-
Notifications
You must be signed in to change notification settings - Fork 1
/
solver.py
221 lines (157 loc) · 7.39 KB
/
solver.py
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
import os
import random
import numpy as np
from PIL import Image
import math
import sys
import torch
import torch.nn as nn
import torch.optim as optim
from torch.utils.data import DataLoader
from adamp import AdamP
from tensorboardX import SummaryWriter
from dataset import TrainDataset, TestDataset
import datetime as datetimes
import time as times
from util import *
from ops import *
import shutil
time = datetimes.datetime.now().strftime('%m.%d-%H:%M:%S')
# torch.manual_seed(42)
class Solver():
def __init__(self, model, cfg):
self.device = torch.device("cuda:0" if torch.cuda.is_available() else "cpu")
# if torch.cuda.is_available():
# torch.cuda.manual_seed(42)
self.refiner = model(scale=cfg.scale)
if cfg.loss_fn in ["MSE"]:
self.loss_fn = nn.MSELoss()
elif cfg.loss_fn in ["L1"]:
self.loss_fn = nn.L1Loss()
elif cfg.loss_fn in ["SmoothL1"]:
self.loss_fn = nn.SmoothL1Loss()
self.optim = AdamP(
filter(lambda p: p.requires_grad, self.refiner.parameters()),
cfg.lr)
self.train_data = TrainDataset(cfg.train_data_path,
scale=cfg.scale,
size=cfg.patch_size)
self.train_loader = DataLoader(self.train_data,
batch_size=cfg.batch_size,
num_workers=1,
shuffle=True, drop_last=True)
self.refiner = self.refiner.to(self.device)
self.loss_fn = self.loss_fn
self.folder_name = str(cfg.loss_fn) + '_' + str(cfg.batch_size) + '_' + str(cfg.max_steps)[0] + 'K' + '_' + \
str(cfg.lr) + '_' + str(cfg.scale)
checkpoint_folder = 'logs/{}/checkpoints'.format(self.folder_name)
mkdir(checkpoint_folder)
if cfg.resume:
PATH = os.path.join("logs", self.folder_name, "checkpoints")
all_checkpoints = list(sorted(os.listdir(PATH)))
if len(all_checkpoints) > 0:
PATH = os.path.join(PATH, all_checkpoints[-1])
print("=> loading checkpoint '{}'".format(PATH))
checkpoint = torch.load(PATH)
self.refiner.load_state_dict(checkpoint['model_state_dict'])
self.optim.load_state_dict(checkpoint['optimizer_state_dict'])
self.step = checkpoint['step']
self.best_psnr = checkpoint["best_psnr"]
else:
print("=> no checkpoint at '{}'".format(PATH))
self.best_psnr = 0
self.step = 0
else:
self.best_psnr = 0
self.step = 0
self.cfg = cfg
self.writer = SummaryWriter(log_dir=os.path.join("logs/{}/tensorboard/".format(self.folder_name)))
if cfg.verbose:
num_params = 0
for param in self.refiner.parameters():
num_params += param.nelement()
print("Number of parameters for scale X{}: {}".format(cfg.scale, num_params))
def fit(self):
cfg = self.cfg
refiner = nn.DataParallel(self.refiner,
device_ids=range(cfg.num_gpu))
self.mean_content = 0.
self.mean_l1 = 0.
learning_rate = cfg.lr
while True:
for inputs in self.train_loader:
self.refiner.train()
total_loss = []
scale = cfg.scale
hr, lr = inputs[-1][0], inputs[-1][1]
hr = hr.to(self.device)
lr = lr.to(self.device)
sr_main = refiner(lr, scale)
loss = self.loss_fn(sr_main, hr)
self.optim.zero_grad()
loss.backward()
nn.utils.clip_grad_norm_(self.refiner.parameters(), cfg.clip)
self.optim.step()
self.mean_l1 += loss
learning_rate = self.decay_learning_rate()
for param_group in self.optim.param_groups:
param_group["lr"] = learning_rate
self.step += 1
sys.stdout.write("\r==>>Steps:[%d/ %d] Total:[%.6f] "
% (self.step, cfg.max_steps, loss.item()))
self.writer.add_scalar('Loss', loss.data.cpu().numpy(), global_step=self.step)
if cfg.verbose and self.step % cfg.print_interval == 0:
with open('logs/{}/'.format(self.folder_name) + 'logs.txt', 'a') as f:
PATH = os.path.join('logs/{}/checkpoints/'.format(self.folder_name),
"{}_{:06d}.pth.tar".format(cfg.ckpt_name, self.step))
t1 = times.time()
mean_psnr = self.evaluate(cfg.valid_data_path, scale=cfg.scale, num_step=self.step)
t2 = times.time()
self.writer.add_scalar("PSNR_{}x:".format(scale), mean_psnr, self.step)
print('-- PSNR_x{}: {:.5f} -- Total_Loss: {:.5f}\n'
.format(scale, mean_psnr, (self.mean_l1) / cfg.print_interval))
torch.save({'step': self.step, 'model_state_dict': self.refiner.state_dict(),
'optimizer_state_dict': self.optim.state_dict(), 'best_psnr': self.best_psnr}, PATH)
f.write('Step: {}'
'--> PSNR_x{}:{:.5f} -->{:.3f}m\n'
.format(self.step, scale, mean_psnr, ((t2 - t1)/60)))
self.mean_l1 = 0.
self.mean_content = 0.
if self.step > cfg.max_steps: break
def evaluate(self, test_data_dir, scale=2, num_step=0):
cfg = self.cfg
mean_psnr = 0
self.refiner.eval()
test_data = TestDataset(test_data_dir, scale=scale)
test_loader = DataLoader(test_data, batch_size=1, num_workers=1, shuffle=True)
for step, inputs in enumerate(test_loader):
hr = inputs[0]
lr = inputs[1]
name = inputs[2][0]
lr = lr.to(self.device)
hr = hr.to(self.device)
# run refine process in here!
sr = self.refiner(lr, scale)
psnr = calc_psnr(sr, hr, scale, 1, benchmark=True)
mean_psnr += psnr / len(test_data)
return mean_psnr
def load(self, path):
self.refiner.load_state_dict(torch.load(path))
splited = path.split(".")[0].split("_")[-1]
try:
self.step = int(path.split(".")[0].split("_")[-1])
except ValueError:
self.step = 0
print("Load pretrained {} model".format(path))
def save(self, ckpt_name):
save_path = os.path.join(
'logs/{}/checkpoints/'.format(self.folder_name), "{}_{}.pth".format(ckpt_name, self.step))
torch.save(self.refiner.state_dict(), save_path)
def decay_learning_rate(self):
lr = self.cfg.lr * (0.5 ** (self.step // self.cfg.decay))
return lr
def save_checkpoint(self, is_best, filename='checkpoint.pth.tar'):
save_path = os.path.join(self.cfg.logdir, self.folder_name) + '/'
torch.save(self.refiner, save_path + filename)
if is_best:
shutil.copyfile(save_path + filename, save_path + 'model_best.pth.tar')