-
Notifications
You must be signed in to change notification settings - Fork 1
/
train.py
180 lines (146 loc) · 6.9 KB
/
train.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
import os
import GPUtil
# Get a list of available GPUs
gpus = GPUtil.getGPUs()
# Select the GPU with the lowest utilization
chosen_gpu = None
for gpu in gpus:
if not chosen_gpu:
chosen_gpu = gpu
elif gpu.memoryUtil < chosen_gpu.memoryUtil:
chosen_gpu = gpu
# Set CUDA device to the selected GPU
os.environ["CUDA_VISIBLE_DEVICES"] = str(chosen_gpu.id)
import numpy as np
import os
import torch
import torch.utils.data as data
from torch import optim, nn
from tensorboardX import SummaryWriter
from data_readers.train_data_loaders import TrainSeqData
from utils.evaluate import PerceptualLoss
from pytorch_msssim import SSIM
from utils.configs import set_configs
import argparse
from model_v2e2v import V2E2VNet
import matplotlib.pyplot as plt
class Train:
def __init__(self, cfgs, device):
self.device = device
self.model_name = '{}_C{}_{}_{}_fc{}_{}_{}'\
.format(cfgs.model_name, cfgs.C, cfgs.pl, cfgs.ps, cfgs.cutoff_hz, cfgs.ql, cfgs.qs)
self.path_to_model = os.path.join(cfgs.path_to_model, self.model_name)
if not os.path.exists(self.path_to_model):
os.makedirs(self.path_to_model)
self.model = V2E2VNet(cfgs=cfgs, image_dim=cfgs.image_dim, device=device).to(device)
print(self.model)
self.v2e_params = {'C': cfgs.C,
'ps': cfgs.ps,
'pl': cfgs.pl,
'cutoff_hz': cfgs.cutoff_hz,
'qs': cfgs.qs,
'ql': cfgs.ql,
'refractory_period_s': cfgs.refractory_period_s
}
if cfgs.load_epoch_for_train:
## load trained E2V model using the new V2E generator
checkpoint = torch.load(os.path.join(self.path_to_model, '{}_{}.pth.tar'
.format(self.model_name, cfgs.load_epoch_for_train)))
self.model.load_state_dict(checkpoint['state_dict'], strict=True)
else:
## load pretrained E2V model using normal events
checkpoint = torch.load(cfgs.path_to_e2v, map_location='cuda:0')
self.model.e2v_net.load_state_dict(checkpoint['state_dict'], strict=True)
# Training data
path_to_train_data = cfgs.path_to_train_data
train_data = TrainSeqData(os.path.join(path_to_train_data, 'train_v2e2v.txt'), cfgs.path_to_train_data, cfgs.len_sequence, cfgs.num_pack_frames)
self.train_loader = data.DataLoader(train_data,batch_size=cfgs.batch_size,shuffle=cfgs.shuffle,num_workers=4)
# Training details
lr = cfgs.lr*(0.9**np.floor(cfgs.load_epoch_for_train/10.))
self.optimizer = optim.Adam(self.model.parameters(),lr=lr) #Adam
self.scheduler = optim.lr_scheduler.StepLR(self.optimizer, step_size=10, gamma=0.9)
# Loss
self.lpips_loss_fn = PerceptualLoss(net='vgg', device=device)
self.L1_loss_fn = nn.L1Loss().to(device)
self.ssim_loss_fn = SSIM(data_range=1, size_average=True, channel=1, nonnegative_ssim=False).to(device)
# Save training results
self.is_SummaryWriter = cfgs.is_SummaryWriter
if self.is_SummaryWriter:
self.writer = SummaryWriter('./summary/{}'\
.format(self.model_name))
def run_train(self, cfgs):
for epoch in range(cfgs.load_epoch_for_train, cfgs.epochs):
lr = self.scheduler.get_last_lr()[0]
print('lr:', lr)
self.train_recurrent(epoch, cfgs)
self.scheduler.step()
torch.save({'epoch': epoch+1,
'state_dict': self.model.state_dict(),
'v2e_params': self.v2e_params},
os.path.join(self.path_to_model, '{}_{}.pth.tar'\
.format(self.model_name, epoch+1)))
def train_recurrent(self, epoch, cfgs):
torch.cuda.empty_cache()
self.model.train()
batch_num =len(self.train_loader)
loss = 0
prev_img = None
state = None
for batch_idx, train_data in enumerate(self.train_loader):
seq_timestamps, seq_images, seq_gt_images = train_data
loss = 0
state = None
for s in range(len(seq_timestamps)):
timestamps = seq_timestamps[s].to(self.device)
images = seq_images[s].to(self.device)
gt_images = seq_gt_images[s].to(self.device)
if s == 0:
loss = 0
prev_img = torch.zeros_like(gt_images)
state = None
output, state= \
self.model(images, timestamps, prev_img, state, batch_idx)
output = torch.clamp(output, min=1e-7, max=1-1e-7)
prev_img = output.clone()
loss_lpips = self.lpips_loss_fn(output, gt_images, normalize=True)
loss_l1 = self.L1_loss_fn(output, gt_images)
loss_ssim = 1 - self.ssim_loss_fn(output, gt_images)
loss = loss_lpips + loss_l1 + loss_ssim
#loss
if self.is_SummaryWriter:
self.writer.add_scalar('LPIPS', loss_lpips, batch_num*epoch+batch_idx)
self.writer.add_scalar('MSE', loss_l1, batch_num*epoch+batch_idx)
self.writer.add_scalar('SSIM', loss_ssim, batch_num*epoch+batch_idx)
self.writer.add_scalar('loss', loss, batch_num*epoch+batch_idx)
if cfgs.display_train:
# for name, parms in self.model.named_parameters():
# print('-->name:', name, '-->grad_requirs:',parms.requires_grad)#, \
#' -->grad_value:',parms.grad, '-->value', parms.data)
plt.subplot(1,2,1)
plt.imshow(gt_images.cpu().data[0,0,...], cmap='gray')
plt.axis('off')
plt.title('GT')
plt.subplot(1,2,2)
plt.imshow(output.cpu().data[0,0,...], cmap='gray')
plt.axis('off')
plt.title('I_rec')
plt.show()
self.optimizer.zero_grad()
loss.backward(retain_graph=False)
self.optimizer.step()
if batch_idx%50 == 0:
print('Train Epoch: {} [{}/{} ({:.0f}%)]\tloss: {:.6f}'.format(\
epoch+1, batch_idx*self.train_loader.batch_size, len(self.train_loader.dataset),\
100.*batch_idx/len(self.train_loader), loss.data)) # .data.cpu().numpy()
if __name__ == '__main__':
device = torch.device("cuda:0" if torch.cuda.is_available() else "cpu")
print('device: ', device)
## Parameter configuration
parser = argparse.ArgumentParser(
description='Training options')
set_configs(parser)
cfgs = parser.parse_args()
cfgs.shuffle = True
# Training
model_train = Train(cfgs, device)
model_train.run_train(cfgs)