-
Notifications
You must be signed in to change notification settings - Fork 1
/
Copy pathtest.py
183 lines (146 loc) · 6.66 KB
/
test.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
import gc
import torch
import torch.nn as nn
import torch.nn.functional as functional
from torch.autograd import Variable
import torchvision
from dataset import MyData, MyTestData
from model import Feature, Deconv
import cv2
from tensorboardX import SummaryWriter
from datetime import datetime
import os
import pdb
from myfunc import make_image_grid
import time
import matplotlib.pyplot as plt
import math
# ECSSD
train_root = '/home/wbm/桌面/未命名文件夹/RFCN-master/ECSSD/Img' # training dataset
val_root = '/home/wbm/桌面/未命名文件夹/RFCN-master/ECSSD/GT' # validation dataset
check_root = './rfcn_parameters' # save checkpoint parameters
val_output_root = '/home/wbm/桌面/未命名文件夹/RFCN-master/rfcn_validation/ECSSD' # save validation results
bsize = 1 # batch size
iter_num = 20 # training iterations
r_num = 3 # recurrence
ptag = '/home/wbm/桌面/未命名文件夹/RFCN-master/ECSSD/MR' # prior map
# # SOD
# train_root = '/home/wbm/桌面/未命名文件夹/RFCN-master/SOD/images' # training dataset
# val_root = '/home/wbm/桌面/未命名文件夹/RFCN-master/SOD/GT' # validation dataset
# check_root = './rfcn_parameters' # save checkpoint parameters
# val_output_root = '/home/wbm/桌面/未命名文件夹/RFCN-master/rfcn_validation/SOD' # save validation results
# bsize = 1 # batch size
# iter_num = 20 # training iterations
# r_num = 3 # recurrence
# ptag = '/home/wbm/桌面/未命名文件夹/RFCN-master/SOD/MR' # prior map
# PASCALS
# train_root = '/home/wbm/桌面/未命名文件夹/RFCN-master/PASCAL-S/images' # training dataset
# val_root = '/home/wbm/桌面/未命名文件夹/RFCN-master/PASCAL-S/groundtruth' # validation dataset
# check_root = './rfcn_parameters' # save checkpoint parameters
# val_output_root = '/home/wbm/桌面/未命名文件夹/RFCN-master/rfcn_validation/PASCAL-S' # save validation results
# bsize = 1 # batch size
# iter_num = 20 # training iterations
# r_num = 3 # recurrence
# ptag = '/home/wbm/桌面/未命名文件夹/RFCN-master/PASCAL-S/MR' # prior map
# train_root = '/home/wbm/桌面/未命名文件夹/RFCN-master/testdata/images' # training dataset
# val_root = '/home/wbm/桌面/未命名文件夹/RFCN-master/testdata/GT' # validation dataset
# check_root = '/home/wbm/桌面/未命名文件夹/RFCN-master/rfcn_parameters' # save checkpoint parameters
# val_output_root = '/home/wbm/桌面/未命名文件夹/RFCN-master/rfcn_validation/testdata' # save validation results
# bsize = 1 # batch size
# iter_num = 20 # training iterations
# r_num = 3 # recurrence
# ptag = '/home/wbm/桌面/未命名文件夹/RFCN-master/testdata/MR' # prior map
std = [.229, .224, .225]
mean = [.485, .456, .406]
# 画图tensorboardx
os.system('rm -rf ./runs/*')
writer = SummaryWriter('./runs/'+datetime.now().strftime('%B%d %H:%M:%S'))
if not os.path.exists('./runs'):
os.mkdir('./runs')
if not os.path.exists(check_root):
os.mkdir(check_root)
if not os.path.exists(val_output_root):
os.mkdir(val_output_root)
# models
feature = Feature()
# feature.cuda()
deconv = Deconv()
# deconv.cuda()
# 参数部分，直接在此处调
# 使用保存的模型
feature.load_state_dict(torch.load('/home/wbm/桌面/未命名文件夹/RFCN-master/rfcn_parameters/feature-epoch-0-step-800.pth'))
deconv.load_state_dict(torch.load('/home/wbm/桌面/未命名文件夹/RFCN-master/rfcn_parameters/deconv-epoch-0-step-800.pth'))
train_loader = torch.utils.data.DataLoader(
MyData(train_root, transform=True, ptag=ptag),
batch_size=bsize, shuffle=True, num_workers=4, pin_memory=True)
val_loader = torch.utils.data.DataLoader(
MyTestData(val_root, transform=True, ptag=ptag),
batch_size=1, shuffle=True, num_workers=4, pin_memory=True)
# 损失函数
criterion = nn.BCEWithLogitsLoss()
# 优化器
optimizer_deconv = torch.optim.Adam(deconv.parameters(), lr=1e-3)
optimizer_feature = torch.optim.Adam(feature.parameters(), lr=1e-4)
istep = 0
# 验证
def validation(val_loader, output_root, feature, deconv):
if not os.path.exists(output_root):
os.mkdir(output_root)
for ib, (data, prior, img_name, img_size) in enumerate(val_loader):
# print ib
prior = prior.unsqueeze(1)
data = torch.cat((data, prior), 1)
# inputs = Variable(data).cuda()
inputs = Variable(data)
feats = feature(inputs)
feats = feats[-3:]
feats = feats[::-1]
msk = deconv(feats)
msk = functional.upsample(msk, scale_factor=4)
msk = functional.sigmoid(msk)
mask = msk.data[0, 0].cpu().numpy()
mask = cv2.resize(mask, dsize=(img_size[0][0], img_size[1][0]))
plt.imsave(os.path.join(output_root, img_name[0]+'.png'), mask, cmap='gray')
for it in range(iter_num):
for ib, (data, prior, lbl) in enumerate(train_loader):
# prior = Variable(prior.unsqueeze(1)).cuda()
prior = Variable(prior.unsqueeze(1))
# inputs = Variable(data).cuda()
inputs = Variable(data)
# lbl = Variable(lbl.unsqueeze(1)).cuda()
lbl = Variable(lbl.unsqueeze(1))
loss = 0
for ir in range(r_num): # 循环
inputs4c = torch.cat((inputs, prior), 1)
feats = feature(inputs4c)
feats = feats[-3:]
feats = feats[::-1]
msk = deconv(feats)
msk = functional.upsample(msk, scale_factor=4)
prior = functional.sigmoid(msk)
loss += criterion(msk, lbl)
deconv.zero_grad()
feature.zero_grad()
loss.backward()
optimizer_feature.step()
optimizer_deconv.step()
# visulize
image = make_image_grid(inputs.data[:, :3], mean, std)
writer.add_image('Image', torchvision.utils.make_grid(image), ib)
msk = functional.sigmoid(msk)
mask1 = msk.data
mask1 = mask1.repeat(1, 3, 1, 1)
writer.add_image('Image2', torchvision.utils.make_grid(mask1), ib)
acc = math.e ** (0 - loss)
print('loss: %.4f, acc %.4f, (epoch: %d, step: %d)' % (loss.data[0], acc, it, ib))
writer.add_scalar('loss', loss.data[0], istep)
writer.add_scalar('acc', acc.data[0], istep)
istep += 1
del inputs, msk, lbl, loss, feats, mask1, image, acc
gc.collect()
if ib % 24 == 0:
filename = ('%s/deconv-epoch-%d-step-%d.pth' % (check_root, it, ib))
torch.save(deconv.state_dict(), filename)
filename = ('%s/feature-epoch-%d-step-%d.pth' % (check_root, it, ib))
torch.save(feature.state_dict(), filename)
print('save: (epoch: %d, step: %d)' % (it, ib))