2020-11-05 17:04:17 +00:00
|
|
|
import os
|
|
|
|
|
|
|
|
import torch
|
|
|
|
import torch.nn as nn
|
|
|
|
import torch.nn.functional as F
|
|
|
|
import torchvision
|
|
|
|
from torch.utils.checkpoint import checkpoint_sequential
|
|
|
|
|
|
|
|
from models.archs.arch_util import make_layer, default_init_weights, ConvGnSilu, ConvGnLelu
|
2020-11-05 20:31:34 +00:00
|
|
|
from models.archs.srg2_classic import Interpolate
|
2020-11-05 17:04:17 +00:00
|
|
|
from utils.util import checkpoint
|
|
|
|
|
|
|
|
|
|
|
|
class ResidualDenseBlock(nn.Module):
|
|
|
|
"""Residual Dense Block.
|
|
|
|
|
|
|
|
Used in RRDB block in ESRGAN.
|
|
|
|
|
|
|
|
Args:
|
|
|
|
mid_channels (int): Channel number of intermediate features.
|
|
|
|
growth_channels (int): Channels for each growth.
|
|
|
|
"""
|
|
|
|
|
|
|
|
def __init__(self, mid_channels=64, growth_channels=32):
|
|
|
|
super(ResidualDenseBlock, self).__init__()
|
|
|
|
for i in range(5):
|
|
|
|
out_channels = mid_channels if i == 4 else growth_channels
|
|
|
|
self.add_module(
|
|
|
|
f'conv{i+1}',
|
|
|
|
nn.Conv2d(mid_channels + i * growth_channels, out_channels, 3,
|
|
|
|
1, 1))
|
|
|
|
self.lrelu = nn.LeakyReLU(negative_slope=0.2, inplace=True)
|
|
|
|
for i in range(5):
|
|
|
|
default_init_weights(getattr(self, f'conv{i+1}'), 0.1)
|
|
|
|
|
|
|
|
|
|
|
|
def forward(self, x, identity=None):
|
|
|
|
if identity is None:
|
|
|
|
identity = x
|
|
|
|
x1 = self.lrelu(self.conv1(x))
|
|
|
|
x2 = self.lrelu(self.conv2(torch.cat((x, x1), 1)))
|
|
|
|
x3 = self.lrelu(self.conv3(torch.cat((x, x1, x2), 1)))
|
|
|
|
x4 = self.lrelu(self.conv4(torch.cat((x, x1, x2, x3), 1)))
|
|
|
|
x5 = self.conv5(torch.cat((x, x1, x2, x3, x4), 1))
|
|
|
|
return x5 * 0.2 + identity
|
|
|
|
|
|
|
|
|
|
|
|
class RRDBWithBypassAndLatent(nn.Module):
|
2020-11-06 05:13:05 +00:00
|
|
|
def __init__(self, mid_channels, growth_channels=32):
|
2020-11-05 17:04:17 +00:00
|
|
|
super(RRDBWithBypassAndLatent, self).__init__()
|
|
|
|
self.latent_join = nn.Sequential(ConvGnLelu(mid_channels*2, mid_channels*2, activation=True, norm=False, bias=False),
|
|
|
|
ConvGnLelu(mid_channels*2, mid_channels, activation=False, norm=False, bias=False))
|
|
|
|
self.rdb1 = ResidualDenseBlock(mid_channels, growth_channels)
|
|
|
|
self.rdb2 = ResidualDenseBlock(mid_channels, growth_channels)
|
|
|
|
self.rdb3 = ResidualDenseBlock(mid_channels, growth_channels)
|
|
|
|
self.bypass = nn.Sequential(ConvGnSilu(mid_channels*2, mid_channels, kernel_size=3, bias=True, activation=True, norm=True),
|
|
|
|
ConvGnSilu(mid_channels, mid_channels//2, kernel_size=3, bias=False, activation=True, norm=False),
|
|
|
|
ConvGnSilu(mid_channels//2, 1, kernel_size=3, bias=False, activation=False, norm=False),
|
|
|
|
nn.Sigmoid())
|
|
|
|
|
2020-11-06 05:13:05 +00:00
|
|
|
def forward(self, x, latent):
|
2020-11-05 17:04:17 +00:00
|
|
|
out = self.latent_join(torch.cat([x, latent], dim=1))
|
|
|
|
out = self.rdb1(out, x)
|
|
|
|
out = self.rdb2(out)
|
|
|
|
out = self.rdb3(out)
|
|
|
|
bypass = self.bypass(torch.cat([x, out], dim=1))
|
|
|
|
self.bypass_map = bypass.detach().clone()
|
2020-11-05 20:31:34 +00:00
|
|
|
residual = out * .2 * bypass
|
|
|
|
return residual + x, residual
|
2020-11-05 17:04:17 +00:00
|
|
|
|
|
|
|
|
2020-11-06 05:13:05 +00:00
|
|
|
class ConvLatentEncoder(nn.Module):
|
|
|
|
def __init__(self, nf):
|
|
|
|
super(ConvLatentEncoder, self).__init__()
|
|
|
|
latent_filters = [nf * 4, nf * 2, nf]
|
|
|
|
layers = []
|
|
|
|
for i in range(len(latent_filters)-1):
|
|
|
|
layers.append(nn.Sequential(
|
|
|
|
ConvGnLelu(latent_filters[i], latent_filters[i], kernel_size=1, activation=True, bias=False, norm=True),
|
|
|
|
Interpolate(2),
|
|
|
|
ConvGnLelu(latent_filters[i], latent_filters[i+1], kernel_size=1, activation=True, bias=False, norm=True)))
|
|
|
|
self.final = nn.Sequential(
|
|
|
|
ConvGnLelu(nf, nf, kernel_size=1, activation=True, bias=True, norm=True),
|
|
|
|
ConvGnLelu(nf, nf, kernel_size=1, activation=False, bias=True, norm=False))
|
|
|
|
self.layers = nn.ModuleList(layers)
|
|
|
|
|
|
|
|
def forward(self, latents):
|
|
|
|
assert len(latents) == 3
|
|
|
|
out = torch.zeros_like(latents[0])
|
|
|
|
for i in range(2):
|
|
|
|
out = out + latents[i]
|
|
|
|
out = self.layers[i](out)
|
|
|
|
out = out + latents[2]
|
|
|
|
return self.final(out)
|
|
|
|
|
|
|
|
|
2020-11-05 17:04:17 +00:00
|
|
|
class RRDBNetWithLatent(nn.Module):
|
2020-11-05 20:31:34 +00:00
|
|
|
# 8-layer MLP in the vein of StyleGAN.
|
|
|
|
def create_linear_latent_encoder(self, latent_size):
|
|
|
|
return nn.Sequential(nn.Linear(latent_size, latent_size),
|
|
|
|
nn.BatchNorm1d(latent_size),
|
|
|
|
nn.LeakyReLU(negative_slope=0.2, inplace=True),
|
|
|
|
nn.Linear(latent_size, latent_size),
|
|
|
|
nn.BatchNorm1d(latent_size),
|
|
|
|
nn.LeakyReLU(negative_slope=0.2, inplace=True),
|
|
|
|
nn.Linear(latent_size, latent_size),
|
|
|
|
nn.BatchNorm1d(latent_size),
|
|
|
|
nn.LeakyReLU(negative_slope=0.2, inplace=True),
|
|
|
|
nn.Linear(latent_size, latent_size),
|
|
|
|
nn.BatchNorm1d(latent_size),
|
|
|
|
nn.LeakyReLU(negative_slope=0.2, inplace=True),
|
|
|
|
nn.Linear(latent_size, latent_size),
|
|
|
|
nn.BatchNorm1d(latent_size),
|
|
|
|
nn.LeakyReLU(negative_slope=0.2, inplace=True),
|
|
|
|
nn.Linear(latent_size, latent_size),
|
|
|
|
nn.BatchNorm1d(latent_size),
|
|
|
|
nn.LeakyReLU(negative_slope=0.2, inplace=True),
|
|
|
|
nn.Linear(latent_size, latent_size),
|
|
|
|
nn.BatchNorm1d(latent_size),
|
|
|
|
nn.LeakyReLU(negative_slope=0.2, inplace=True),
|
|
|
|
nn.Linear(latent_size, latent_size),
|
|
|
|
nn.BatchNorm1d(latent_size),
|
|
|
|
nn.LeakyReLU(negative_slope=0.2, inplace=True))
|
|
|
|
|
|
|
|
# Creates a 2D latent by iterating through the provided latent_filters and doubling the
|
|
|
|
# image size each step.
|
|
|
|
def create_conv_latent_encoder(self, latent_filters):
|
2020-11-06 05:13:05 +00:00
|
|
|
return ConvLatentEncoder(latent_filters)
|
2020-11-05 20:31:34 +00:00
|
|
|
|
2020-11-05 17:04:17 +00:00
|
|
|
def __init__(self,
|
|
|
|
in_channels,
|
|
|
|
out_channels,
|
|
|
|
mid_channels=64,
|
|
|
|
num_blocks=23,
|
|
|
|
growth_channels=32,
|
|
|
|
blocks_per_checkpoint=4,
|
2020-11-06 05:13:05 +00:00
|
|
|
scale=4):
|
2020-11-05 17:04:17 +00:00
|
|
|
super(RRDBNetWithLatent, self).__init__()
|
|
|
|
self.num_blocks = num_blocks
|
|
|
|
self.blocks_per_checkpoint = blocks_per_checkpoint
|
|
|
|
self.scale = scale
|
|
|
|
self.in_channels = in_channels
|
2020-11-06 05:13:05 +00:00
|
|
|
self.nf = mid_channels
|
2020-11-05 17:04:17 +00:00
|
|
|
first_conv_stride = 1 if in_channels <= 4 else scale
|
|
|
|
first_conv_ksize = 3 if first_conv_stride == 1 else 7
|
|
|
|
first_conv_padding = 1 if first_conv_stride == 1 else 3
|
|
|
|
self.conv_first = nn.Conv2d(in_channels, mid_channels, first_conv_ksize, first_conv_stride, first_conv_padding)
|
|
|
|
self.body = make_layer(
|
|
|
|
RRDBWithBypassAndLatent,
|
|
|
|
num_blocks,
|
|
|
|
mid_channels=mid_channels,
|
2020-11-06 05:13:05 +00:00
|
|
|
growth_channels=growth_channels)
|
2020-11-05 17:04:17 +00:00
|
|
|
self.conv_body = nn.Conv2d(mid_channels, mid_channels, 3, 1, 1)
|
|
|
|
# upsample
|
|
|
|
self.conv_up1 = nn.Conv2d(mid_channels, mid_channels, 3, 1, 1)
|
|
|
|
self.conv_up2 = nn.Conv2d(mid_channels, mid_channels, 3, 1, 1)
|
|
|
|
self.conv_hr = nn.Conv2d(mid_channels, mid_channels, 3, 1, 1)
|
|
|
|
self.conv_last = nn.Conv2d(mid_channels, out_channels, 3, 1, 1)
|
|
|
|
self.lrelu = nn.LeakyReLU(negative_slope=0.2, inplace=True)
|
|
|
|
|
2020-11-06 05:13:05 +00:00
|
|
|
self.latent_encoder = self.create_conv_latent_encoder(mid_channels)
|
2020-11-05 17:04:17 +00:00
|
|
|
|
|
|
|
for m in [
|
|
|
|
self.conv_first, self.conv_body, self.conv_up1,
|
|
|
|
self.conv_up2, self.conv_hr, self.conv_last
|
|
|
|
]:
|
|
|
|
default_init_weights(m, 0.1)
|
|
|
|
|
|
|
|
def forward(self, x, latent=None, ref=None):
|
2020-11-05 20:31:34 +00:00
|
|
|
latent_was_none = latent
|
2020-11-05 17:04:17 +00:00
|
|
|
if latent is None:
|
2020-11-06 05:13:05 +00:00
|
|
|
mults = [4, 2, 1]
|
|
|
|
b, f, h, w = x.shape
|
|
|
|
latent = [torch.randn((b, self.nf * m, h // m, w // m), dtype=torch.float, device=x.device) for m in mults]
|
|
|
|
latent = self.latent_encoder(latent)
|
2020-11-05 20:31:34 +00:00
|
|
|
if latent_was_none is None:
|
|
|
|
self.latent_mean = torch.mean(latent).detach().cpu()
|
|
|
|
self.latent_std = torch.std(latent).detach().cpu()
|
|
|
|
self.latent_var = torch.var(latent).detach().cpu()
|
2020-11-05 17:04:17 +00:00
|
|
|
if self.in_channels > 4:
|
|
|
|
x_lg = F.interpolate(x, scale_factor=self.scale, mode="bicubic")
|
|
|
|
if ref is None:
|
|
|
|
ref = torch.zeros_like(x_lg)
|
|
|
|
x_lg = torch.cat([x_lg, ref], dim=1)
|
|
|
|
else:
|
|
|
|
x_lg = x
|
|
|
|
feat = self.conv_first(x_lg)
|
|
|
|
body_feat = feat
|
2020-11-05 20:31:34 +00:00
|
|
|
self.block_residual_means = []
|
|
|
|
self.block_residual_stds = []
|
2020-11-05 17:04:17 +00:00
|
|
|
for bl in self.body:
|
2020-11-05 20:31:34 +00:00
|
|
|
body_feat, residual = checkpoint(bl, body_feat, latent)
|
|
|
|
self.block_residual_means.append(torch.mean(residual).cpu())
|
|
|
|
self.block_residual_stds.append(torch.std(residual).cpu())
|
2020-11-05 17:04:17 +00:00
|
|
|
body_feat = self.conv_body(body_feat)
|
|
|
|
feat = feat + body_feat
|
|
|
|
# upsample
|
|
|
|
feat = self.lrelu(
|
|
|
|
self.conv_up1(F.interpolate(feat, scale_factor=2, mode='nearest')))
|
|
|
|
if self.scale == 4:
|
|
|
|
feat = self.lrelu(
|
|
|
|
self.conv_up2(F.interpolate(feat, scale_factor=2, mode='nearest')))
|
|
|
|
else:
|
|
|
|
feat = self.lrelu(self.conv_up2(feat))
|
|
|
|
out = self.conv_last(self.lrelu(self.conv_hr(feat)))
|
|
|
|
return out
|
|
|
|
|
|
|
|
def visual_dbg(self, step, path):
|
|
|
|
for i, bm in enumerate(self.body):
|
|
|
|
torchvision.utils.save_image(bm.bypass_map.cpu().float(), os.path.join(path, "%i_bypass_%i.png" % (step, i+1)))
|
|
|
|
|
2020-11-05 20:31:34 +00:00
|
|
|
def get_debug_values(self, s, n):
|
|
|
|
blk_stds, blk_means = {}, {}
|
|
|
|
for i, (s, m) in enumerate(zip(self.block_residual_stds, self.block_residual_means)):
|
|
|
|
blk_stds['block_%i' % (i+1,)] = s
|
|
|
|
blk_means['block_%i' % (i+1,)] = m
|
|
|
|
return {'encoded_latent_mean': self.latent_mean,
|
|
|
|
'encoded_latent_std': self.latent_std,
|
|
|
|
'encoded_latent_var': self.latent_var,
|
|
|
|
'blocks_mean': blk_means,
|
|
|
|
'blocks_std': blk_stds}
|
|
|
|
|
2020-11-05 17:04:17 +00:00
|
|
|
|
|
|
|
# Based heavily on the same VGG arch used for the discriminator.
|
|
|
|
class LatentEstimator(nn.Module):
|
|
|
|
# input_img_factor = multiplier to support images over 128x128. Only certain factors are supported.
|
2020-11-07 03:38:23 +00:00
|
|
|
def __init__(self, in_nc, nf, overwrite_levels=[]):
|
2020-11-05 17:04:17 +00:00
|
|
|
super(LatentEstimator, self).__init__()
|
2020-11-07 03:38:23 +00:00
|
|
|
self.overwrite_levels = overwrite_levels
|
2020-11-05 17:04:17 +00:00
|
|
|
# [64, 128, 128]
|
|
|
|
self.conv0_0 = nn.Conv2d(in_nc, nf, 3, 1, 1, bias=True)
|
|
|
|
self.conv0_1 = nn.Conv2d(nf, nf, 4, 2, 1, bias=False)
|
|
|
|
self.bn0_1 = nn.BatchNorm2d(nf, affine=True)
|
|
|
|
# [64, 64, 64]
|
|
|
|
self.conv1_0 = nn.Conv2d(nf, nf * 2, 3, 1, 1, bias=False)
|
|
|
|
self.bn1_0 = nn.BatchNorm2d(nf * 2, affine=True)
|
|
|
|
self.conv1_1 = nn.Conv2d(nf * 2, nf * 2, 4, 2, 1, bias=False)
|
|
|
|
self.bn1_1 = nn.BatchNorm2d(nf * 2, affine=True)
|
2020-11-06 05:13:05 +00:00
|
|
|
self.d1p1 = ConvGnLelu(nf * 2, nf, kernel_size=1, activation=True, norm=True, bias=True)
|
|
|
|
self.d1p2 = ConvGnLelu(nf, nf, kernel_size=1, activation=False, norm=False, bias=True)
|
|
|
|
|
2020-11-05 17:04:17 +00:00
|
|
|
# [128, 32, 32]
|
|
|
|
self.conv2_0 = nn.Conv2d(nf * 2, nf * 4, 3, 1, 1, bias=False)
|
|
|
|
self.bn2_0 = nn.BatchNorm2d(nf * 4, affine=True)
|
|
|
|
self.conv2_1 = nn.Conv2d(nf * 4, nf * 4, 4, 2, 1, bias=False)
|
|
|
|
self.bn2_1 = nn.BatchNorm2d(nf * 4, affine=True)
|
2020-11-06 05:13:05 +00:00
|
|
|
self.d2p1 = ConvGnLelu(nf * 4, nf * 2, kernel_size=1, activation=True, norm=True, bias=True)
|
|
|
|
self.d2p2 = ConvGnLelu(nf * 2, nf * 2, kernel_size=1, activation=False, norm=False, bias=True)
|
|
|
|
|
2020-11-05 17:04:17 +00:00
|
|
|
# [256, 16, 16]
|
|
|
|
self.conv3_0 = nn.Conv2d(nf * 4, nf * 8, 3, 1, 1, bias=False)
|
|
|
|
self.bn3_0 = nn.BatchNorm2d(nf * 8, affine=True)
|
|
|
|
self.conv3_1 = nn.Conv2d(nf * 8, nf * 8, 4, 2, 1, bias=False)
|
|
|
|
self.bn3_1 = nn.BatchNorm2d(nf * 8, affine=True)
|
2020-11-06 05:13:05 +00:00
|
|
|
self.d3p1 = ConvGnLelu(nf * 8, nf * 4, kernel_size=1, activation=True, norm=True, bias=True)
|
|
|
|
self.d3p2 = ConvGnLelu(nf * 4, nf * 4, kernel_size=1, activation=False, norm=False, bias=True)
|
2020-11-05 17:04:17 +00:00
|
|
|
|
2020-11-06 05:13:05 +00:00
|
|
|
self.lrelu = nn.LeakyReLU(.2, inplace=True)
|
2020-11-05 20:31:34 +00:00
|
|
|
self.tanh = nn.Tanh()
|
2020-11-05 17:04:17 +00:00
|
|
|
|
|
|
|
def compute_body(self, x):
|
2020-11-06 05:13:05 +00:00
|
|
|
fea = self.lrelu(self.bn1_0(self.conv1_0(x)))
|
2020-11-05 17:04:17 +00:00
|
|
|
fea = self.lrelu(self.bn1_1(self.conv1_1(fea)))
|
2020-11-06 05:13:05 +00:00
|
|
|
o1 = self.tanh(self.d1p2(self.d1p1(fea)))
|
2020-11-05 17:04:17 +00:00
|
|
|
|
|
|
|
fea = self.lrelu(self.bn2_0(self.conv2_0(fea)))
|
|
|
|
fea = self.lrelu(self.bn2_1(self.conv2_1(fea)))
|
2020-11-06 05:13:05 +00:00
|
|
|
o2 = self.tanh(self.d2p2(self.d2p1(fea)))
|
2020-11-05 17:04:17 +00:00
|
|
|
|
|
|
|
fea = self.lrelu(self.bn3_0(self.conv3_0(fea)))
|
|
|
|
fea = self.lrelu(self.bn3_1(self.conv3_1(fea)))
|
2020-11-06 05:13:05 +00:00
|
|
|
o3 = self.tanh(self.d3p2(self.d3p1(fea)))
|
2020-11-05 17:04:17 +00:00
|
|
|
|
2020-11-06 05:13:05 +00:00
|
|
|
return o3, o2, o1
|
2020-11-05 17:04:17 +00:00
|
|
|
|
|
|
|
def forward(self, x):
|
2020-11-06 05:13:05 +00:00
|
|
|
fea = self.lrelu(self.conv0_0(x))
|
|
|
|
fea = self.lrelu(self.bn0_1(self.conv0_1(fea)))
|
|
|
|
out = list(checkpoint(self.compute_body, fea))
|
2020-11-07 03:38:23 +00:00
|
|
|
for lvl in self.overwrite_levels:
|
|
|
|
out[lvl] = torch.zeros_like(out[lvl])
|
2020-11-06 05:13:05 +00:00
|
|
|
self.latent_mean = torch.mean(out[-1])
|
|
|
|
self.latent_std = torch.std(out[-1])
|
|
|
|
self.latent_var = torch.var(out[-1])
|
2020-11-05 20:31:34 +00:00
|
|
|
return out
|
|
|
|
|
|
|
|
def get_debug_values(self, s, n):
|
|
|
|
return {'latent_estimator_mean': self.latent_mean,
|
|
|
|
'latent_estimator_std': self.latent_std,
|
|
|
|
'latent_estimator_var': self.latent_var}
|
|
|
|
|