-
Notifications
You must be signed in to change notification settings - Fork 1
/
train_PDG_DA.py
228 lines (218 loc) · 11.1 KB
/
train_PDG_DA.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
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
from __future__ import print_function
import argparse
import torch
import torch.nn as nn
import torch.optim as optim
import sklearn.metrics as metrics
import tqdm
from torch.optim.lr_scheduler import CosineAnnealingLR
from data.dataloader import *
from model.PDG import best
from model.pointnet import PointNet
from model.dgcnn import DGCNN
import numpy as np
from torch.utils.data import DataLoader
import time
import random
from utils.loss import SupConLoss
from torch.utils.tensorboard import SummaryWriter
class IOStream():
def __init__(self, path):
self.f = open(path, 'a')
def cprint(self, text):
print(text)
self.f.write(text+'\n')
self.f.flush()
def close(self):
self.f.close()
def _init_():
if not os.path.exists('checkpoint'):
os.makedirs('checkpoint')
if not os.path.exists('checkpoint/' + args.exp_name):
os.makedirs('checkpoint/' + args.exp_name)
if not os.path.exists('checkpoint/' + args.exp_name + '/' + 'models'):
os.makedirs('checkpoint/' + args.exp_name + '/' + 'models')
# os.system('cp main.py checkpoint' + '/' + args.exp_name + '/' + 'main.py.backup')
# os.system('cp model.py checkpoint' + '/' + args.exp_name + '/' + 'model.py.backup')
# os.system('cp util.py checkpoint' + '/' + args.exp_name + '/' + 'util.py.backup')
# os.system('cp data.py checkpoint' + '/' + args.exp_name + '/' + 'data.py.backup')
def entropy_prob(x):
entropy = -(torch.softmax(x,-1)*torch.log_softmax(x,-1)).sum(-1)
return entropy.mean()
def train(args, io):
print('this')
writer = SummaryWriter('./checkpoint/' + args.exp_name)
source_train_loader = DataLoader(myBatchPaddingData_DA(pc_input_num=args.num_points, status='train', aug=True, swapax=swapax_src,
pc_root='data/data_DA/'+args.src_dataset,sample_num=1,batch_size=1), num_workers=12,
batch_size=32, shuffle=True, drop_last=True, pin_memory=True)
target_test_loader = DataLoader(PaddingData(pc_input_num=args.num_points, status='test', aug=False, swapax=swapax_trg,
pc_root='data/data_DA/'+args.trg_dataset), num_workers=4,
batch_size=32, shuffle=True, drop_last=False, pin_memory=True)
target_test_loader2 = DataLoader(PaddingData(pc_input_num=args.num_points, status='test', aug=False, swapax=swapax_trg2,
pc_root='data/data_DA/' + args.trg_dataset2), num_workers=4,
batch_size=32, shuffle=True, drop_last=False, pin_memory=True)
device = torch.device("cuda" if args.cuda else "cpu")
# Try to load models
# if args.model == 'pointnet':
# model = PointNet(args,output_channels=num_class).to(device)
# elif args.model == 'dgcnn':
# model = DGCNN(args,output_channels=num_class).to(device)
model = best(args, output_channels=num_class).to(device)
opt = optim.Adam(model.parameters(), lr=args.lr, weight_decay=1e-4)
scheduler = CosineAnnealingLR(opt, args.epochs,1e-5)
CSL = SupConLoss(temperature=0.07, contrast_mode='all', base_temperature=0.07)
best_test_acc = 0
best_test_acc_bg = 0
last_five_epoch = []
last_five_epoch_bg = []
for i in range(args.epochs):
train_loss = 0.0
count = 0.0
train_pred = []
train_true = []
model.train()
for data_s in source_train_loader:
idx,data, pn, label, data_ori, pn_ori = data_s
data,pn,data_ori,pn_ori, label = data.to(device).transpose(1, 2),pn.to(device),data_ori.to(
device).transpose(1, 2), pn_ori.to(device),label.to(device).squeeze()
batch_size = data.size()[0]
label_score_L = label.unsqueeze(1).repeat(1, 8).reshape(-1)
opt.zero_grad()
out1,out2,score1, score2, p1,f1 = model(data_ori, pn_ori, data, pn,test=False)
loss_entropy2 = nn.CrossEntropyLoss()(score1,label_score_L)+nn.CrossEntropyLoss()(score2,label_score_L)
loss_entropy = nn.CrossEntropyLoss()(out1,label)+nn.CrossEntropyLoss()(out2,label)
loss_CSL = CSL(p1)
loss_CSL_all = CSL(f1,label)
loss = loss_entropy+0.05*loss_entropy2+0.01*loss_CSL+0.01*loss_CSL_all
loss.backward()
opt.step()
preds = out1.max(dim=1)[1]
count += batch_size
train_loss += loss.item() * batch_size
train_true.append(label.cpu().numpy())
train_pred.append(preds.detach().cpu().numpy())
train_true = np.concatenate(train_true)
train_pred = np.concatenate(train_pred)
train_acc = metrics.accuracy_score(train_true, train_pred)
avg_per_class_acc = metrics.balanced_accuracy_score(train_true, train_pred)
train_loss = train_loss * 1.0 / count
outstr ='Source_train epoch %d, loss: %.6f, test acc: %.6f, test avg acc: %.6f' % (i,train_loss, train_acc, avg_per_class_acc)
io.cprint(outstr)
scheduler.step()
if i > -1 :
loss_t, acc_t, per_class_acc_t = val(model, target_test_loader, device)
loss_t2, acc_t2, per_class_acc_t2 = val(model, target_test_loader2, device)
outstr2 = 'Target_test epoch %d, loss: %.6f, test acc: %.6f, test avg acc: %.6f' % (i,loss_t, acc_t, per_class_acc_t)
io.cprint(outstr2)
outstr3 = 'Target_test_bg epoch %d, loss: %.6f, test acc: %.6f, test avg acc: %.6f' % (
i, loss_t2, acc_t2, per_class_acc_t2)
io.cprint(outstr3)
# writer.add_scalar('train_acc:',acc,i)
# writer.add_scalar('test_acc:', acc_bg, i)
if acc_t >= best_test_acc:
best_test_acc = acc_t
if acc_t2 >= best_test_acc_bg:
best_test_acc_bg = acc_t2
if i >= 150:
torch.save(model.state_dict(), 'checkpoint/%s/models/ori_model_%d.t7' % (args.exp_name, i))
io.cprint('best_target_test_acc::::::::::::::::::::::%5f' % (best_test_acc))
io.cprint('best_target_bg_test_acc::::::::::::::::::::::%5f' % (best_test_acc_bg))
last_five_epoch.append(acc_t)
last_five_epoch_bg.append(acc_t2)
io.cprint('final 5 epoch acc: %5f' % (np.mean(np.array(last_five_epoch)[-5:-1])))
io.cprint('final 5 epoch acc bg: %5f' % (np.mean(np.array(last_five_epoch_bg)[-5:-1])))
writer.close()
def val(model,data_loader,device):
test_loss = 0.0
count = 0.0
model.eval()
test_pred = []
test_true = []
iter_test = iter(data_loader)
for i in tqdm.trange(len(data_loader)):
idx,data, pn, label = iter_test.next()
data, label = data.to(device), label.to(device).squeeze()
data = data.squeeze()
batch_size = data.size()[0]
with torch.no_grad():
logits,_ = model(data)
loss = nn.CrossEntropyLoss()(logits, label)
preds = logits.max(dim=1)[1]
count += batch_size
test_loss += loss.item() * batch_size
test_true.append(label.cpu().numpy())
test_pred.append(preds.detach().cpu().numpy())
test_true = np.concatenate(test_true)
test_pred = np.concatenate(test_pred)
test_acc = metrics.accuracy_score(test_true, test_pred)
avg_per_class_acc = metrics.balanced_accuracy_score(test_true, test_pred)
test_loss = test_loss * 1.0 / count
return test_loss, test_acc, avg_per_class_acc
if __name__ == "__main__":
# Training settings
parser = argparse.ArgumentParser(description='Point Cloud Recognition')
parser.add_argument('--exp_name', type=str, default='test', metavar='N',
help='Name of the experiment')
parser.add_argument('--model', type=str, default='dgcnn', metavar='N',
choices=['pointnet', 'dgcnn'],
help='Model to use, [pointnet, dgcnn]')
parser.add_argument('--src_dataset', type=str, default='modelnet', metavar='N',
choices=['modelnet,shapenet,scannet'])
parser.add_argument('--trg_dataset', type=str, default='shapenet', metavar='N',
choices=['modelnet,shapenet,scannet'])
parser.add_argument('--trg_dataset2', type=str, default='scannet', metavar='N',
choices=['modelnet,shapenet,scannet'])
parser.add_argument('--batch_size', type=int, default=32, metavar='batch_size',
help='Size of batch)')
parser.add_argument('--test_batch_size', type=int, default=32, metavar='batch_size',
help='Size of batch)')
parser.add_argument('--epochs', type=int, default=160, metavar='N',
help='number of episode to train ')
parser.add_argument('--use_sgd', type=bool, default=False,
help='Use SGD')
parser.add_argument('--lr', type=float, default=0.001, metavar='LR',
help='learning rate (default: 0.001, 0.1 if using sgd)')
parser.add_argument('--momentum', type=float, default=0.9, metavar='M',
help='SGD momentum (default: 0.9)')
parser.add_argument('--no_cuda', type=bool, default=False,
help='enables CUDA training')
parser.add_argument('--seed', type=int, default=1, metavar='S',
help='random seed (default: 1)')
parser.add_argument('--eval', type=bool, default=False,
help='evaluate the model')
parser.add_argument('--num_points', type=int, default=1024,
help='num of points to use')
parser.add_argument('--dropout', type=float, default=0.5,
help='dropout rate')
parser.add_argument('--mixup_params', type=float, default=1.0, help='a,b in beta distribution')
parser.add_argument('--emb_dims', type=int, default=1024, metavar='N',
help='Dimension of embeddings')
parser.add_argument('--k', type=int, default=20, metavar='N',
help='Num of nearest neighbors to use')
parser.add_argument('--model_path', type=str, default='./checkpoint/pointnet_6part/models/best_ori_model.t7',
metavar='N',
help='Pretrained model path')
args = parser.parse_args()
_init_()
def seed_torch(seed=0):
random.seed(seed)
os.environ['PYTHONHASHSEED'] = str(seed)
np.random.seed(seed)
torch.manual_seed(seed)
torch.cuda.manual_seed(seed)
# torch.cuda.manual_seed_all(seed) # if you are using multi-GPU.
# torch.backends.cudnn.benchmark = False
# torch.backends.cudnn.deterministic = True
# torch.backends.cudnn.enabled = True
io = IOStream('checkpoint/' + args.exp_name + '/run.log')
io.cprint(str(args))
num_class = 10
swapax_src = args.src_dataset == 'modelnet'
swapax_trg = args.trg_dataset == 'modelnet'
swapax_trg2 = args.trg_dataset2 == 'modelnet'
args.cuda = not args.no_cuda and torch.cuda.is_available()
seed_torch(1)
start_time = time.time()
train(args, io)
end_time = time.time()
print('training time%.6f' % ((end_time - start_time) / 60))