-
Notifications
You must be signed in to change notification settings - Fork 5
/
utils.py
207 lines (173 loc) · 6.44 KB
/
utils.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
import yaml
import os
import torch
from torch.backends import cudnn
import random
import importlib
from Models.BaseModel import BaseModel
import numpy as np
import data_process
import torchvision
import time
from sklearn.metrics import f1_score
from sklearn.metrics import confusion_matrix
def read_config(yaml_path):
with open(yaml_path,'r') as imf:
config = yaml.load(imf.read())
return config
def save_checkpoint(state,config):
expr_dir = os.path.join(config['checkpoint_dir'],config['experiment_name'])
if not os.path.exists(expr_dir):
os.makedirs(expr_dir)
epoch = state['epoch']
save_dir = os.path.join(expr_dir,str(epoch)+'.pth')
torch.save(state,save_dir)
def set_seed(seed, cuda=True):
random.seed(seed)
np.random.seed(seed)
torch.manual_seed(seed)
os.environ['PYTHONHASHSEED'] = str(seed)
if cuda:
torch.cuda.manual_seed(seed)
if seed == 0:
torch.backends.cudnn.deterministic = True
torch.backends.cudnn.benchmark = False
def init(config,local_rank,use_ddp):
# cudnn.benchmark = False # if benchmark=True, deterministic will be False
# cudnn.deterministic = True
# torch.manual_seed(config['seed']) # 为CPU设置随机种子
# torch.cuda.manual_seed(config['seed']) # 为当前GPU设置随机种子
# torch.cuda.manual_seed_all(config['seed']) # 为所有GPU设置随机种子
# random.seed(config['seed'])
config['local_rank'] = local_rank
config['use_ddp'] = use_ddp
set_seed(config['seed'])
os.environ['CUDA_VISIBLE_DEVICES'] = config['gpu_ids']
if config['use_ddp']:
torch.cuda.set_device(local_rank)
torch.distributed.init_process_group(backend='nccl')
print_options(config)
def print_options(config):
expr_dir = os.path.join(config['checkpoint_dir'], config['experiment_name'])
if not os.path.exists(expr_dir) and ((config['local_rank'] == 0 and config['use_ddp']) or not config['use_ddp']):
os.makedirs(expr_dir)
message = '--------------------Options----------------------\n'
for k in list(config.keys()):
val = config[k]
comment = str(k)+':\t'
if val == None:
comment += 'None\n'
else:
comment += str(val) + '\n'
message += comment
message += '--------------------End----------------------\n'
phase = 'train' if not config['eval'] else 'val'
file_name = '{}_{}_opt.txt'.format(phase,get_time())
path = os.path.join(config['checkpoint_dir'],config['experiment_name'],file_name)
with open(path,'w') as imf:
imf.write(message)
def get_time():
return str(time.strftime("%Y_%m_%d_%H_%M_%S",time.localtime()))
def create_model(config):
model_name = 'Models.' + config['model_name']
modellib = importlib.import_module(model_name)
model = None
target_model_name = config['model_name']
for name,cls in modellib.__dict__.items():
if name.lower() == target_model_name.lower() and issubclass(cls,BaseModel):
model = cls
if model == None:
print("In %s.py, there should be a subclass of BaseModel with class name that matches %s in lowercase." % (
model_name, target_model_name))
exit(0)
instance = model(config)
return instance
def create_dataset(config,phase=None):
'''
:param config:
:param phase: 'train,val,test' look at config
:return:
'''
dataset_name = 'data_process.' + config['dataset_name']
modellib = importlib.import_module(dataset_name)
dataset = None
target_dataset_name = config['dataset_name']
for name,cls in modellib.__dict__.items():
if name.lower() == target_dataset_name:
dataset = cls
if dataset == None:
print("In %s.py, there should has class name that matches %s in lowercase." % (
dataset_name, target_dataset_name))
exit(0)
instance = dataset(config,phase)
return instance
def img2label(colormap,config):
cm2lbl = np.zeros(config['img_size'] ** 3,dtype='int64')
for i, cm in enumerate(colormap):
cm2lbl[(cm[0] * 256 + cm[1]) * 256 + cm[2]] = i
return cm2lbl
def iou(pred, target,n_class):
ious = []
for cls in range(n_class):
pred_inds = pred == cls
target_inds = target == cls
intersection = pred_inds[target_inds].sum()
union = pred_inds.int().sum() + target_inds.int().sum() - intersection
if union == 0:
ious.append(0) # if there is no ground truth, do not include in evaluation
else:
ious.append(float(intersection) / max(float(union), 1))
# print("cls", cls, pred_inds.sum(), target_inds.sum(), intersection, float(intersection) / max(union, 1))
return ious
def cal_matrics(pred,label):
p = torch.flatten(pred)
l = torch.flatten(label)
p = p.detach().cpu().numpy().tolist()
l = l.detach().cpu().numpy().tolist()
tn1, fp1, fn1, tp1 = confusion_matrix(l, p, labels=[0, 1]).flatten()
f1 = (2 * tp1) / (2 * tp1 + fn1 + fp1)
iou = tp1 / (fn1+fp1+tp1)
return f1,iou
def cal_f1_score(pred,label):
'''
:param pred: [1,w,h]
:param label: [1,w,h]
:return: float f1 score
'''
p = torch.flatten(pred)
l = torch.flatten(label)
p = p.detach().cpu().numpy().tolist()
l = l.detach().cpu().numpy().tolist()
#tn1, fp1, fn1, tp1 = confusion_matrix(l, p, labels=[0, 1]).flatten()
#f1 = (2 * tp1) / (2 * tp1 + fn1 + fp1)
#return f1
f1 = f1_score(l,p,average=None)
if len(f1) >= 2:
return f1[1]
else:
return f1[0]
def save_img(img,img_name,config):
dir_path = os.path.join(config['save_img_dir'],config['experiment_name'])
file_path = os.path.join(dir_path, img_name)
preffix,suffix = os.path.split(file_path)
if not os.path.exists(preffix):
os.makedirs(preffix)
torchvision.utils.save_image(img,file_path)
def accuracy(output,target,topk=(1,)):
with torch.no_grad():
maxk = max(topk)
batch_size = target.size(0)
_,pred = output.topk(maxk,1,True,True)
pred = pred.t()
correct = pred.eq(target.view(1,-1).expand_as(pred))
res = []
for k in topk:
correct_k = correct[:k].contiguous().view(-1).float().sum(0,keepdim=True)
res.append(correct_k.mul_(100.0 / batch_size).item())
return res
# if __name__ == '__main__':
# yaml_path = './configs/init.yaml'
# config = read_config(yaml_path)
#
# model = create_dataset(config)
# debug = 0