-
Notifications
You must be signed in to change notification settings - Fork 4
/
Copy pathtrain_gated_aff.py
197 lines (158 loc) · 8.35 KB
/
train_gated_aff.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
import os
import numpy as np
import torch
from torch.backends import cudnn
from torch.utils.data import DataLoader
cudnn.enabled = True
from torchvision import transforms
import voc12.data
import tool.pyutils as pyutils
import tool.torchutils as torchutils
import tool.imutils as imutils
import argparse
import importlib
import torch.nn.functional as F
from model_fts_gated_regularized import ModelLossSemsegGatedCRF
os.environ["CUDA_VISIBLE_DEVICES"] = '0,1'
def _unfold(img, radius):
assert img.dim() == 4, 'Unfolding requires NCHW batch'
N, C, H, W = img.shape
diameter = 2 * radius + 1
return F.unfold(img, diameter, 1, radius).view(N, C, diameter, diameter, H, W)
if __name__ == '__main__':
parser = argparse.ArgumentParser()
parser.add_argument("--batch_size", default=4, type=int)
parser.add_argument("--max_epoches", default=12, type=int)
parser.add_argument("--network", default="network.resnet38_aff_gated", type=str)
parser.add_argument("--lr", default=0.01, type=float)
parser.add_argument('--radius', type=int, default=4)
parser.add_argument("--num_workers", default=8, type=int)
parser.add_argument("--wt_dec", default=5e-4, type=float)
parser.add_argument("--train_list", default="voc12/train_aug.txt", type=str)
parser.add_argument("--session_name", default="res_aff_box", type=str)
parser.add_argument("--crop_size", default=448, type=int)
# parser.add_argument("--weights", default='./netWeights/ilsvrc-cls_rna-a1_cls1000_ep-0001.params', type=str)
parser.add_argument("--weights", default='./netWeights/resnet38_aff_SEAM.pth', type=str)
parser.add_argument("--voc12_root", default='/data/zbf_data/dataset/VOCdevkit/VOC2012', type=str)
parser.add_argument("--label_dir",
default='./data/Init_Label/SEAM_Box',
type=str)
args = parser.parse_args()
pyutils.Logger(args.session_name + '.log')
print(vars(args))
model = getattr(importlib.import_module(args.network), 'Net')()
print(model)
train_dataset = voc12.data.VOC12AffGtDataset_NoExtract(args.train_list, label_dir=args.label_dir,
voc12_root=args.voc12_root, cropsize=args.crop_size, radius=5,
joint_transform_list=[
None,
None,
imutils.RandomCrop(args.crop_size),
imutils.RandomHorizontalFlip()
],
img_transform_list=[
transforms.ColorJitter(brightness=0.3, contrast=0.3,
saturation=0.3, hue=0.1),
np.asarray,
model.normalize,
imutils.HWC_to_CHW
],
label_transform_list=[
None,
None,
None,
imutils.LabelResize(args.crop_size//8)
])
train_data_loader = DataLoader(train_dataset, batch_size=args.batch_size, shuffle=True, num_workers=args.num_workers,
pin_memory=True, drop_last=True)
max_step = len(train_dataset) // args.batch_size * args.max_epoches
param_groups = model.get_parameter_groups()
optimizer = torchutils.PolyOptimizer([
{'params': param_groups[0], 'lr': args.lr, 'weight_decay': args.wt_dec},
{'params': param_groups[1], 'lr': 2*args.lr, 'weight_decay': 0},
{'params': param_groups[2], 'lr': 10*args.lr, 'weight_decay': args.wt_dec},
{'params': param_groups[3], 'lr': 20*args.lr, 'weight_decay': 0}
], lr=args.lr, weight_decay=args.wt_dec, max_step=max_step)
if args.weights[-7:] == '.params':
import network.resnet38d
assert args.network == "network.resnet38_aff_gated"
weights_dict = network.resnet38d.convert_mxnet_to_torch(args.weights)
else:
weights_dict = torch.load(args.weights)
model.load_state_dict(weights_dict, strict=False)
model = torch.nn.DataParallel(model).cuda()
model.train()
avg_meter = pyutils.AverageMeter('loss', 'bg_loss', 'fg_loss', 'neg_loss', 'crf_loss', 'bg_cnt', 'fg_cnt',
'neg_cnt')
gatedcrf = ModelLossSemsegGatedCRF()
timer = pyutils.Timer("Session started: ")
radius = args.radius
for ep in range(args.max_epoches):
for iter, pack in enumerate(train_data_loader):
features, aff, aff_crf = model.forward(pack[0], radius)
aff = aff.unsqueeze(dim=1)
label = pack[1].cuda().unsqueeze(dim=1)
label = _unfold(label,radius=radius)
N, _, H, W = features.shape
crf_img = F.interpolate(pack[0], [H, W], mode='bilinear', align_corners=False)
crf_input = {'rgb': crf_img.cuda()}
label_center = label[:, :, radius, radius, :, :].view(N, 1, 1, 1, H, W)
label_center = label_center.expand_as(label)
aff_label = torch.eq(label, label_center)
aff_label_ignore = torch.eq(label,255)
aff_label[aff_label_ignore == 1] = 255
aff_label[label_center == 255] = 255
bg_pos_label_need = torch.zeros_like(aff_label)
bg_pos_label_need[aff_label == 1] = 1
bg_pos_label_need[label_center != 0] = 0
bg_pos_label_need[aff_label_ignore == 1] = 0
fg_pos_label_need = torch.zeros_like(aff_label)
fg_pos_label_need[aff_label == 1] = 1
fg_pos_label_need[label_center == 0] = 0
fg_pos_label_need[aff_label_ignore ==1] = 0
fg_pos_label_need[label_center == 255] = 0
neg_label_need = torch.zeros_like(aff_label)
neg_label_need[aff_label==0] =1
neg_label_need[aff_label_ignore==1] = 0
neg_label_need[label_center==255] = 0
bg_count = torch.sum(bg_pos_label_need).float() + 1e-5
fg_count = torch.sum(fg_pos_label_need).float() + 1e-5
neg_count = torch.sum(neg_label_need).float() + 1e-5
bg_loss = torch.sum(- bg_pos_label_need.float() * torch.log(aff + 1e-5)) / bg_count
fg_loss = torch.sum(- fg_pos_label_need.float() * torch.log(aff + 1e-5)) / fg_count
neg_loss = torch.sum(- neg_label_need.float() * torch.log(1. + 1e-5 - aff)) / neg_count
croppings = pack[2].cuda().unsqueeze(dim=1)
croppings = _unfold(croppings, radius=radius)
croppings_center = croppings[:, :, radius, radius, :, :].view(N, 1, 1, 1, H, W)
croppings_center = croppings_center.expand_as(croppings)
croppings_ignore = torch.eq(croppings, 0)
croppings = torch.ones_like(croppings)
croppings[croppings_ignore == 1] = 0
croppings[croppings_center == 0] = 0
out_gatedcrf = gatedcrf(aff_crf, [{'weight': 1, 'xy': 6, 'rgb': 0.1}], radius, crf_input, H, W,
mask_src=croppings.float())
crf_loss = 3*out_gatedcrf['loss']
loss = bg_loss / 4 + fg_loss / 4 + neg_loss / 2 + crf_loss
optimizer.zero_grad()
loss.backward()
optimizer.step()
avg_meter.add({
'loss': loss.item(),
'bg_loss': bg_loss.item(), 'fg_loss': fg_loss.item(), 'neg_loss': neg_loss.item(),
'crf_loss': crf_loss.item(),
'bg_cnt': bg_count.item(), 'fg_cnt': fg_count.item(), 'neg_cnt': neg_count.item()
})
if (optimizer.global_step - 1) % 5 == 0:
timer.update_progress(optimizer.global_step / max_step)
print('Iter:%5d/%5d' % (optimizer.global_step - 1, max_step),
'loss:%.4f %.4f %.4f %.4f %.4f' % avg_meter.get('loss', 'crf_loss', 'bg_loss', 'fg_loss',
'neg_loss'),
'cnt:%.0f %.0f %.0f' % avg_meter.get('bg_cnt', 'fg_cnt', 'neg_cnt'),
'imps:%.1f' % ((iter + 1) * args.batch_size / timer.get_stage_elapsed()),
'Fin:%s' % (timer.str_est_finish()),
'lr: %.4f' % (optimizer.param_groups[0]['lr']), flush=True)
avg_meter.pop()
else:
print('')
timer.reset_stage()
torch.save(model.module.state_dict(), args.session_name + '.pth')