DL-Art-School/codes/models/SRGAN_model.py

1046 lines
57 KiB
Python
Raw Normal View History

2019-08-23 13:42:47 +00:00
import logging
from collections import OrderedDict
import torch
import torch.nn as nn
from torch.nn.parallel import DataParallel, DistributedDataParallel
import models.networks as networks
import models.lr_scheduler as lr_scheduler
from models.base_model import BaseModel
from models.loss import GANLoss, FDPLLoss
from apex import amp
from data.weight_scheduler import get_scheduler_for_opt
from .archs.SPSR_arch import ImageGradient, ImageGradientNoPadding
import torch.nn.functional as F
import glob
import random
import torchvision.utils as utils
import os
2019-08-23 13:42:47 +00:00
logger = logging.getLogger('base')
2020-08-23 23:32:03 +00:00
class GaussianBlur(nn.Module):
def __init__(self):
super(GaussianBlur, self).__init__()
# Set these to whatever you want for your gaussian filter
kernel_size = 3
sigma = 2
# Create a x, y coordinate grid of shape (kernel_size, kernel_size, 2)
x_cord = torch.arange(kernel_size)
x_grid = x_cord.repeat(kernel_size).view(kernel_size, kernel_size)
y_grid = x_grid.t()
xy_grid = torch.stack([x_grid, y_grid], dim=-1)
mean = (kernel_size - 1) / 2.
variance = sigma ** 2.
# Calculate the 2-dimensional gaussian kernel which is
# the product of two gaussian distributions for two different
# variables (in this case called x and y)
gaussian_kernel = (1. / (2. * 3.1415926 * variance)) * \
torch.exp(
-torch.sum((xy_grid - mean) ** 2., dim=-1) / \
(2 * variance)
)
# Make sure sum of values in gaussian kernel equals 1.
gaussian_kernel = gaussian_kernel / torch.sum(gaussian_kernel)
# Reshape to 2d depthwise convolutional weight
gaussian_kernel = gaussian_kernel.view(1, 1, kernel_size, kernel_size)
gaussian_kernel = gaussian_kernel.repeat(3, 1, 1, 1)
self.gaussian_filter = nn.Conv2d(in_channels=3, out_channels=3,
kernel_size=kernel_size, groups=3, bias=False)
self.gaussian_filter.weight.data = gaussian_kernel
self.gaussian_filter.weight.requires_grad = False
def forward(self, x):
return self.gaussian_filter(x)
2019-08-23 13:42:47 +00:00
class SRGANModel(BaseModel):
def __init__(self, opt):
super(SRGANModel, self).__init__(opt)
if opt['dist']:
self.rank = torch.distributed.get_rank()
else:
self.rank = -1 # non dist training
train_opt = opt['train']
self.spsr_enabled = 'spsr' in opt['model']
2019-08-23 13:42:47 +00:00
# define networks and load pretrained models
self.netG = networks.define_G(opt).to(self.device)
if self.is_train:
self.netD = networks.define_D(opt).to(self.device)
if self.spsr_enabled:
2020-08-26 00:11:53 +00:00
logger.info("Defining grad net...")
2020-08-25 23:58:20 +00:00
self.netD_grad = networks.define_D(opt, wrap=True).to(self.device) # D_grad
2019-08-23 13:42:47 +00:00
if 'network_C' in opt.keys():
self.netC = networks.define_G(opt, net_key='network_C').to(self.device)
# The corruptor net is fixed. Lock 'her down.
self.netC.eval()
for p in self.netC.parameters():
p.requires_grad = True
else:
self.netC = None
2020-07-24 18:26:44 +00:00
self.mega_batch_factor = 1
self.disjoint_data = False
2019-08-23 13:42:47 +00:00
# define losses, optimizer and scheduler
if self.is_train:
2020-05-02 23:46:59 +00:00
self.mega_batch_factor = train_opt['mega_batch_factor']
if self.mega_batch_factor is None:
self.mega_batch_factor = 1
2019-08-23 13:42:47 +00:00
# G pixel loss
if train_opt['pixel_weight'] > 0:
l_pix_type = train_opt['pixel_criterion']
if l_pix_type == 'l1':
self.cri_pix = nn.L1Loss().to(self.device)
elif l_pix_type == 'l2':
self.cri_pix = nn.MSELoss().to(self.device)
else:
raise NotImplementedError('Loss type [{:s}] not recognized.'.format(l_pix_type))
self.l_pix_w = train_opt['pixel_weight']
else:
logger.info('Remove pixel loss.')
self.cri_pix = None
# FDPL loss.
if 'fdpl_loss' in train_opt.keys():
fdpl_opt = train_opt['fdpl_loss']
self.fdpl_weight = fdpl_opt['weight']
self.fdpl_enabled = self.fdpl_weight > 0
if self.fdpl_enabled:
self.cri_fdpl = FDPLLoss(fdpl_opt['data_mean'], self.device)
else:
self.fdpl_enabled = False
if self.spsr_enabled:
spsr_opt = train_opt['spsr']
self.branch_pretrain = spsr_opt['branch_pretrain'] if spsr_opt['branch_pretrain'] else 0
self.branch_init_iters = spsr_opt['branch_init_iters'] if spsr_opt['branch_init_iters'] else 1
if spsr_opt['gradient_pixel_weight'] > 0:
self.cri_pix_grad = nn.MSELoss().to(self.device)
self.l_pix_grad_w = spsr_opt['gradient_pixel_weight']
else:
self.cri_pix_grad = None
if spsr_opt['gradient_gan_weight'] > 0:
self.cri_grad_gan = GANLoss(train_opt['gan_type'], 1.0, 0.0).to(self.device)
self.l_gan_grad_w = spsr_opt['gradient_gan_weight']
else:
self.cri_grad_gan = None
if spsr_opt['pixel_branch_weight'] > 0:
l_pix_type = spsr_opt['pixel_branch_criterion']
if l_pix_type == 'l1':
self.cri_pix_branch = nn.L1Loss().to(self.device)
elif l_pix_type == 'l2':
self.cri_pix_branch = nn.MSELoss().to(self.device)
else:
raise NotImplementedError('Loss type [{:s}] not recognized.'.format(l_pix_type))
self.l_pix_branch_w = spsr_opt['pixel_branch_weight']
else:
logger.info('Remove G_grad pixel loss.')
self.cri_pix_branch = None
2019-08-23 13:42:47 +00:00
# G feature loss
if train_opt['feature_weight'] and train_opt['feature_weight'] > 0:
# For backwards compatibility, use a scheduler definition instead. Remove this at some point.
l_fea_type = train_opt['feature_criterion']
if l_fea_type == 'l1':
self.cri_fea = nn.L1Loss().to(self.device)
elif l_fea_type == 'l2':
self.cri_fea = nn.MSELoss().to(self.device)
else:
raise NotImplementedError('Loss type [{:s}] not recognized.'.format(l_fea_type))
sched_params = {
'type': 'fixed',
'weight': train_opt['feature_weight']
}
self.l_fea_sched = get_scheduler_for_opt(sched_params)
elif train_opt['feature_scheduler']:
self.l_fea_sched = get_scheduler_for_opt(train_opt['feature_scheduler'])
2019-08-23 13:42:47 +00:00
l_fea_type = train_opt['feature_criterion']
if l_fea_type == 'l1':
self.cri_fea = nn.L1Loss().to(self.device)
elif l_fea_type == 'l2':
self.cri_fea = nn.MSELoss().to(self.device)
else:
raise NotImplementedError('Loss type [{:s}] not recognized.'.format(l_fea_type))
else:
logger.info('Remove feature loss.')
self.cri_fea = None
if self.cri_fea: # load VGG perceptual loss
2020-08-23 23:32:03 +00:00
self.use_corrupted_feature_input = train_opt['corrupted_feature_input'] if 'corrupted_feature_input' in train_opt.keys() else False
if self.use_corrupted_feature_input:
2020-08-23 23:39:02 +00:00
logger.info("Corrupting inputs into the feature network..")
2020-08-23 23:52:18 +00:00
self.feature_corruptor = GaussianBlur().to(self.device)
2020-08-22 19:08:33 +00:00
self.netF = networks.define_F(use_bn=False).to(self.device)
self.lr_netF = None
if 'lr_fea_path' in train_opt.keys():
2020-08-22 19:08:33 +00:00
self.lr_netF = networks.define_F(use_bn=False, load_path=train_opt['lr_fea_path']).to(self.device)
self.disjoint_data = True
2019-08-23 13:42:47 +00:00
if opt['dist']:
2019-09-01 14:14:29 +00:00
pass # do not need to use DistributedDataParallel for netF
2019-08-23 13:42:47 +00:00
else:
self.netF = DataParallel(self.netF)
if self.lr_netF:
self.lr_netF = DataParallel(self.lr_netF)
2019-08-23 13:42:47 +00:00
# You can feed in a list of frozen pre-trained discriminators. These are treated the same as feature losses.
self.fixed_disc_nets = []
if 'fixed_discriminators' in opt.keys():
for opt_fdisc in opt['fixed_discriminators'].keys():
netFD = networks.define_fixed_D(opt['fixed_discriminators'][opt_fdisc]).to(self.device)
if opt['dist']:
pass # do not need to use DistributedDataParallel for netF
else:
netFD = DataParallel(netFD)
self.fixed_disc_nets.append(netFD)
2019-08-23 13:42:47 +00:00
# GD gan loss
self.cri_gan = GANLoss(train_opt['gan_type'], 1.0, 0.0).to(self.device)
self.l_gan_w = train_opt['gan_weight']
# D_update_ratio and D_init_iters
self.D_update_ratio = train_opt['D_update_ratio'] if train_opt['D_update_ratio'] else 1
self.D_init_iters = train_opt['D_init_iters'] if train_opt['D_init_iters'] else 0
2020-07-27 04:44:24 +00:00
self.G_warmup = train_opt['G_warmup'] if train_opt['G_warmup'] else -1
self.D_noise_theta = train_opt['D_noise_theta_init'] if train_opt['D_noise_theta_init'] else 0
self.D_noise_final = train_opt['D_noise_final_it'] if train_opt['D_noise_final_it'] else 0
self.D_noise_theta_floor = train_opt['D_noise_theta_floor'] if train_opt['D_noise_theta_floor'] else 0
self.corruptor_swapout_steps = train_opt['corruptor_swapout_steps'] if train_opt['corruptor_swapout_steps'] else 500
self.corruptor_usage_prob = train_opt['corruptor_usage_probability'] if train_opt['corruptor_usage_probability'] else .5
2019-08-23 13:42:47 +00:00
# optimizers
# G optimizer
2019-08-23 13:42:47 +00:00
wd_G = train_opt['weight_decay_G'] if train_opt['weight_decay_G'] else 0
optim_params = []
if train_opt['lr_scheme'] == 'ProgressiveMultiStepLR':
optim_params = self.netG.get_param_groups()
else:
for k, v in self.netG.named_parameters(): # can optimize for a part of the model
if v.requires_grad:
optim_params.append(v)
else:
if self.rank <= 0:
logger.warning('Params [{:s}] will not optimize.'.format(k))
2019-08-23 13:42:47 +00:00
self.optimizer_G = torch.optim.Adam(optim_params, lr=train_opt['lr_G'],
weight_decay=wd_G,
betas=(train_opt['beta1_G'], train_opt['beta2_G']))
self.optimizers.append(self.optimizer_G)
# D optimizer
optim_params = []
for k, v in self.netD.named_parameters(): # can optimize for a part of the model
if v.requires_grad:
optim_params.append(v)
else:
if self.rank <= 0:
logger.warning('Params [{:s}] will not optimize.'.format(k))
2019-08-23 13:42:47 +00:00
wd_D = train_opt['weight_decay_D'] if train_opt['weight_decay_D'] else 0
self.optimizer_D = torch.optim.Adam(optim_params, lr=train_opt['lr_D'],
2019-08-23 13:42:47 +00:00
weight_decay=wd_D,
betas=(train_opt['beta1_D'], train_opt['beta2_D']))
self.optimizers.append(self.optimizer_D)
self.disc_optimizers.append(self.optimizer_D)
2019-08-23 13:42:47 +00:00
if self.spsr_enabled:
# D_grad optimizer
optim_params = []
for k, v in self.netD_grad.named_parameters(): # can optimize for a part of the model
if v.requires_grad:
optim_params.append(v)
else:
if self.rank <= 0:
logger.warning('Params [{:s}] will not optimize.'.format(k))
# D
wd_D = train_opt['weight_decay_D'] if train_opt['weight_decay_D'] else 0
self.optimizer_D_grad = torch.optim.Adam(optim_params, lr=train_opt['lr_D'],
weight_decay=wd_D,
betas=(train_opt['beta1_D'], train_opt['beta2_D']))
self.optimizers.append(self.optimizer_D_grad)
self.disc_optimizers.append(self.optimizer_D_grad)
if self.spsr_enabled:
self.get_grad_nopadding = ImageGradientNoPadding().to(self.device)
2020-08-25 23:03:18 +00:00
[self.netG, self.netD, self.netD_grad, self.get_grad_nopadding], \
[self.optimizer_G, self.optimizer_D, self.optimizer_D_grad] = \
2020-08-25 23:05:16 +00:00
amp.initialize([self.netG, self.netD, self.netD_grad, self.get_grad_nopadding],
[self.optimizer_G, self.optimizer_D, self.optimizer_D_grad],
opt_level=self.amp_level, num_losses=3)
else:
# AMP
[self.netG, self.netD], [self.optimizer_G, self.optimizer_D] = \
amp.initialize([self.netG, self.netD], [self.optimizer_G, self.optimizer_D], opt_level=self.amp_level, num_losses=3)
2020-04-22 06:39:31 +00:00
# DataParallel
if opt['dist']:
self.netG = DistributedDataParallel(self.netG, device_ids=[torch.cuda.current_device()],
find_unused_parameters=True)
2020-04-22 06:39:31 +00:00
else:
self.netG = DataParallel(self.netG)
if self.is_train:
if opt['dist']:
self.netD = DistributedDataParallel(self.netD,
device_ids=[torch.cuda.current_device()],
find_unused_parameters=True)
2020-08-25 22:06:27 +00:00
if self.spsr_enabled:
self.netD_grad = DistributedDataParallel(self.netD_grad,
device_ids=[torch.cuda.current_device()],
find_unused_parameters=True)
2020-04-22 06:39:31 +00:00
else:
self.netD = DataParallel(self.netD)
2020-08-25 22:06:27 +00:00
if self.spsr_enabled:
self.netD_grad = DataParallel(self.netD_grad)
2020-04-22 06:39:31 +00:00
self.netG.train()
self.netD.train()
if self.spsr_enabled:
self.netD_grad.train()
2020-04-22 06:39:31 +00:00
2019-08-23 13:42:47 +00:00
# schedulers
if train_opt['lr_scheme'] == 'MultiStepLR':
# This is a recent change. assert to make sure any legacy configs dont find their way here.
assert 'gen_lr_steps' in train_opt.keys() and 'disc_lr_steps' in train_opt.keys()
self.schedulers.append(
lr_scheduler.MultiStepLR_Restart(self.optimizer_G, train_opt['gen_lr_steps'],
restarts=train_opt['restarts'],
weights=train_opt['restart_weights'],
gamma=train_opt['lr_gamma'],
clear_state=train_opt['clear_state'],
force_lr=train_opt['force_lr']))
for o in self.disc_optimizers:
2019-08-23 13:42:47 +00:00
self.schedulers.append(
lr_scheduler.MultiStepLR_Restart(o, train_opt['disc_lr_steps'],
2019-08-23 13:42:47 +00:00
restarts=train_opt['restarts'],
weights=train_opt['restart_weights'],
gamma=train_opt['lr_gamma'],
2020-06-07 22:56:05 +00:00
clear_state=train_opt['clear_state'],
force_lr=train_opt['force_lr']))
elif train_opt['lr_scheme'] == 'ProgressiveMultiStepLR':
# Only supported when there are two optimizers: G and D.
assert len(self.optimizers) == 2
self.schedulers.append(lr_scheduler.ProgressiveMultiStepLR(self.optimizer_G, train_opt['gen_lr_steps'],
self.netG.module.get_progressive_starts(),
train_opt['lr_gamma']))
for o in self.disc_optimizers:
self.schedulers.append(lr_scheduler.ProgressiveMultiStepLR(o, train_opt['disc_lr_steps'],
[0],
train_opt['lr_gamma']))
2019-08-23 13:42:47 +00:00
elif train_opt['lr_scheme'] == 'CosineAnnealingLR_Restart':
for optimizer in self.optimizers:
self.schedulers.append(
lr_scheduler.CosineAnnealingLR_Restart(
optimizer, train_opt['T_period'], eta_min=train_opt['eta_min'],
restarts=train_opt['restarts'], weights=train_opt['restart_weights']))
else:
raise NotImplementedError('MultiStepLR learning rate scheme is enough.')
self.log_dict = OrderedDict()
# Swapout params
self.swapout_G_freq = train_opt['swapout_G_freq'] if train_opt['swapout_G_freq'] else 0
self.swapout_G_duration = 0
self.swapout_D_freq = train_opt['swapout_D_freq'] if train_opt['swapout_D_freq'] else 0
self.swapout_D_duration = 0
self.swapout_duration = train_opt['swapout_duration'] if train_opt['swapout_duration'] else 0
# GAN LQ image params
self.gan_lq_img_use_prob = train_opt['gan_lowres_use_probability'] if train_opt['gan_lowres_use_probability'] else 0
self.img_debug_steps = opt['logger']['img_debug_steps'] if 'img_debug_steps' in opt['logger'].keys() else 50
else:
self.netF = networks.define_F(use_bn=False).to(self.device)
self.cri_fea = nn.L1Loss().to(self.device)
2020-07-27 04:44:24 +00:00
2020-08-26 00:11:53 +00:00
#self.print_network() # print network
2019-08-23 13:42:47 +00:00
self.load() # load G and D if needed
self.load_random_corruptor()
2019-08-23 13:42:47 +00:00
# Setting this to false triggers SRGAN to call the models update_model() function on the first iteration.
self.updated = True
2019-08-23 13:42:47 +00:00
def feed_data(self, data, need_GT=True):
_profile = True
if _profile:
from time import time
_t = time()
# Corrupt the data with the given corruptor, if specified.
self.fed_LQ = data['LQ'].to(self.device)
if self.netC and random.random() < self.corruptor_usage_prob:
with torch.no_grad():
corrupted_L = self.netC(self.fed_LQ)[0].detach()
else:
corrupted_L = self.fed_LQ
self.var_L = torch.chunk(corrupted_L, chunks=self.mega_batch_factor, dim=0)
2019-08-23 13:42:47 +00:00
if need_GT:
2020-05-02 23:46:59 +00:00
self.var_H = [t.to(self.device) for t in torch.chunk(data['GT'], chunks=self.mega_batch_factor, dim=0)]
input_ref = data['ref'] if 'ref' in data.keys() else data['GT']
2020-05-02 23:46:59 +00:00
self.var_ref = [t.to(self.device) for t in torch.chunk(input_ref, chunks=self.mega_batch_factor, dim=0)]
input_pix = data['PIX'] if 'pix' in data.keys() else data['GT']
self.pix = [t.to(self.device) for t in torch.chunk(input_pix, chunks=self.mega_batch_factor, dim=0)]
2019-08-23 13:42:47 +00:00
if 'GAN' in data.keys():
self.gan_img = [t.to(self.device) for t in torch.chunk(data['GAN'], chunks=self.mega_batch_factor, dim=0)]
else:
# If not provided, use provided LQ for anyplace where the GAN would have been used.
self.gan_img = self.var_L
if not self.updated:
self.netG.module.update_model(self.optimizer_G, self.schedulers[0])
self.updated = True
2019-08-23 13:42:47 +00:00
def optimize_parameters(self, step):
_profile = False
if _profile:
from time import time
_t = time()
# Some generators have variants depending on the current step.
if hasattr(self.netG.module, "update_for_step"):
self.netG.module.update_for_step(step, os.path.join(self.opt['path']['models'], ".."))
if hasattr(self.netD.module, "update_for_step"):
self.netD.module.update_for_step(step, os.path.join(self.opt['path']['models'], ".."))
2019-08-23 13:42:47 +00:00
# G
for p in self.netD.parameters():
p.requires_grad = False
if self.spsr_enabled:
for p in self.netD_grad.parameters():
p.requires_grad = False
2019-08-23 13:42:47 +00:00
self.swapout_D(step)
self.swapout_G(step)
# Turning off G-grad is required to enable mega-batching and D_update_ratio to work together for some reason.
if step % self.D_update_ratio == 0 and step >= self.D_init_iters:
if self.spsr_enabled and self.branch_pretrain and step < self.branch_init_iters:
for k, v in self.netG.named_parameters():
2020-08-08 03:03:39 +00:00
if v.dtype != torch.int64 and v.dtype != torch.bool:
v.requires_grad = '_branch_pretrain' in k
else:
for p in self.netG.parameters():
if p.dtype != torch.int64 and p.dtype != torch.bool:
p.requires_grad = True
else:
for p in self.netG.parameters():
p.requires_grad = False
# Calculate a standard deviation for the gaussian noise to be applied to the discriminator, termed noise-theta.
2020-05-15 02:45:06 +00:00
if self.D_noise_final == 0:
noise_theta = 0
else:
2020-05-15 02:45:06 +00:00
noise_theta = (self.D_noise_theta - self.D_noise_theta_floor) * (self.D_noise_final - min(step, self.D_noise_final)) / self.D_noise_final + self.D_noise_theta_floor
if _profile:
print("Misc setup %f" % (time() - _t,))
_t = time()
2020-08-23 23:32:03 +00:00
self.optimizer_G.zero_grad()
self.fake_GenOut = []
self.fea_GenOut = []
self.fake_H = []
self.spsr_grad_GenOut = []
var_ref_skips = []
for var_L, var_LGAN, var_H, var_ref, pix in zip(self.var_L, self.gan_img, self.var_H, self.var_ref, self.pix):
if self.spsr_enabled:
using_gan_img = False
# SPSR models have outputs from three different branches.
fake_H_branch, fake_GenOut, grad_LR = self.netG(var_L)
fea_GenOut = fake_GenOut
self.spsr_grad_GenOut.append(fake_H_branch)
# Get image gradients for later use.
fake_H_grad = self.get_grad_nopadding(fake_GenOut)
else:
if random.random() > self.gan_lq_img_use_prob:
fea_GenOut, fake_GenOut = self.netG(var_L)
using_gan_img = False
else:
fea_GenOut, fake_GenOut = self.netG(var_LGAN)
using_gan_img = True
if _profile:
print("Gen forward %f" % (time() - _t,))
_t = time()
self.fake_GenOut.append(fake_GenOut.detach())
self.fea_GenOut.append(fea_GenOut.detach())
2020-05-02 23:46:59 +00:00
l_g_total = 0
if step % self.D_update_ratio == 0 and step >= self.D_init_iters:
fea_w = self.l_fea_sched.get_weight_for_step(step)
l_g_pix_log = None
l_g_fea_log = None
l_g_fdpl = None
l_g_fea_log = None
if self.cri_pix and not using_gan_img: # pixel loss
l_g_pix = self.l_pix_w * self.cri_pix(fea_GenOut, pix)
l_g_pix_log = l_g_pix / self.l_pix_w
2020-05-02 23:46:59 +00:00
l_g_total += l_g_pix
if self.spsr_enabled and self.cri_pix_grad: # gradient pixel loss
if self.disjoint_data:
grad_truth = self.get_grad_nopadding(var_L)
grad_pred = F.interpolate(fake_H_grad, size=grad_truth.shape[2:], mode="nearest")
else:
grad_truth = self.get_grad_nopadding(var_H)
grad_pred = fake_H_grad
l_g_pix_grad = self.l_pix_grad_w * self.cri_pix_grad(grad_pred, grad_truth)
l_g_total += l_g_pix_grad
if self.spsr_enabled and self.cri_pix_branch: # branch pixel loss
if self.disjoint_data:
grad_truth = self.get_grad_nopadding(var_L)
grad_pred = F.interpolate(fake_H_branch, size=grad_truth.shape[2:], mode="nearest")
else:
grad_truth = self.get_grad_nopadding(var_H)
grad_pred = fake_H_branch
l_g_pix_grad_branch = self.l_pix_branch_w * self.cri_pix_branch(grad_pred, grad_truth)
l_g_total += l_g_pix_grad_branch
if self.fdpl_enabled and not using_gan_img:
l_g_fdpl = self.cri_fdpl(fea_GenOut, pix)
l_g_total += l_g_fdpl * self.fdpl_weight
if self.cri_fea and not using_gan_img and fea_w > 0: # feature loss
if self.lr_netF is not None:
real_fea = self.lr_netF(var_L, interpolate_factor=self.opt['scale'])
2020-08-23 23:32:03 +00:00
elif self.use_corrupted_feature_input:
cor_Pix = F.interpolate(self.feature_corruptor(pix), size=var_L.shape[2:])
real_fea = self.netF(cor_Pix).detach()
else:
real_fea = self.netF(pix).detach()
2020-08-23 23:32:03 +00:00
if self.use_corrupted_feature_input:
fake_fea = self.netF(F.interpolate(self.feature_corruptor(fea_GenOut), size=var_L.shape[2:]))
else:
fake_fea = self.netF(fea_GenOut)
l_g_fea = fea_w * self.cri_fea(fake_fea, real_fea)
l_g_fea_log = l_g_fea / fea_w
2020-05-02 23:46:59 +00:00
l_g_total += l_g_fea
2019-08-23 13:42:47 +00:00
if _profile:
print("Fea forward %f" % (time() - _t,))
_t = time()
# Note to future self: The BCELoss(0, 1) and BCELoss(0, 0) = .6931
# Effectively this means that the generator has only completely "won" when l_d_real and l_d_fake is
# equal to this value. If I ever come up with an algorithm that tunes fea/gan weights automatically,
# it should target this
2020-07-31 23:03:20 +00:00
l_g_fix_disc = torch.zeros(1, requires_grad=False, device=self.device).squeeze()
for fixed_disc in self.fixed_disc_nets:
2020-07-31 22:59:23 +00:00
weight = fixed_disc.module.fdisc_weight
real_fea = fixed_disc(pix).detach()
fake_fea = fixed_disc(fea_GenOut)
2020-07-31 23:03:20 +00:00
l_g_fix_disc = l_g_fix_disc + weight * self.cri_fea(fake_fea, real_fea)
l_g_total += l_g_fix_disc
if self.l_gan_w > 0:
2020-08-08 03:03:39 +00:00
if self.opt['train']['gan_type'] in ['gan', 'pixgan', 'pixgan_fea', 'crossgan']:
if self.opt['train']['gan_type'] == 'crossgan':
pred_g_fake = self.netD(fake_GenOut, var_L)
else:
pred_g_fake = self.netD(fake_GenOut)
l_g_gan = self.l_gan_w * self.cri_gan(pred_g_fake, True)
elif self.opt['train']['gan_type'] == 'ragan':
pred_d_real = self.netD(var_ref).detach()
pred_g_fake = self.netD(fake_GenOut)
l_g_gan = self.l_gan_w * (
self.cri_gan(pred_d_real - torch.mean(pred_g_fake), False) +
self.cri_gan(pred_g_fake - torch.mean(pred_d_real), True)) / 2
l_g_gan_log = l_g_gan / self.l_gan_w
l_g_total += l_g_gan
2019-08-23 13:42:47 +00:00
if self.spsr_enabled and self.cri_grad_gan:
2020-08-08 03:03:39 +00:00
if self.opt['train']['gan_type'] == 'crossgan':
2020-08-12 14:45:49 +00:00
pred_g_fake_grad = self.netD_grad(fake_H_grad, var_L)
pred_g_fake_grad_branch = self.netD_grad(fake_H_branch, var_L)
2020-08-08 03:03:39 +00:00
else:
2020-08-12 14:45:49 +00:00
pred_g_fake_grad = self.netD_grad(fake_H_grad)
pred_g_fake_grad_branch = self.netD_grad(fake_H_branch)
2020-08-08 03:03:39 +00:00
if self.opt['train']['gan_type'] in ['gan', 'pixgan', 'pixgan_fea', 'crossgan']:
l_g_gan_grad = self.l_gan_grad_w * self.cri_grad_gan(pred_g_fake_grad, True)
2020-08-06 18:11:40 +00:00
l_g_gan_grad_branch = self.l_gan_grad_w * self.cri_grad_gan(pred_g_fake_grad_branch, True)
elif self.opt['train']['gan_type'] == 'ragan':
2020-08-05 22:47:21 +00:00
pred_g_real_grad = self.netD_grad(self.get_grad_nopadding(var_ref)).detach()
l_g_gan_grad = self.l_gan_grad_w * (
self.cri_gan(pred_g_real_grad - torch.mean(pred_g_fake_grad), False) +
self.cri_gan(pred_g_fake_grad - torch.mean(pred_g_real_grad), True)) / 2
l_g_gan_grad_branch = self.l_gan_grad_w * (
self.cri_gan(pred_g_real_grad - torch.mean(pred_g_fake_grad_branch), False) +
self.cri_gan(pred_g_fake_grad_branch - torch.mean(pred_g_real_grad), True)) / 2
l_g_total += l_g_gan_grad + l_g_gan_grad_branch
# Scale the loss down by the batch factor.
l_g_total_log = l_g_total
l_g_total = l_g_total / self.mega_batch_factor
2020-05-02 23:46:59 +00:00
with amp.scale_loss(l_g_total, self.optimizer_G, loss_id=0) as l_g_total_scaled:
l_g_total_scaled.backward()
if _profile:
print("Gen backward %f" % (time() - _t,))
_t = time()
2020-05-02 23:46:59 +00:00
self.optimizer_G.step()
2019-08-23 13:42:47 +00:00
if _profile:
print("Gen step %f" % (time() - _t,))
_t = time()
2019-08-23 13:42:47 +00:00
# D
2020-07-27 04:44:24 +00:00
if self.l_gan_w > 0 and step >= self.G_warmup:
for p in self.netD.parameters():
if p.dtype != torch.int64 and p.dtype != torch.bool:
p.requires_grad = True
2019-08-23 13:42:47 +00:00
noise = torch.randn_like(var_ref) * noise_theta
noise.to(self.device)
2020-07-10 15:45:34 +00:00
real_disc_images = []
fake_disc_images = []
2020-08-08 03:03:39 +00:00
for fake_GenOut, var_LGAN, var_L, var_H, var_ref, pix in zip(self.fake_GenOut, self.gan_img, self.var_L, self.var_H, self.var_ref, self.pix):
if random.random() > self.gan_lq_img_use_prob:
fake_H = fake_GenOut.clone().detach().requires_grad_(False)
else:
# Re-compute generator outputs with the GAN inputs.
with torch.no_grad():
if self.spsr_enabled:
_, fake_H, _ = self.netG(var_LGAN)
else:
_, fake_H = self.netG(var_LGAN)
fake_H = fake_H.detach()
if _profile:
print("Gen forward for disc %f" % (time() - _t,))
_t = time()
# Apply noise to the inputs to slow discriminator convergence.
var_ref = var_ref + noise
fake_H = fake_H + noise
2020-07-16 16:18:30 +00:00
l_d_fea_real = 0
l_d_fea_fake = 0
2020-08-04 16:25:13 +00:00
self.optimizer_D.zero_grad()
if self.opt['train']['gan_type'] == 'pixgan_fea':
# Compute a feature loss which is added to the GAN loss computed later to guide the discriminator better.
disc_fea_scale = .1
_, fea_real = self.netD(var_ref, output_feature_vector=True)
actual_fea = self.netF(var_ref)
l_d_fea_real = self.cri_fea(fea_real, actual_fea) * disc_fea_scale / self.mega_batch_factor
_, fea_fake = self.netD(fake_H, output_feature_vector=True)
actual_fea = self.netF(fake_H)
l_d_fea_fake = self.cri_fea(fea_fake, actual_fea) * disc_fea_scale / self.mega_batch_factor
2020-08-08 03:03:39 +00:00
if self.opt['train']['gan_type'] == 'crossgan':
# need to forward and backward separately, since batch norm statistics differ
2020-08-08 03:03:39 +00:00
# real
pred_d_real = self.netD(var_ref, var_L)
l_d_real = self.cri_gan(pred_d_real, True)
l_d_real_log = l_d_real
# fake
pred_d_fake = self.netD(fake_H, var_L)
l_d_fake = self.cri_gan(pred_d_fake, False)
l_d_fake_log = l_d_fake
# mismatched
mismatched_L = torch.roll(var_L, shifts=1, dims=0)
pred_d_real_mismatched = self.netD(var_ref, mismatched_L)
pred_d_fake_mismatched = self.netD(fake_H, mismatched_L)
l_d_mismatched = (self.cri_gan(pred_d_real_mismatched, False) + self.cri_gan(pred_d_fake_mismatched, False)) / 2
l_d_total = (l_d_real + l_d_fake + l_d_mismatched) / 3
l_d_total = l_d_total / self.mega_batch_factor
with amp.scale_loss(l_d_total, self.optimizer_D, loss_id=1) as l_d_total_scaled:
l_d_total_scaled.backward()
elif self.opt['train']['gan_type'] == 'gan':
# real
pred_d_real = self.netD(var_ref)
l_d_real = self.cri_gan(pred_d_real, True) / self.mega_batch_factor
l_d_real_log = l_d_real * self.mega_batch_factor
# fake
pred_d_fake = self.netD(fake_H)
l_d_fake = self.cri_gan(pred_d_fake, False) / self.mega_batch_factor
l_d_fake_log = l_d_fake * self.mega_batch_factor
2020-08-04 16:25:13 +00:00
l_d_total = (l_d_real + l_d_fake) / 2
with amp.scale_loss(l_d_total, self.optimizer_D, loss_id=1) as l_d_total_scaled:
l_d_total_scaled.backward()
2020-08-08 03:03:39 +00:00
elif 'pixgan' in self.opt['train']['gan_type']:
2020-07-31 22:38:14 +00:00
pixdisc_channels, pixdisc_output_reduction = self.netD.module.pixgan_parameters()
disc_output_shape = (var_ref.shape[0], pixdisc_channels, var_ref.shape[2] // pixdisc_output_reduction, var_ref.shape[3] // pixdisc_output_reduction)
b, _, w, h = var_ref.shape
real = torch.ones((b, pixdisc_channels, w, h), device=var_ref.device)
fake = torch.zeros((b, pixdisc_channels, w, h), device=var_ref.device)
2020-08-08 03:03:39 +00:00
if not self.disjoint_data:
# randomly determine portions of the image to swap to keep the discriminator honest.
SWAP_MAX_DIM = w // 4
SWAP_MIN_DIM = 16
assert SWAP_MAX_DIM > 0
if random.random() > .5: # Make this only happen half the time. Earlier experiments had it happen
# more often and the model was "cheating" by using the presence of
# easily discriminated fake swaps to count the entire generated image
# as fake.
random_swap_count = random.randint(0, 4)
for i in range(random_swap_count):
# Make the swap across fake_H and var_ref
swap_x, swap_y = random.randint(0, w - SWAP_MIN_DIM), random.randint(0, h - SWAP_MIN_DIM)
swap_w, swap_h = random.randint(SWAP_MIN_DIM, SWAP_MAX_DIM), random.randint(SWAP_MIN_DIM, SWAP_MAX_DIM)
if swap_x + swap_w > w:
swap_w = w - swap_x
if swap_y + swap_h > h:
swap_h = h - swap_y
t = fake_H[:, :, swap_x:(swap_x+swap_w), swap_y:(swap_y+swap_h)].clone()
fake_H[:, :, swap_x:(swap_x+swap_w), swap_y:(swap_y+swap_h)] = var_ref[:, :, swap_x:(swap_x+swap_w), swap_y:(swap_y+swap_h)]
var_ref[:, :, swap_x:(swap_x+swap_w), swap_y:(swap_y+swap_h)] = t
real[:, :, swap_x:(swap_x+swap_w), swap_y:(swap_y+swap_h)] = 0.0
fake[:, :, swap_x:(swap_x+swap_w), swap_y:(swap_y+swap_h)] = 1.0
# Interpolate down to the dimensionality that the discriminator uses.
real = F.interpolate(real, size=disc_output_shape[2:], mode="bilinear", align_corners=False)
fake = F.interpolate(fake, size=disc_output_shape[2:], mode="bilinear", align_corners=False)
# We're also assuming that this is exactly how the flattened discriminator output is generated.
real = real.view(-1, 1)
fake = fake.view(-1, 1)
# real
pred_d_real = self.netD(var_ref)
l_d_real = self.cri_gan(pred_d_real, real) / self.mega_batch_factor
l_d_real_log = l_d_real * self.mega_batch_factor
l_d_real += l_d_fea_real
# fake
pred_d_fake = self.netD(fake_H)
l_d_fake = self.cri_gan(pred_d_fake, fake) / self.mega_batch_factor
l_d_fake_log = l_d_fake * self.mega_batch_factor
l_d_fake += l_d_fea_fake
2020-08-04 16:25:13 +00:00
l_d_total = (l_d_real + l_d_fake) / 2
with amp.scale_loss(l_d_total, self.optimizer_D, loss_id=1) as l_d_total_scaled:
l_d_total_scaled.backward()
2020-07-10 15:45:34 +00:00
pdr = pred_d_real.detach() + torch.abs(torch.min(pred_d_real))
pdr = pdr / torch.max(pdr)
real_disc_images.append(pdr.view(disc_output_shape))
pdf = pred_d_fake.detach() + torch.abs(torch.min(pred_d_fake))
pdf = pdf / torch.max(pdf)
fake_disc_images.append(pdf.view(disc_output_shape))
elif self.opt['train']['gan_type'] == 'ragan':
2020-08-05 22:47:21 +00:00
pred_d_fake = self.netD(fake_H)
pred_d_real = self.netD(var_ref)
l_d_real = self.cri_gan(pred_d_real - torch.mean(pred_d_fake), True)
2020-08-05 17:03:06 +00:00
l_d_real_log = l_d_real
l_d_fake = self.cri_gan(pred_d_fake - torch.mean(pred_d_real), False)
2020-08-05 17:03:06 +00:00
l_d_fake_log = l_d_fake
l_d_total = (l_d_real + l_d_fake) / 2
l_d_total /= self.mega_batch_factor
with amp.scale_loss(l_d_total, self.optimizer_D, loss_id=1) as l_d_total_scaled:
l_d_total_scaled.backward()
var_ref_skips.append(var_ref.detach())
self.fake_H.append(fake_H.detach())
self.optimizer_D.step()
2019-08-23 13:42:47 +00:00
if _profile:
print("Disc step %f" % (time() - _t,))
_t = time()
# D_grad.
if self.spsr_enabled and self.cri_grad_gan and step >= self.G_warmup:
for p in self.netD_grad.parameters():
p.requires_grad = True
self.optimizer_D_grad.zero_grad()
2020-08-12 14:45:49 +00:00
for var_L, var_ref, fake_H, fake_H_grad_branch in zip(self.var_L, var_ref_skips, self.fake_H, self.spsr_grad_GenOut):
2020-08-05 22:47:21 +00:00
fake_H_grad = self.get_grad_nopadding(fake_H).detach()
var_ref_grad = self.get_grad_nopadding(var_ref)
fake_H_grad_branch = fake_H_grad_branch.detach() + noise
2020-08-12 14:45:49 +00:00
if self.opt['train']['gan_type'] == 'crossgan':
pred_d_real_grad = self.netD_grad(var_ref_grad, var_L)
pred_d_fake_grad = self.netD_grad(fake_H_grad, var_L) # Tensor already detached above.
# var_ref and fake_H already has noise added to it. We **must** add noise to fake_H_grad_branch too.
pred_d_fake_grad_branch = self.netD_grad(fake_H_grad_branch, var_L)
else:
pred_d_real_grad = self.netD_grad(var_ref_grad)
pred_d_fake_grad = self.netD_grad(fake_H_grad) # Tensor already detached above.
# var_ref and fake_H already has noise added to it. We **must** add noise to fake_H_grad_branch too.
pred_d_fake_grad_branch = self.netD_grad(fake_H_grad_branch)
if self.opt['train']['gan_type'] == 'gan' or self.opt['train']['gan_type'] == 'crossgan':
l_d_real_grad = self.cri_gan(pred_d_real_grad, True)
l_d_fake_grad = (self.cri_gan(pred_d_fake_grad, False) + self.cri_gan(pred_d_fake_grad_branch, False)) / 2
elif self.opt['train']['gan_type'] == 'pixgan':
real = torch.ones_like(pred_d_real_grad)
fake = torch.zeros_like(pred_d_fake_grad)
2020-08-04 16:25:13 +00:00
l_d_real_grad = self.cri_grad_gan(pred_d_real_grad, real)
l_d_fake_grad = (self.cri_grad_gan(pred_d_fake_grad, fake) + \
self.cri_grad_gan(pred_d_fake_grad_branch, fake)) / 2
elif self.opt['train']['gan_type'] == 'ragan':
2020-08-05 22:47:21 +00:00
l_d_real_grad = self.cri_grad_gan(pred_d_real_grad - torch.mean(pred_d_fake_grad), True)
l_d_fake_grad = (self.cri_grad_gan(pred_d_fake_grad - torch.mean(pred_d_real_grad), False) + \
self.cri_grad_gan(pred_d_fake_grad_branch - torch.mean(pred_d_real_grad), False)) / 2
l_d_total_grad = (l_d_real_grad + l_d_fake_grad) / 2
l_d_total_grad /= self.mega_batch_factor
with amp.scale_loss(l_d_total_grad, self.optimizer_D_grad, loss_id=2) as l_d_total_grad_scaled:
l_d_total_grad_scaled.backward()
self.optimizer_D_grad.step()
2020-05-02 23:46:59 +00:00
# Log sample images from first microbatch.
2020-08-25 22:00:49 +00:00
if step % self.img_debug_steps == 0 and self.rank <= 0:
sample_save_path = os.path.join(self.opt['path']['models'], "..", "temp")
os.makedirs(os.path.join(sample_save_path, "hr"), exist_ok=True)
os.makedirs(os.path.join(sample_save_path, "lr"), exist_ok=True)
os.makedirs(os.path.join(sample_save_path, "gen_fea"), exist_ok=True)
os.makedirs(os.path.join(sample_save_path, "gen"), exist_ok=True)
os.makedirs(os.path.join(sample_save_path, "disc_fake"), exist_ok=True)
os.makedirs(os.path.join(sample_save_path, "pix"), exist_ok=True)
2020-07-10 15:45:34 +00:00
os.makedirs(os.path.join(sample_save_path, "disc"), exist_ok=True)
os.makedirs(os.path.join(sample_save_path, "ref"), exist_ok=True)
if self.spsr_enabled:
os.makedirs(os.path.join(sample_save_path, "gen_grad"), exist_ok=True)
# fed_LQ is not chunked.
for i in range(self.mega_batch_factor):
utils.save_image(self.var_H[i].cpu(), os.path.join(sample_save_path, "hr", "%05i_%02i.png" % (step, i)))
utils.save_image(self.var_L[i].cpu(), os.path.join(sample_save_path, "lr", "%05i_%02i.png" % (step, i)))
utils.save_image(self.pix[i].cpu(), os.path.join(sample_save_path, "pix", "%05i_%02i.png" % (step, i)))
utils.save_image(self.fake_GenOut[i].cpu(), os.path.join(sample_save_path, "gen", "%05i_%02i.png" % (step, i)))
utils.save_image(self.fea_GenOut[i].cpu(), os.path.join(sample_save_path, "gen_fea", "%05i_%02i.png" % (step, i)))
if self.spsr_enabled:
utils.save_image(self.spsr_grad_GenOut[i].cpu(), os.path.join(sample_save_path, "gen_grad", "%05i_%02i.png" % (step, i)))
2020-07-27 04:44:24 +00:00
if self.l_gan_w > 0 and step >= self.G_warmup and 'pixgan' in self.opt['train']['gan_type']:
utils.save_image(var_ref_skips[i].cpu(), os.path.join(sample_save_path, "ref", "%05i_%02i.png" % (step, i)))
utils.save_image(self.fake_H[i], os.path.join(sample_save_path, "disc_fake", "fake%05i_%02i.png" % (step, i)))
utils.save_image(F.interpolate(fake_disc_images[i], scale_factor=4), os.path.join(sample_save_path, "disc", "fake%05i_%02i.png" % (step, i)))
utils.save_image(F.interpolate(real_disc_images[i], scale_factor=4), os.path.join(sample_save_path, "disc", "real%05i_%02i.png" % (step, i)))
2020-05-02 23:46:59 +00:00
# Log metrics
if step % self.D_update_ratio == 0 and step >= self.D_init_iters:
if self.cri_pix and l_g_pix_log is not None:
self.add_log_entry('l_g_pix', l_g_pix_log.detach().item())
if self.fdpl_enabled and l_g_fdpl is not None:
self.add_log_entry('l_g_fdpl', l_g_fdpl.detach().item())
if self.cri_fea and l_g_fea_log is not None:
self.add_log_entry('feature_weight', fea_w)
self.add_log_entry('l_g_fea', l_g_fea_log.detach().item())
self.add_log_entry('l_g_fix_disc', l_g_fix_disc.detach().item())
if self.l_gan_w > 0:
self.add_log_entry('l_g_gan', l_g_gan_log.detach().item())
self.add_log_entry('l_g_total', l_g_total_log.detach().item())
2020-07-27 04:25:59 +00:00
if self.opt['train']['gan_type'] == 'pixgan_fea':
self.add_log_entry('l_d_fea_fake', l_d_fea_fake.detach().item() * self.mega_batch_factor)
self.add_log_entry('l_d_fea_real', l_d_fea_real.detach().item() * self.mega_batch_factor)
self.add_log_entry('l_d_fake_total', l_d_fake.detach().item() * self.mega_batch_factor)
self.add_log_entry('l_d_real_total', l_d_real.detach().item() * self.mega_batch_factor)
2020-08-08 03:03:39 +00:00
if self.opt['train']['gan_type'] == 'crossgan':
self.add_log_entry('l_d_mismatched', l_d_mismatched.detach().item())
if self.spsr_enabled:
2020-08-03 22:21:05 +00:00
if self.cri_pix_grad:
self.add_log_entry('l_g_pix_grad_branch', l_g_pix_grad.detach().item())
if self.cri_pix_branch:
self.add_log_entry('l_g_pix_grad_branch', l_g_pix_grad_branch.detach().item())
if self.cri_grad_gan:
self.add_log_entry('l_g_gan_grad', l_g_gan_grad.detach().item() / self.l_gan_grad_w)
self.add_log_entry('l_g_gan_grad_branch', l_g_gan_grad_branch.detach().item() / self.l_gan_grad_w)
2020-07-27 04:44:24 +00:00
if self.l_gan_w > 0 and step >= self.G_warmup:
self.add_log_entry('l_d_real', l_d_real_log.detach().item())
self.add_log_entry('l_d_fake', l_d_fake_log.detach().item())
2020-07-16 16:45:50 +00:00
self.add_log_entry('D_fake', torch.mean(pred_d_fake.detach()))
self.add_log_entry('D_diff', torch.mean(pred_d_fake.detach()) - torch.mean(pred_d_real.detach()))
2020-08-25 22:42:59 +00:00
if self.spsr_enabled and self.cri_grad_gan:
self.add_log_entry('l_d_real_grad', l_d_real_grad.detach().item())
self.add_log_entry('l_d_fake_grad', l_d_fake_grad.detach().item())
2020-08-05 22:47:21 +00:00
self.add_log_entry('D_fake_grad', torch.mean(pred_d_fake_grad.detach()))
self.add_log_entry('D_diff_grad', torch.mean(pred_d_fake_grad.detach()) - torch.mean(pred_d_real_grad.detach()))
# Log learning rates.
for i, pg in enumerate(self.optimizer_G.param_groups):
self.add_log_entry('gen_lr_%i' % (i,), pg['lr'])
for i, pg in enumerate(self.optimizer_D.param_groups):
self.add_log_entry('disc_lr_%i' % (i,), pg['lr'])
if step % self.corruptor_swapout_steps == 0 and step > 0:
self.load_random_corruptor()
# Allows the log to serve as an easy-to-use rotating buffer.
def add_log_entry(self, key, value):
key_it = "%s_it" % (key,)
log_rotating_buffer_size = 50
if key not in self.log_dict.keys():
self.log_dict[key] = []
self.log_dict[key_it] = 0
if len(self.log_dict[key]) < log_rotating_buffer_size:
self.log_dict[key].append(value)
else:
self.log_dict[key][self.log_dict[key_it] % log_rotating_buffer_size] = value
self.log_dict[key_it] += 1
def pick_rand_prev_model(self, model_suffix):
previous_models = glob.glob(os.path.join(self.opt['path']['models'], "*_%s.pth" % (model_suffix,)))
if len(previous_models) <= 1:
return None
# Just a note: this intentionally includes the swap model in the list of possibilities.
return previous_models[random.randint(0, len(previous_models)-1)]
2020-07-03 21:18:57 +00:00
def compute_fea_loss(self, real, fake):
with torch.no_grad():
real = real.unsqueeze(dim=0).to(self.device)
fake = fake.unsqueeze(dim=0).to(self.device)
2020-07-03 21:18:57 +00:00
real_fea = self.netF(real).detach()
fake_fea = self.netF(fake)
return self.cri_fea(fake_fea, real_fea).item()
# Called before verification/checkpoint to ensure we're using the real models and not a swapout variant.
def force_restore_swapout(self):
if self.swapout_D_duration > 0:
logger.info("Swapping back to current D model: %s" % (self.stashed_D,))
self.load_network(self.stashed_D, self.netD, self.opt['path']['strict_load'])
self.stashed_D = None
self.swapout_D_duration = 0
if self.swapout_G_duration > 0:
logger.info("Swapping back to current G model: %s" % (self.stashed_G,))
self.load_network(self.stashed_G, self.netG, self.opt['path']['strict_load'])
self.stashed_G = None
self.swapout_G_duration = 0
def swapout_D(self, step):
if self.swapout_D_duration > 0:
self.swapout_D_duration -= 1
if self.swapout_D_duration == 0:
# Swap back.
logger.info("Swapping back to current D model: %s" % (self.stashed_D,))
self.load_network(self.stashed_D, self.netD, self.opt['path']['strict_load'])
self.stashed_D = None
elif self.swapout_D_freq != 0 and step % self.swapout_D_freq == 0:
swapped_model = self.pick_rand_prev_model('D')
if swapped_model is not None:
logger.info("Swapping to previous D model: %s" % (swapped_model,))
self.stashed_D = self.save_network(self.netD, 'D', 'swap_model')
self.load_network(swapped_model, self.netD, self.opt['path']['strict_load'])
self.swapout_D_duration = self.swapout_duration
def swapout_G(self, step):
if self.swapout_G_duration > 0:
self.swapout_G_duration -= 1
if self.swapout_G_duration == 0:
# Swap back.
logger.info("Swapping back to current G model: %s" % (self.stashed_G,))
self.load_network(self.stashed_G, self.netG, self.opt['path']['strict_load'])
self.stashed_G = None
elif self.swapout_G_freq != 0 and step % self.swapout_G_freq == 0:
swapped_model = self.pick_rand_prev_model('G')
if swapped_model is not None:
logger.info("Swapping to previous G model: %s" % (swapped_model,))
self.stashed_G = self.save_network(self.netG, 'G', 'swap_model')
self.load_network(swapped_model, self.netG, self.opt['path']['strict_load'])
self.swapout_G_duration = self.swapout_duration
2019-08-23 13:42:47 +00:00
def test(self):
self.netG.eval()
with torch.no_grad():
if self.spsr_enabled:
self.fake_H_branch = []
self.fake_GenOut = []
self.grad_LR = []
fake_H_branch, fake_GenOut, grad_LR = self.netG(self.var_L[0])
self.fake_H_branch.append(fake_H_branch)
self.fake_GenOut.append(fake_GenOut)
self.grad_LR.append(grad_LR)
else:
self.fake_GenOut = [self.netG(self.var_L[0])]
2019-08-23 13:42:47 +00:00
self.netG.train()
# Fetches a summary of the log.
def get_current_log(self, step):
return_log = {}
for k in self.log_dict.keys():
if not isinstance(self.log_dict[k], list):
continue
return_log[k] = sum(self.log_dict[k]) / len(self.log_dict[k])
# Some generators can do their own metric logging.
if hasattr(self.netG.module, "get_debug_values"):
return_log.update(self.netG.module.get_debug_values(step))
if hasattr(self.netD.module, "get_debug_values"):
return_log.update(self.netD.module.get_debug_values(step))
return return_log
2019-08-23 13:42:47 +00:00
def get_current_visuals(self, need_GT=True):
out_dict = OrderedDict()
out_dict['LQ'] = self.var_L[0].detach().float().cpu()
gen_batch = self.fake_GenOut[0]
if isinstance(gen_batch, tuple):
gen_batch = gen_batch[0]
out_dict['rlt'] = gen_batch.detach().float().cpu()
2019-08-23 13:42:47 +00:00
if need_GT:
out_dict['GT'] = self.var_H[0].detach().float().cpu()
if self.spsr_enabled:
out_dict['SR_branch'] = self.fake_H_branch[0].float().cpu()
out_dict['LR_grad'] = self.grad_LR[0].float().cpu()
2019-08-23 13:42:47 +00:00
return out_dict
def print_network(self):
# Generator
s, n = self.get_network_description(self.netG)
if isinstance(self.netG, nn.DataParallel) or isinstance(self.netG, DistributedDataParallel):
net_struc_str = '{} - {}'.format(self.netG.__class__.__name__,
self.netG.module.__class__.__name__)
else:
net_struc_str = '{}'.format(self.netG.__class__.__name__)
if self.rank <= 0:
logger.info('Network G structure: {}, with parameters: {:,d}'.format(net_struc_str, n))
logger.info(s)
if self.is_train:
# Discriminator
s, n = self.get_network_description(self.netD)
if isinstance(self.netD, nn.DataParallel) or isinstance(self.netD,
DistributedDataParallel):
net_struc_str = '{} - {}'.format(self.netD.__class__.__name__,
self.netD.module.__class__.__name__)
else:
net_struc_str = '{}'.format(self.netD.__class__.__name__)
if self.rank <= 0:
logger.info('Network D structure: {}, with parameters: {:,d}'.format(
net_struc_str, n))
logger.info(s)
if self.cri_fea: # F, Perceptual Network
s, n = self.get_network_description(self.netF)
if isinstance(self.netF, nn.DataParallel) or isinstance(
self.netF, DistributedDataParallel):
net_struc_str = '{} - {}'.format(self.netF.__class__.__name__,
self.netF.module.__class__.__name__)
else:
net_struc_str = '{}'.format(self.netF.__class__.__name__)
if self.rank <= 0:
logger.info('Network F structure: {}, with parameters: {:,d}'.format(
net_struc_str, n))
logger.info(s)
def load(self):
load_path_G = self.opt['path']['pretrain_model_G']
if load_path_G is not None:
logger.info('Loading model for G [{:s}] ...'.format(load_path_G))
self.load_network(load_path_G, self.netG, self.opt['path']['strict_load'])
load_path_D = self.opt['path']['pretrain_model_D']
if self.opt['is_train'] and load_path_D is not None:
logger.info('Loading model for D [{:s}] ...'.format(load_path_D))
self.load_network(load_path_D, self.netD, self.opt['path']['strict_load'])
if self.spsr_enabled:
load_path_D_grad = self.opt['path']['pretrain_model_D_grad']
if self.opt['is_train'] and load_path_D_grad is not None:
logger.info('Loading pretrained model for D_grad [{:s}] ...'.format(load_path_D_grad))
self.load_network(load_path_D_grad, self.netD_grad)
2019-08-23 13:42:47 +00:00
def load_random_corruptor(self):
if self.netC is None:
return
corruptor_files = glob.glob(os.path.join(self.opt['path']['pretrained_corruptors_dir'], "*.pth"))
corruptor_to_load = corruptor_files[random.randint(0, len(corruptor_files)-1)]
logger.info('Swapping corruptor to: %s' % (corruptor_to_load,))
self.load_network(corruptor_to_load, self.netC, self.opt['path']['strict_load'])
2019-08-23 13:42:47 +00:00
def save(self, iter_step):
self.save_network(self.netG, 'G', iter_step)
self.save_network(self.netD, 'D', iter_step)
2020-08-04 16:25:13 +00:00
if self.spsr_enabled:
self.save_network(self.netD_grad, 'D_grad', iter_step)