-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathosvos_scribble.py
186 lines (153 loc) · 8.45 KB
/
osvos_scribble.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
import os
import copy
import timeit
import torch
import torch.optim as optim
from torchvision import transforms
from torch.utils.data import DataLoader
from torch.autograd import Variable
import scipy.misc as sm
import numpy as np
from davisinteractive import utils as interactive_utils
from dataloaders import davis_2017 as db
from mypath import Path
import networks.vgg_osvos as vo
from dataloaders import custom_transforms as tr
from layers.osvos_layers import class_balanced_cross_entropy_loss
class OSVOSScribble(object):
def __init__(self, parent_model, save_model_dir, gpu_id, time_budget, save_result_dir=None):
self.save_model_dir = save_model_dir
self.parent_model = parent_model
self.save_res_dir = save_result_dir
self.net = vo.OSVOS(pretrained=0)
if gpu_id >= 0:
torch.cuda.set_device(device=gpu_id)
self.net.cuda()
self.gpu_id = gpu_id
self.time_budget = time_budget
self.meanval = (104.00699, 116.66877, 122.67892)
self.train_batch = 1
self.test_batch = 1
self.prev_models = {}
self.parent_model_state = torch.load(os.path.join(Path.models_dir(), self.parent_model),
map_location=lambda storage, loc: storage)
def train(self, first_frame, n_interaction, obj_id, scribbles_data, scribble_iter, subset, use_previous_mask=False):
nAveGrad = 1
num_workers = 4
train_batch = min(n_interaction, self.train_batch)
frames_list = interactive_utils.scribbles.annotated_frames_object(scribbles_data, obj_id)
scribbles_list = scribbles_data['scribbles']
seq_name = scribbles_data['sequence']
if obj_id == 1 and n_interaction == 1:
self.prev_models = {}
# Network definition
if n_interaction == 1:
print('Loading weights from: {}'.format(self.parent_model))
self.net.load_state_dict(self.parent_model_state)
self.prev_models[obj_id] = None
else:
print('Loading weights from previous network: objId-{}_interaction-{}_scribble-{}.pth'
.format(obj_id, n_interaction-1, scribble_iter))
self.net.load_state_dict(self.prev_models[obj_id])
lr = 1e-8
wd = 0.0002
optimizer = optim.SGD([
{'params': [pr[1] for pr in self.net.stages.named_parameters() if 'weight' in pr[0]], 'weight_decay': wd},
{'params': [pr[1] for pr in self.net.stages.named_parameters() if 'bias' in pr[0]], 'lr': lr * 2},
{'params': [pr[1] for pr in self.net.side_prep.named_parameters() if 'weight' in pr[0]], 'weight_decay': wd},
{'params': [pr[1] for pr in self.net.side_prep.named_parameters() if 'bias' in pr[0]], 'lr': lr * 2},
{'params': [pr[1] for pr in self.net.upscale.named_parameters() if 'weight' in pr[0]], 'lr': 0},
{'params': [pr[1] for pr in self.net.upscale_.named_parameters() if 'weight' in pr[0]], 'lr': 0},
{'params': self.net.fuse.weight, 'lr': lr / 100, 'weight_decay': wd},
{'params': self.net.fuse.bias, 'lr': 2 * lr / 100},
], lr=lr, momentum=0.9)
prev_mask_path = os.path.join(self.save_res_dir, 'interaction-{}'.format(n_interaction-1),
'scribble-{}'.format(scribble_iter))
composed_transforms_tr = transforms.Compose([tr.SubtractMeanImage(self.meanval),
tr.CustomScribbleInteractive(scribbles_list, first_frame,
use_previous_mask=use_previous_mask,
previous_mask_path=prev_mask_path),
tr.RandomHorizontalFlip(),
tr.ScaleNRotate(rots=(-30, 30), scales=(.75, 1.25)),
tr.ToTensor()])
# Training dataset and its iterator
db_train = db.DAVIS2017(split=subset, transform=composed_transforms_tr,
custom_frames=frames_list, seq_name=seq_name,
obj_id=obj_id, no_gt=True, retname=True)
trainloader = DataLoader(db_train, batch_size=train_batch, shuffle=True, num_workers=num_workers)
num_img_tr = len(trainloader)
loss_tr = []
aveGrad = 0
start_time = timeit.default_timer()
# Main Training and Testing Loop
epoch = 0
while 1:
# One training epoch
running_loss_tr = 0
for ii, sample_batched in enumerate(trainloader):
inputs, gts, void = sample_batched['image'], sample_batched['scribble_gt'], sample_batched[
'scribble_void_pixels']
# Forward-Backward of the mini-batch
inputs, gts, void = Variable(inputs), Variable(gts), Variable(void)
if self.gpu_id >= 0:
inputs, gts, void = inputs.cuda(), gts.cuda(), void.cuda()
outputs = self.net.forward(inputs)
# Compute the fuse loss
loss = class_balanced_cross_entropy_loss(outputs[-1], gts, size_average=False, void_pixels=void)
running_loss_tr += loss.item()
# Print stuff
if epoch % 10 == 0:
running_loss_tr /= num_img_tr
loss_tr.append(running_loss_tr)
print('[Epoch: %d, numImages: %5d]' % (epoch + 1, ii + 1))
print('Loss: %f' % running_loss_tr)
# writer.add_scalar('data/total_loss_epoch', running_loss_tr, epoch)
# Backward the averaged gradient
loss /= nAveGrad
loss.backward()
aveGrad += 1
# Update the weights once in nAveGrad forward passes
if aveGrad % nAveGrad == 0:
# writer.add_scalar('data/total_loss_iter', loss.data[0], ii + num_img_tr * epoch)
optimizer.step()
optimizer.zero_grad()
aveGrad = 0
epoch += train_batch
stop_time = timeit.default_timer()
if stop_time - start_time > self.time_budget:
break
# Save the model into dictionary
self.prev_models[obj_id] = copy.deepcopy(self.net.state_dict())
def test(self, sequence, n_interaction, obj_id, subset, scribble_iter=0):
if self.save_res_dir:
save_dir_res = os.path.join(self.save_res_dir, 'interaction-{}'.format(n_interaction),
'scribble-{}'.format(scribble_iter),
sequence, str(obj_id))
if not os.path.exists(save_dir_res):
os.makedirs(save_dir_res)
composed_transforms_ts = transforms.Compose([tr.SubtractMeanImage(self.meanval),
tr.ToTensor()])
# Testing dataset and its iterator
db_test = db.DAVIS2017(split=subset, transform=composed_transforms_ts, seq_name=sequence, no_gt=True, retname=True)
testloader = DataLoader(db_test, batch_size=self.test_batch, shuffle=False, num_workers=2)
print('Testing Network for obj_id={}'.format(obj_id))
print('Loading weights from objId-{}_interaction-{}_scribble-{}.pth'
.format(obj_id, n_interaction, scribble_iter))
# Main Testing Loop
masks = []
for ii, sample_batched in enumerate(testloader):
img, gt, meta = sample_batched['image'], sample_batched['gt'], sample_batched['meta']
# Forward of the mini-batch
inputs, gts = Variable(img, volatile=True), Variable(gt, volatile=True)
if self.gpu_id >= 0:
inputs, gts = inputs.cuda(), gts.cuda()
outputs = self.net.forward(inputs)[-1].cpu().data.numpy()
for jj in range(int(inputs.size()[0])):
pred = np.transpose(outputs[jj, :, :, :], (1, 2, 0))
pred = 1 / (1 + np.exp(-pred))
pred = np.squeeze(pred)
if self.save_res_dir:
# Save the result, attention to the index jj
sm.imsave(os.path.join(save_dir_res, os.path.basename(meta['frame_id'][jj]) + '.png'), pred)
masks.append(pred)
return masks