-
Notifications
You must be signed in to change notification settings - Fork 26
/
train_jigsaw.py
executable file
·176 lines (146 loc) · 7.88 KB
/
train_jigsaw.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
# Copyright (c) 2020. Hanchen Wang, [email protected]
# ref: https://github.com/AnTao97/dgcnn.pytorch/blob/master/main_semseg.py
# ref: https://github.com/yanx27/Pointnet_Pointnet2_pytorch/blob/master/train_semseg.py
import os, sys, torch, argparse, importlib, shutil
sys.path.append('models')
sys.path.append('utils')
from torch.optim.lr_scheduler import CosineAnnealingLR, StepLR
from Torch_Utility import weights_init, bn_momentum_adjust
from ModelNetDataLoader import ModelNetJigsawDataLoader
from torch.utils.tensorboard import SummaryWriter
from torch.utils.data import DataLoader
from TrainLogger import TrainLogger
from tqdm import tqdm
def parse_args():
parser = argparse.ArgumentParser('3D Point Cloud Jigsaw Puzzles')
''' === Training === '''
parser.add_argument('--log_dir', type=str, help='log folder [default: ]')
parser.add_argument('--gpu', type=str, default='0', help='GPU [default: 0]')
parser.add_argument('--batch_size', type=int, default=32, help='batch size [default: 32]')
parser.add_argument('--epoch', default=200, type=int, help='training epochs [default: 200]')
parser.add_argument('--lr', default=0.0001, type=float, help='learning rate [default: 1e-4]')
parser.add_argument('--optimizer', type=str, default='Adam', help='optimiser [default: Adam]')
parser.add_argument('--momentum', type=float, default=0.9, help='SGD momentum [default: 0.9]')
parser.add_argument('--lr_decay', type=float, default=0.7, help='lr decay rate [default: 0.7]')
parser.add_argument('--bn_decay', action='store_true', help='BN Momentum Decay [default: False]')
parser.add_argument('--xavier_init', action='store_true', help='Xavier weight init [default: False]')
parser.add_argument('--scheduler', type=str, default='step', help='lr decay scheduler [default: step]')
parser.add_argument('--model', type=str, default='pointnet_jigsaw', help='model [default: pointnet_jigsaw]')
parser.add_argument('--step_size', type=int, default=20, help='decay steps for lr [default: every 20 epochs]')
''' === Model === '''
parser.add_argument('--k', type=int, default=20, help='num of nearest neighbors to use [default: 20]')
parser.add_argument('--emb_dims', type=int, default=1024, help='dimension of embeddings [default: 1024]')
parser.add_argument('--num_point', type=int, default=1024, help='number of points per object [default: 1024]')
return parser.parse_args()
def main(args):
os.environ["CUDA_DEVICE_ORDER"] = "PCI_BUS_ID"
os.environ["CUDA_VISIBLE_DEVICES"] = args.gpu
NUM_CLASSES = 3 ** 3
DATA_PATH = 'data/modelnet40_ply_hdf5_2048/jigsaw'
TRAIN_DATASET = ModelNetJigsawDataLoader(DATA_PATH, split='train', n_points=args.num_point, k=3)
TEST_DATASET = ModelNetJigsawDataLoader(DATA_PATH, split='test', n_points=args.num_point, k=3)
trainDataLoader = DataLoader(TRAIN_DATASET, batch_size=args.batch_size, shuffle=True, num_workers=4)
testDataLoader = DataLoader(TEST_DATASET, batch_size=args.batch_size, shuffle=False, num_workers=4)
MyLogger = TrainLogger(args, name=args.model.upper(), subfold='jigsaw')
''' === MODEL LOADING === '''
MODEL = importlib.import_module(args.model)
shutil.copy(os.path.abspath(__file__), MyLogger.log_dir)
shutil.copy('./models/%s.py' % args.model, MyLogger.log_dir)
writer = SummaryWriter(os.path.join(MyLogger.experiment_dir, 'runs'))
# allow multiple GPUs
device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
classifier = MODEL.get_model(args=args, num_class=NUM_CLASSES, num_channel=3).to(device)
criterion = MODEL.get_loss().to(device)
classifier = torch.nn.DataParallel(classifier)
print('=' * 33)
print('Using %d GPU,' % torch.cuda.device_count(), 'Indices: %s' % args.gpu)
print('=' * 33)
if args.xavier_init:
classifier = classifier.apply(weights_init)
MyLogger.logger.info("Using Xavier Weight Initialisation")
if args.optimizer == 'Adam':
optimizer = torch.optim.Adam(
classifier.parameters(),
lr=args.lr,
betas=(0.9, 0.999),
eps=1e-08,
weight_decay=1e-4)
MyLogger.logger.info("Using Adam Optimiser")
else:
optimizer = torch.optim.SGD(
classifier.parameters(),
lr=1000 * args.lr,
momentum=args.momentum)
MyLogger.logger.info("Using SGD Optimiser")
LEARNING_RATE_CLIP = 1e-5
MOMENTUM_ORIGINAL = 0.1
MOMENTUM_DECAY = 0.5
MOMENTUM_DECAY_STEP = args.step_size
if args.scheduler == 'cos':
scheduler = CosineAnnealingLR(optimizer, T_max=args.epoch, eta_min=1e-3)
else:
scheduler = StepLR(optimizer, step_size=args.step_size, gamma=0.7)
for epoch in range(MyLogger.epoch, args.epoch + 1):
''' === Training === '''
MyLogger.epoch_init(training=True)
for points, target in tqdm(trainDataLoader, total=len(trainDataLoader), smoothing=0.9):
points, target = points.transpose(2, 1).float().cuda(), target.view(-1, 1)[:, 0].long().cuda()
classifier.train()
optimizer.zero_grad()
if args.model == 'pointnet_jigsaw':
pred, trans_feat = classifier(points)
pred = pred.contiguous().view(-1, NUM_CLASSES)
loss = criterion(pred, target, trans_feat)
else:
pred = classifier(points)
pred = pred.contiguous().view(-1, NUM_CLASSES)
loss = criterion(pred, target)
loss.backward()
optimizer.step()
# pdb.set_trace()
MyLogger.step_update(pred.data.max(1)[1].cpu().numpy(),
target.long().cpu().numpy(),
loss.cpu().detach().numpy())
MyLogger.epoch_summary(writer=writer, training=True)
''' === Evaluation === '''
with torch.no_grad():
classifier.eval()
MyLogger.epoch_init(training=False)
for points, target in tqdm(testDataLoader, total=len(testDataLoader), smoothing=0.9):
points, target = points.transpose(2, 1).float().cuda(), target.view(-1, 1)[:, 0].long().cuda()
if args.model == 'pointnet_jigsaw':
pred, trans_feat = classifier(points)
pred = pred.contiguous().view(-1, NUM_CLASSES)
loss = criterion(pred, target, trans_feat)
else:
pred = classifier(points)
pred = pred.contiguous().view(-1, NUM_CLASSES)
loss = criterion(pred, target)
MyLogger.step_update(pred.data.max(1)[1].cpu().numpy(),
target.long().cpu().numpy(),
loss.cpu().detach().numpy())
MyLogger.epoch_summary(writer=writer, training=False)
if MyLogger.save_model:
state = {
'step': MyLogger.step,
'epoch': MyLogger.best_instance_epoch,
'instance_acc': MyLogger.best_instance_acc,
'model_state_dict': classifier.state_dict(),
'optimizer_state_dict': optimizer.state_dict(),
}
torch.save(state, MyLogger.savepath)
scheduler.step()
if args.scheduler == 'step':
for param_group in optimizer.param_groups:
if optimizer.param_groups[0]['lr'] < LEARNING_RATE_CLIP:
param_group['lr'] = LEARNING_RATE_CLIP
if args.bn_decay:
momentum = MOMENTUM_ORIGINAL * (MOMENTUM_DECAY ** (epoch // MOMENTUM_DECAY_STEP))
if momentum < 0.01:
momentum = 0.01
print('BN momentum updated to: %f' % momentum)
classifier = classifier.apply(lambda x: bn_momentum_adjust(x, momentum))
MyLogger.train_summary()
if __name__ == '__main__':
args = parse_args()
main(args)