLeakyReLU, fused_leaky_relu, upfirdn2d


class PixelNorm(nn.Module):
    def __init__(self):
        super().__init__()

    def forward(self, input):
        return input * torch.rsqrt(torch.mean(input ** 2, dim=1, keepdim=True) + 1e-8)


def make_kernel(k):
    k = torch.tensor(k, dtype=torch.float32)

    if k.ndim == 1:
        k = k[None, :] * k[:, None]

    k /= k.sum()

    return k


class Upsample(nn.Module):
    def __init__(self, kernel, factor=2):
        super().__init__()

        self.factor = factor
        kernel = make_kernel(kernel) * (factor ** 2)
        self.register_buffer('kernel', kernel)

        p = kernel.shape[0] - factor

        pad0 = (p + 1) // 2 + factor - 1
        pad1 = p // 2

        self.pad = (pad0, pad1)

    def forward(self, input):
        out = upfirdn2d(input, self.kernel, up=self.factor, down=1, pad=self.pad)

        return out


class Downsample(nn.Module):
    def __init__(self, kernel, factor=2):
        super().__init__()

        self.factor = factor
        kernel = make_kernel(kernel)
        self.register_buffer('kernel', kernel)

        p = kernel.shape[0] - factor

        pad0 = (p + 1) // 2
        pad1 = p // 2

        self.pad = (pad0, pad1)

    def forward(self, input):
        out = upfirdn2d(input, self.kernel, up=1, down=self.factor, pad=self.pad)

        return out


class Blur(nn.Module):
    def __init__(self, kernel, pad, upsample_factor=1):
        super().__init__()

        kernel = make_kernel(kernel)

        if upsample_factor > 1:
            kernel = kernel * (upsample_factor ** 2)

        self.register_buffer('kernel', kernel)

        self.pad = pad

    def forward(self, input):
        out = upfirdn2d(input, self.kernel, pad=self.pad)

        return out


class EqualConv2d(nn.Module):
    def __init__(
            self, in_channel, out_channel, kernel_size, stride=1, padding=0, bias=True, dilation=1 ## modified
    ):
        super().__init__()

        self.weight = nn.Parameter(
            torch.randn(out_channel, in_channel, kernel_size, kernel_size)
        )
        self.scale = 1 / math.sqrt(in_channel * kernel_size ** 2)

        self.stride = stride
        self.padding = padding
        self.dilation = dilation ## modified

        if bias:
            self.bias = nn.Parameter(torch.zeros(out_channel))

        else:
            self.bias = None

    def forward(self, input):
        out = F.conv2d(
            input,
            self.weight * self.scale,
            bias=self.bias,
            stride=self.stride,
            padding=self.padding,
            dilation=self.dilation,  ## modified
        )

        return out

    def __repr__(self):
        return (
            f"{self.__class__.__name__}({self.weight.shape[1]}, {self.weight.shape[0]},"
            f" {self.weight.shape[2]}, stride={self.stride}, padding={self.padding}, dilation={self.dilation})" ## modified
        )


class EqualLinear(nn.Module):
    def __init__(
            self, in_dim, out_dim, bias=True, bias_init=0, lr_mul=1, activation=None
    ):
        super().__init__()

        self.weight = nn.Parameter(torch.randn(out_dim, in_dim).div_(lr_mul))

        if bias:
            self.bias = nn.Parameter(torch.zeros(out_dim).fill_(bias_init))

        else:
            self.bias = None

        self.activation = activation

        self.scale = (1 / math.sqrt(in_dim)) * lr_mul
        self.lr_mul = lr_mul

    def forward(self, input):
        if self.activation:
            out = F.linear(input, self.weight * self.scale)
            out = fused_leaky_relu(out, self.bias * self.lr_mul)

        else:
            out = F.linear(
                input, self.weight * self.scale, bias=self.bias * self.lr_mul
            )

        return out

    def __repr__(self):
        return (
            f'{self.__class__.__name__}({self.weight.shape[1]}, {self.weight.shape[0]})'
        )


class ScaledLeakyReLU(nn.Module):
    def __init__(self, negative_slope=0.2):
        super().__init__()

        self.negative_slope = negative_slope

    def forward(self, input):
        out = F.leaky_relu(input, negative_slope=self.negative_slope)

        return out * math.sqrt(2)


class ModulatedConv2d(nn.Module):
    def __init__(
            self,
            in_channel,
            out_channel,
            kernel_size,
            style_dim,
            demodulate=True,
            upsample=False,
            downsample=False,
            blur_kernel=[1, 3, 3, 1],
            dilation=1, ##### modified
    ):
        super().__init__()

        self.eps = 1e-8
        self.kernel_size = kernel_size
        self.in_channel = in_channel
        self.out_channel = out_channel
        self.upsample = upsample
        self.downsample = downsample
        self.dilation = dilation ##### modified

        if upsample:
            factor = 2
            p = (len(blur_kernel) - factor) - (kernel_size - 1)
            pad0 = (p + 1) // 2 + factor - 1
            pad1 = p // 2 + 1

            self.blur = Blur(blur_kernel, pad=(pad0, pad1), upsample_factor=factor)
            
            # to simulate transconv + blur
            # we use dilated transposed conv with blur kernel as weight + dilated transconv 
            if dilation > 1: ##### modified
                blur_weight = torch.randn(1, 1, 3, 3) * 0  + 1
                blur_weight[:,:,0,1] = 2
                blur_weight[:,:,1,0] = 2
                blur_weight[:,:,1,2] = 2
                blur_weight[:,:,2,1] = 2
                blur_weight[:,:,1,1] = 4
                blur_weight = blur_weight / 16.0 
                self.register_buffer("blur_weight", blur_weight)

        if downsample:
            factor = 2
            p = (len(blur_kernel) - factor) + (kernel_size - 1)
            pad0 = (p + 1) // 2
            pad1 = p // 2

            self.blur = Blur(blur_kernel, pad=(pad0, pad1))

        fan_in = in_channel * kernel_size ** 2
        self.scale = 1 / math.sqrt(fan_in)
        self.padding = kernel_size // 2 + dilation - 1 ##### modified

        self.weight = nn.Parameter(
            torch.randn(1, out_channel, in_channel, kernel_size, kernel_size)
        )

        self.modulation = EqualLinear(style_dim, in_channel, bias_init=1)

        self.demodulate = demodulate

    def __repr__(self):
        return (
            f'{self.__class__.__name__}({self.in_channel}, {self.out_channel}, {self.kernel_size}, '
            f'upsample={self.upsample}, downsample={self.downsample})'
        )

    def forward(self, input, style):
        batch, in_channel, height, width = input.shape

        style = self.modulation(style).view(batch, 1, in_channel, 1, 1)
        weight = self.scale * self.weight * style

        if self.demodulate:
            demod = torch.rsqrt(weight.pow(2).sum([2, 3, 4]) + 1e-8)
            weight = weight * demod.view(batch, self.out_channel, 1, 1, 1)

        weight = weight.view(
            batch * self.out_channel, in_channel, self.kernel_size, self.kernel_size
        )

        if self.upsample:
            input = input.view(1, batch * in_channel, height, width)
            weight = weight.view(
                batch, self.out_channel, in_channel, self.kernel_size, self.kernel_size
            )
            weight = weight.transpose(1, 2).reshape(
                batch * in_channel, self.out_channel, self.kernel_size, self.kernel_size
            )
            
            if self.dilation > 1: ##### modified
                # to simulate out = self.blur(out)
                out = F.conv_transpose2d(
                    input, self.blur_weight.repeat(batch*in_channel,1,1,1), padding=0, groups=batch*in_channel, dilation=self.dilation//2)
                # to simulate the next line
                out = F.conv_transpose2d(
                    out, weight, padding=self.dilation, groups=batch, dilation=self.dilation//2)
                _, _, height, width = out.shape
                out = out.view(batch, self.out_channel, height, width)
                return out            
            
            out = F.conv_transpose2d(input, weight, padding=0, stride=2, groups=batch)
            _, _, height, width = out.shape
            out = out.view(batch, self.out_channel, height, width)
            out = self.blur(out)

        elif self.downsample:
            input = self.blur(input)
            _, _, height, width = input.shape
            input = input.view(1, batch * in_channel, height, width)
            out = F.conv2d(input, weight, padding=0, stride=2, groups=batch)
            _, _, height, width = out.shape
            out = out.view(batch, self.out_channel, height, width)

        else:
            input = input.view(1, batch * in_channel, height, width)
            out = F.conv2d(input, weight, padding=self.padding, groups=batch, dilation=self.dilation)  ##### modified
            _, _, height, width = out.shape
            out = out.view(batch, self.out_channel, height, width)

        return out


class NoiseInjection(nn.Module):
    def __init__(self):
        super().__init__()

        self.weight = nn.Parameter(torch.zeros(1))

    def forward(self, image, noise=None):
        if noise is None:
            batch, _, height, width = image.shape
            noise = image.new_empty(batch, 1, height, width).normal_()
        else:  ##### modified, to make the resolution matches
            batch, _, height, width = image.shape
            _, _, height1, width1 = noise.shape
            if height != height1 or width != width1:
                noise = F.adaptive_avg_pool2d(noise, (height, width))

        return image + self.weight * noise


class ConstantInput(nn.Module):
    def __init__(self, channel, size=4):
        super().__init__()

        self.input = nn.Parameter(torch.randn(1, channel, size, size))

    def forward(self, input):
        batch = input.shape[0]
        out = self.input.repeat(batch, 1, 1, 1)

        return out


class StyledConv(nn.Module):
    def __init__(
            self,
            in_channel,
            out_channel,
            kernel_size,
            style_dim,
            upsample=False,
            blur_kernel=[1, 3, 3, 1],
            demodulate=True,
            dilation=1,  ##### modified
    ):
        super().__init__()

        self.conv = ModulatedConv2d(
            in_channel,
            out_channel,
            kernel_size,
            style_dim,
            upsample=upsample,
            blur_kernel=blur_kernel,
            demodulate=demodulate,
            dilation=dilation,  ##### modified
        )

        self.noise = NoiseInjection()
        self.activate = FusedLeakyReLU(out_channel)

    def forward(self, input, style, noise=None):
        out = self.conv(input, style)
        out = self.noise(out, noise=noise)
        out = self.activate(out)

        return out


class ToRGB(nn.Module):
    def __init__(self, in_channel, style_dim, upsample=True, blur_kernel=[1, 3, 3, 1], dilation=1):  ##### modified
        super().__init__()

        if upsample:
            self.upsample = Upsample(blur_kernel)

        self.conv = ModulatedConv2d(in_channel, 3, 1, style_dim, demodulate=False)
        self.bias = nn.Parameter(torch.zeros(1, 3, 1, 1))
        
        self.dilation = dilation ##### modified
        if dilation > 1: ##### modified
            blur_weight = torch.randn(1, 1, 3, 3) * 0  + 1
            blur_weight[:,:,0,1] = 2
            blur_weight[:,:,1,0] = 2
            blur_weight[:,:,1,2] = 2
            blur_weight[:,:,2,1] = 2
            blur_weight[:,:,1,1] = 4
            blur_weight = blur_weight / 16.0 
            self.register_buffer("blur_weight", blur_weight)

    def forward(self, input, style, skip=None):
        out = self.conv(input, style)
        out = out + self.bias

        if skip is not None:
            if self.dilation == 1:
                skip = self.upsample(skip)
            else:  ##### modified, to simulate skip = self.upsample(skip)
                batch, in_channel, _, _ = skip.shape
                skip = F.conv2d(skip, self.blur_weight.repeat(in_channel,1,1,1), 
                                padding=self.dilation//2, groups=in_channel, dilation=self.dilation//2)

            out = out + skip

        return out


class Generator(nn.Module):
    def __init__(
            self,
            size,
            style_dim,
            n_mlp,
            channel_multiplier=2,
            blur_kernel=[1, 3, 3, 1],
            lr_mlp=0.01,
    ):
        super().__init__()

        self.size = size

        self.style_dim = style_dim

        layers = [PixelNorm()]

        for i in range(n_mlp):
            layers.append(
                EqualLinear(
                    style_dim, style_dim, lr_mul=lr_mlp, activation='fused_lrelu'
                )
            )

        self.style = nn.Sequential(*layers)

        self.channels = {
            4: 512,
            8: 512,
            16: 512,
            32: 512,
            64: 256 * channel_multiplier,
            128: 128 * channel_multiplier,
            256: 64 * channel_multiplier,
            512: 32 * channel_multiplier,
            1024: 16 * channel_multiplier,
        }

        self.input = ConstantInput(self.channels[4])
        self.conv1 = StyledConv(
            self.channels[4], self.channels[4], 3, style_dim, blur_kernel=blur_kernel, dilation=8 ##### modified
        )
        self.to_rgb1 = ToRGB(self.channels[4], style_dim, upsample=False)

        self.log_size = int(math.log(size, 2))
        self.num_layers = (self.log_size - 2) * 2 + 1

        self.convs = nn.ModuleList()
        self.upsamples = nn.ModuleList()
        self.to_rgbs = nn.ModuleList()
        self.noises = nn.Module()

        in_channel = self.channels[4]

        for layer_idx in range(self.num_layers):
            res = (layer_idx + 5) // 2
            shape = [1, 1, 2 ** res, 2 ** res]
            self.noises.register_buffer(f'noise_{layer_idx}', torch.randn(*shape))

        for i in range(3, self.log_size + 1):
            out_channel = self.channels[2 ** i]

            self.convs.append(
                StyledConv(
                    in_channel,
                    out_channel,
                    3,
                    style_dim,
                    upsample=True,
                    blur_kernel=blur_kernel,
                    dilation=max(1, 32 // (2**(i-1))) ##### modified
                )
            )

            self.convs.append(
                StyledConv(
                    out_channel, out_channel, 3, style_dim, blur_kernel=blur_kernel, dilation=max(1, 32 // (2**i))  ##### modified
                )
            )

            self.to_rgbs.append(ToRGB(out_channel, style_dim, dilation=max(1, 32 // (2**(i-1))))) ##### modified

            in_channel = out_channel

        self.n_latent = self.log_size * 2 - 2

    def make_noise(self):
        device = self.input.input.device

        noises = [torch.randn(1, 1, 2 ** 2, 2 ** 2, device=device)]

        for i in range(3, self.log_size + 1):
            for _ in range(2):
                noises.append(torch.randn(1, 1, 2 ** i, 2 ** i, device=device))

        return noises

    def mean_latent(self, n_latent):
        latent_in = torch.randn(
            n_latent, self.style_dim, device=self.input.input.device
        )
        latent = self.style(latent_in).mean(0, keepdim=True)

        return latent

    def get_latent(self, input):
        return self.style(input)
    
    # styles is the latent code w+
    # first_layer_feature is the first-layer input feature f
    # first_layer_feature_ind indicate which layer of G accepts f (should always=0, the first layer)
    # skip_layer_feature is the encoder features sent by skip connection
    # fusion_block is the network to fuse the encoder feature and decoder feature
    # zero_noise is to force the noise to be zero (to avoid flickers for videos)
    # editing_w is the editing vector v used in video face editing
    def forward(
            self,
            styles,
            return_latents=False,
            return_features=False,
            inject_index=None,
            truncation=1,
            truncation_latent=None,
            input_is_latent=False,
            noise=None,
            randomize_noise=True,
            first_layer_feature = None, ##### modified
            first_layer_feature_ind = 0,  ##### modified
            skip_layer_feature = None,   ##### modified
            fusion_block = None,   ##### modified
            zero_noise = False,   ##### modified
            editing_w = None,   ##### modified
    ):
        if not input_is_latent:
            styles = [self.style(s) for s in styles]

        if zero_noise:
            noise = [
                getattr(self.noises, f'noise_{i}') * 0.0 for i in range(self.num_layers)
            ]
        elif noise is None:
            if randomize_noise:
                noise = [None] * self.num_layers
            else:
                noise = [
                    getattr(self.noises, f'noise_{i}') for i in range(self.num_layers)
                ]

        if truncation < 1:
            style_t = []

            for style in styles:
                style_t.append(
                    truncation_latent + truncation * (style - truncation_latent)
                )

            styles = style_t

        if len(styles) < 2:
            inject_index = self.n_latent

            if styles[0].ndim < 3:
                latent = styles[0].unsqueeze(1).repeat(1, inject_index, 1)
            else:
                latent = styles[0]

        else:
            if inject_index is None:
                inject_index = random.randint(1, self.n_latent - 1)

            latent = styles[0].unsqueeze(1).repeat(1, inject_index, 1)
            latent2 = styles[1].unsqueeze(1).repeat(1, self.n_latent - inject_index, 1)

            latent = torch.cat([latent, latent2], 1)
        
        # w+ + v for video face editing
        if editing_w is not None:  ##### modified
            latent = latent + editing_w
        
        # the original StyleGAN
        if first_layer_feature is None: ##### modified
            out = self.input(latent)
            out = F.adaptive_avg_pool2d(out, 32) ##### modified
            out = self.conv1(out, latent[:, 0], noise=noise[0])
            skip = self.to_rgb1(out, latent[:, 1])       
        # the default StyleGANEX, replacing the first layer of G
        elif first_layer_feature_ind == 0: ##### modified
            out = first_layer_feature[0] ##### modified
            out = self.conv1(out, latent[:, 0], noise=noise[0])
            skip = self.to_rgb1(out, latent[:, 1])    
        # maybe we can also use the second layer of G to accept f?
        else: ##### modified
            out = first_layer_feature[0] ##### modified
            skip = first_layer_feature[1]  ##### modified    

        i = 1
        for conv1, conv2, noise1, noise2, to_rgb in zip(
                self.convs[::2], self.convs[1::2], noise[1::2], noise[2::2], self.to_rgbs
        ):
            # these layers accepts skipped encoder layer, use fusion block to fuse the encoder feature and decoder feature
            if skip_layer_feature and fusion_block and i//2 < len(skip_layer_feature) and i//2 < len(fusion_block):
                if editing_w is None:
                    out, skip = fusion_block[i//2](skip_layer_feature[i//2], out, skip) 
                else:
                    out, skip = fusion_block[i//2](skip_layer_feature[i//2], out, skip, editing_w[:,i]) 
            out = conv1(out, latent[:, i], noise=noise1)
            out = conv2(out, latent[:, i + 1], noise=noise2)
            skip = to_rgb(out, latent[:, i + 2], skip)

            i += 2

        image = skip

        if return_latents:
            return image, latent
        elif return_features:
            return image, out
        else:
            return image, None


class ConvLayer(nn.Sequential):
    def __init__(
            self,
            in_channel,
            out_channel,
            kernel_size,
            downsample=False,
            blur_kernel=[1, 3, 3, 1],
            bias=True,
            activate=True,
            dilation=1, ## modified
    ):
        layers = []

        if downsample:
            factor = 2
            p = (len(blur_kernel) - factor) + (kernel_size - 1)
            pad0 = (p + 1) // 2
            pad1 = p // 2

            layers.append(Blur(blur_kernel, pad=(pad0, pad1)))

            stride = 2
            self.padding = 0

        else:
            stride = 1
            self.padding = kernel_size // 2 + dilation-1 ## modified

        layers.append(
            EqualConv2d(
                in_channel,
                out_channel,
                kernel_size,
                padding=self.padding,
                stride=stride,
                bias=bias and not activate,
                dilation=dilation, ## modified
            )
        )

        if activate:
            if bias:
                layers.append(FusedLeakyReLU(out_channel))

            else:
                layers.append(ScaledLeakyReLU(0.2))

        super().__init__(*layers)


class ResBlock(nn.Module):
    def __init__(self, in_channel, out_channel, blur_kernel=[1, 3, 3, 1]):
        super().__init__()

        self.conv1 = ConvLayer(in_channel, in_channel, 3)
        self.conv2 = ConvLayer(in_channel, out_channel, 3, downsample=True)

        self.skip = ConvLayer(
            in_channel, out_channel, 1, downsample=True, activate=False, bias=False
        )

    def forward(self, input):
        out = self.conv1(input)
        out = self.conv2(out)

        skip = self.skip(input)
        out = (out + skip) / math.sqrt(2)

        return out


class Discriminator(nn.Module):
    def __init__(self, size, channel_multiplier=2, blur_kernel=[1, 3, 3, 1], img_channel=3):
        super().__init__()

        channels = {
            4: 512,
            8: 512,
            16: 512,
            32: 512,
            64: 256 * channel_multiplier,
            128: 128 * channel_multiplier,
            256: 64 * channel_multiplier,
            512: 32 * channel_multiplier,
            1024: 16 * channel_multiplier,
        }

        convs = [ConvLayer(img_channel, channels[size], 1)]

        log_size = int(math.log(size, 2))

        in_channel = channels[size]

        for i in range(log_size, 2, -1):
            out_channel = channels[2 ** (i - 1)]

            convs.append(ResBlock(in_channel, out_channel, blur_kernel))

            in_channel = out_channel

        self.convs = nn.Sequential(*convs)

        self.stddev_group = 4
        self.stddev_feat = 1

        self.final_conv = ConvLayer(in_channel + 1, channels[4], 3)
        self.final_linear = nn.Sequential(
            EqualLinear(channels[4] * 4 * 4, channels[4], activation='fused_lrelu'),
            EqualLinear(channels[4], 1),
        )
        
        self.size = size ##### modified

    def forward(self, input):
        # for input that not satisfies the target size, we crop it to extract a small image of the target size.
        _, _, h, w = input.shape ##### modified
        i, j = torch.randint(0, h+1-self.size, size=(1,)).item(), torch.randint(0, w+1-self.size, size=(1,)).item() ##### modified
        out = self.convs(input[:,:,i:i+self.size,j:j+self.size]) ##### modified

        batch, channel, height, width = out.shape
        group = min(batch, self.stddev_group)
        stddev = out.view(
            group, -1, self.stddev_feat, channel // self.stddev_feat, height, width
        )
        stddev = torch.sqrt(stddev.var(0, unbiased=False) + 1e-8)
        stddev = stddev.mean([2, 3, 4], keepdims=True).squeeze(2)
        stddev = stddev.repeat(group, 1, height, width)
        out = torch.cat([out, stddev], 1)

        out = self.final_conv(out)

        out = out.view(batch, -1)
        out = self.final_linear(out)

        return out