def test(config):
log_dir = os.path.join(config.log_dir, config.name + '_stage_2')
val_path = os.path.join(config.data, "*/test")
val_dataset = MultiviewImgDataset(val_path,
scale_aug=False,
rot_aug=False,
num_views=config.num_views)
val_loader = torch.utils.data.DataLoader(
val_dataset,
batch_size=config.stage2_batch_size,
shuffle=False,
num_workers=0)
pretraining = not config.no_pretraining
cnet = SVCNN(config.name,
nclasses=config.num_classes,
cnn_name=config.cnn_name,
pretraining=pretraining)
cnet_2 = MVCNN(config.name,
cnet,
nclasses=config.num_classes,
cnn_name=config.cnn_name,
num_views=config.num_views)
cnet_2.load(
os.path.join(log_dir, config.snapshot_prefix + str(config.weights)))
optimizer = optim.Adam(cnet_2.parameters(),
lr=config.learning_rate,
weight_decay=config.weight_decay,
betas=(0.9, 0.999))
trainer = ModelNetTrainer(cnet_2,
None,
val_loader,
optimizer,
nn.CrossEntropyLoss(),
config,
log_dir,
num_views=config.num_views)
labels, predictions = trainer.update_validation_accuracy(config.weights,
test=True)
import Evaluation_tools as et
eval_file = os.path.join(config.log_dir, '{}.txt'.format(config.name))
et.write_eval_file(config.data, eval_file, predictions, labels,
config.name)
et.make_matrix(config.data, eval_file, config.log_dir)
def stage1():
log_dir = os.path.join(args.log_path, args.name, args.name+'_stage_1')
create_folder(log_dir)
optimizer = optim.Adam(cnet.parameters(), lr=args.lr, weight_decay=args.weight_decay)
train_dataset = SingleImgDataset(args.train_path, scale_aug=False, rot_aug=False, num_models=n_models_train, num_views=args.num_views)
train_loader = torch.utils.data.DataLoader(train_dataset, batch_size=64, shuffle=True, num_workers=0)
val_dataset = SingleImgDataset(args.val_path, scale_aug=False, rot_aug=False, test_mode=True)
val_loader = torch.utils.data.DataLoader(val_dataset, batch_size=64, shuffle=False, num_workers=0)
print('num_train_files: '+str(len(train_dataset.filepaths)))
print('num_val_files: '+str(len(val_dataset.filepaths)))
trainer = ModelNetTrainer(cnet, train_loader, val_loader, optimizer, nn.CrossEntropyLoss(), 'svcnn', log_dir, num_views=1)
trainer.train(args.epoch)
def train_3d_single():
# STAGE 1
print('Stage_1 begin:')
log_dir = args.name + '_stage_1'
create_folder(log_dir)
svcnn = SVCNN(args.name,
nclasses=40,
pretraining=True,
cnn_name=args.cnn_name)
optimizer = optim.Adam(svcnn.parameters(),
lr=args.lr,
weight_decay=args.weight_decay)
train_file = open(args.single_train_path)
train_list = json.load(train_file)
train_dataset = SingleImgDataset(train_list,
scale_aug=False,
rot_aug=False)
train_loader = torch.utils.data.DataLoader(train_dataset,
batch_size=64,
shuffle=True,
num_workers=0)
test_file = open(args.single_test_path)
test_list = json.load(test_file)
val_dataset = SingleImgDataset(test_list,
scale_aug=False,
rot_aug=False,
test_mode=True)
val_loader = torch.utils.data.DataLoader(val_dataset,
batch_size=64,
shuffle=False,
num_workers=0)
print('num_train_files: ' + str(len(train_dataset.data_list)))
print('num_val_files: ' + str(len(val_dataset.data_list)))
trainer = ModelNetTrainer(svcnn,
train_loader,
val_loader,
optimizer,
nn.CrossEntropyLoss(),
'svcnn',
log_dir,
num_views=1)
trainer.train(10)
return svcnn
def train_3d_multi(svcnn):
print('Stage_2 begin:')
log_dir = args.name + '_stage_2'
create_folder(log_dir)
gvcnn = GVCNN(args.name, svcnn, nclasses=40, num_views=args.num_views)
del svcnn
optimizer = optim.Adam(gvcnn.parameters(),
lr=args.lr,
weight_decay=args.weight_decay,
betas=(0.9, 0.999))
train_file = open(args.multi_train_path)
train_list = json.load(train_file)
train_dataset = MultiviewImgDataset(train_list,
scale_aug=False,
rot_aug=False,
num_views=args.num_views)
train_loader = torch.utils.data.DataLoader(train_dataset,
batch_size=args.batchSize,
shuffle=True,
num_workers=0)
test_file = open(args.multi_test_path)
test_list = json.load(test_file)
val_dataset = MultiviewImgDataset(test_list,
scale_aug=False,
rot_aug=False,
num_views=args.num_views)
val_loader = torch.utils.data.DataLoader(val_dataset,
batch_size=args.batchSize,
shuffle=False,
num_workers=0)
print('num_train_files: ' + str(len(train_dataset.data_list)))
print('num_val_files: ' + str(len(val_dataset.data_list)))
trainer = ModelNetTrainer(gvcnn,
train_loader,
val_loader,
optimizer,
nn.CrossEntropyLoss(),
'gvcnn',
log_dir,
num_views=args.num_views)
trainer.train(15)
log_dir = args.name+'_stage_1'
create_folder(log_dir)
cnet = SVCNN(args.name, nclasses=40, pretraining=pretraining, cnn_name=args.cnn_name)
optimizer = optim.Adam(cnet.parameters(), lr=args.lr, weight_decay=args.weight_decay)
n_models_train = args.num_models*args.num_views
train_dataset = SingleImgDataset(args.train_path, scale_aug=False, rot_aug=False, num_models=n_models_train, num_views=args.num_views)
train_loader = torch.utils.data.DataLoader(train_dataset, batch_size=64, shuffle=True, num_workers=0)
val_dataset = SingleImgDataset(args.val_path, scale_aug=False, rot_aug=False, test_mode=True)
val_loader = torch.utils.data.DataLoader(val_dataset, batch_size=64, shuffle=False, num_workers=0)
print('num_train_files: '+str(len(train_dataset.filepaths)))
print('num_val_files: '+str(len(val_dataset.filepaths)))
trainer = ModelNetTrainer(cnet, train_loader, val_loader, optimizer, nn.CrossEntropyLoss(), 'svcnn', log_dir, num_views=1)
trainer.train(30)
# STAGE 2
log_dir = args.name+'_stage_2'
create_folder(log_dir)
cnet_2 = MVCNN(args.name, cnet, nclasses=40, cnn_name=args.cnn_name, num_views=args.num_views)
del cnet
optimizer = optim.Adam(cnet_2.parameters(), lr=args.lr, weight_decay=args.weight_decay, betas=(0.9, 0.999))
train_dataset = MultiviewImgDataset(args.train_path, scale_aug=False, rot_aug=False, num_models=n_models_train, num_views=args.num_views)
train_loader = torch.utils.data.DataLoader(train_dataset, batch_size=args.batchSize, shuffle=False, num_workers=0)# shuffle needs to be false! it's done within the trainer
val_dataset = MultiviewImgDataset(args.val_path, scale_aug=False, rot_aug=False, num_views=args.num_views,test_mode=True)
val_loader = torch.utils.data.DataLoader(val_dataset, batch_size=args.batchSize, shuffle=False, num_workers=0)
shuffle=False,
num_workers=0)
val_dataset = KmeanImgDataset(args.val_path)
val_loader = torch.utils.data.DataLoader(val_dataset,
batch_size=args.batchSize,
shuffle=False,
num_workers=0)
print('num_train_files: ' + str(len(train_dataset.filepaths)))
print('num_val_files: ' + str(len(val_dataset.filepaths)))
trainer = ModelNetTrainer(cnet_2,
train_loader,
val_loader,
optimizer,
nn.CrossEntropyLoss(),
'mvcnn',
log_dir,
num_views=args.num_views)
trainer.train_kmean_threeview(30)
else: # test
path = '~/mvcnn_pytorch-master_ECCV2018_backup_2019_11_22/MVCNN_kmean_cat_no_sort_128/'
modelfile = 'model-00002.pth'
pretraining = not args.no_pretraining
log_dir = args.name
cnet = SVCNN(args.name,
nclasses=40,
pretraining=pretraining,
cnn_name=args.cnn_name)
def train(config):
log(config.log_file, 'Starting...')
pretraining = not config.no_pretraining
log_dir = config.name
create_folder(config.name)
log(config.log_file, '--------------stage 1--------------')
# STAGE 1
log_dir = os.path.join(config.log_dir, config.name + '_stage_1')
create_folder(log_dir)
cnet = SVCNN(config, pretraining=pretraining)
optimizer = optim.Adam(cnet.parameters(),
lr=config.learning_rate,
weight_decay=config.weight_decay)
train_path = os.path.join(config.data, "*/train")
train_dataset = SingleImgDataset(train_path,
config,
scale_aug=False,
rot_aug=False)
train_loader = torch.utils.data.DataLoader(
train_dataset,
batch_size=config.stage1_batch_size,
shuffle=True,
num_workers=0)
val_path = os.path.join(config.data, "*/test")
val_dataset = SingleImgDataset(val_path,
config,
scale_aug=False,
rot_aug=False,
test_mode=True)
val_loader = torch.utils.data.DataLoader(
val_dataset,
batch_size=config.stage1_batch_size,
shuffle=False,
num_workers=0)
log(config.log_file,
'num_train_files: ' + str(len(train_dataset.filepaths)))
log(config.log_file, 'num_val_files: ' + str(len(val_dataset.filepaths)))
trainer = ModelNetTrainer(cnet,
train_loader,
val_loader,
optimizer,
nn.CrossEntropyLoss(),
config,
log_dir,
num_views=1)
trainer.train(config, config.stage1_batch_size)
#cnet.load(os.path.join(log_dir, config.snapshot_prefix + str(30)))
# STAGE 2
log(config.log_file, '--------------stage 2--------------')
log_dir = os.path.join(config.log_dir, config.name + '_stage_2')
create_folder(log_dir)
cnet_2 = MVCNN(cnet, config)
del cnet
optimizer = optim.Adam(cnet_2.parameters(),
lr=config.learning_rate,
weight_decay=config.weight_decay,
betas=(0.9, 0.999))
train_dataset = MultiviewImgDataset(train_path,
config,
scale_aug=False,
rot_aug=False)
train_loader = torch.utils.data.DataLoader(
train_dataset,
batch_size=config.stage2_batch_size,
shuffle=False,
num_workers=0
) # shuffle needs to be false! it's done within the trainer
val_dataset = MultiviewImgDataset(val_path,
config,
scale_aug=False,
rot_aug=False)
val_loader = torch.utils.data.DataLoader(
val_dataset,
batch_size=config.stage2_batch_size,
shuffle=False,
num_workers=0)
log(config.log_file,
'num_train_files: ' + str(len(train_dataset.filepaths)))
log(config.log_file, 'num_val_files: ' + str(len(val_dataset.filepaths)))
trainer = ModelNetTrainer(cnet_2,
train_loader,
val_loader,
optimizer,
nn.CrossEntropyLoss(),
config,
log_dir,
num_views=config.num_views)
trainer.train(config, config.stage2_batch_size)
val_dataset = SingleImgDataset(args.val_path, scale_aug=False, rot_aug=False, test_mode=True)
val_loader = torch.utils.data.DataLoader(val_dataset, batch_size=64, shuffle=False, num_workers=0)
print('num_train_files: '+str(len(train_dataset.filepaths)))
print('num_val_files: '+str(len(val_dataset.filepaths)))
trainer = ModelNetTrainer(cnet, train_loader, val_loader, optimizer, nn.CrossEntropyLoss(), 'svcnn', log_dir, num_views=1)
trainer.train(args.epoch)
if not args.skip_stage1:
stage1()
# STAGE 2
log_dir = os.path.join(args.log_path, args.name, args.name+'_stage_2')
create_folder(log_dir)
cnet_2 = MVCNN(args.name, cnet, nclasses=args.num_class, cnn_name=args.cnn_name, num_views=args.num_views)
del cnet
optimizer = optim.Adam(cnet_2.parameters(), lr=args.lr, weight_decay=args.weight_decay, betas=(0.9, 0.999))
train_dataset = MultiviewImgDataset(args.train_path, scale_aug=False, rot_aug=False, num_models=n_models_train, num_views=args.num_views, num_class=args.num_class)
train_loader = torch.utils.data.DataLoader(train_dataset, batch_size=args.batchSize, shuffle=False, num_workers=0)# shuffle needs to be false! it's done within the trainer
val_dataset = MultiviewImgDataset(args.val_path, scale_aug=False, rot_aug=False, num_views=args.num_views, num_class=args.num_class)
val_loader = torch.utils.data.DataLoader(val_dataset, batch_size=args.batchSize, shuffle=False, num_workers=0)
print('####stage_2####')
print('num_train_files: '+str(len(train_dataset.filepaths)))
print('num_val_files: '+str(len(val_dataset.filepaths)))
trainer = ModelNetTrainer(cnet_2, train_loader, val_loader, optimizer, nn.CrossEntropyLoss(), 'mvcnn', log_dir, num_views=args.num_views, num_class=args.num_class)
trainer.train(args.epoch)
rot_aug=False,
test_mode=True)
val_loader = torch.utils.data.DataLoader(val_dataset,
batch_size=48,
shuffle=False,
num_workers=10)
# 训练集和测试集分别为9843和2468再乖以12
print('num_train_files: ' + str(len(train_dataset.filepaths)))
print('num_val_files: ' + str(len(val_dataset.filepaths)))
# 这里只是定义一个训练器，记录数据，输出loss和acc, svcnn和num_view=1即只要单个图像输入
trainer = ModelNetTrainer(cnet_,
train_loader,
val_loader,
optimizer,
nn.CrossEntropyLoss(),
'svcnn',
log_dir,
num_views=1)
tic1 = time.clock()
trainer.train(n_epochs=0) # 测试时设为1，看能否完整跑完两个阶段
toc1 = time.clock()
print('The training time of first stage: %d m' % ((toc1 - tic1) / 60))
# STAGE 2
print('###################Stage 2####################')
log_dir = args.name + '_stage_2'
create_folder(log_dir)
# cnet_2与cnet采用相同的网络
cnet_2 = MVCNN(args.name,
val_dataset = KmeanImgDataset(args.val_path,
fea_type=args.fea_type,
dataset=dataset)
val_loader = torch.utils.data.DataLoader(val_dataset,
batch_size=args.batchSize,
shuffle=False,
num_workers=0)
print('num_train_files: ' + str(len(train_dataset.filepaths)))
print('num_val_files: ' + str(len(val_dataset.filepaths)))
trainer = ModelNetTrainer((nem, cnet_2),
train_loader,
val_loader,
optimizer,
nn.CrossEntropyLoss(),
None,
log_dir,
num_views=args.num_views,
class_num=class_num)
trainer.train_nem_mvcnn(args.epoch)
else: # test
path = '/mnt/cloud_disk/huangjj/exp_mvcnn/' + args.name
modelfile = args.modelfile
pretraining = not args.no_pretraining
log_dir = args.name
cnet = SVCNN(args.name,
nclasses=class_num,
pretraining=pretraining,
val_dataset = SingleImgDataset(args.val_path,
scale_aug=False,
rot_aug=False,
test_mode=True)
val_loader = torch.utils.data.DataLoader(val_dataset,
batch_size=64,
shuffle=False,
num_workers=0)
print('num_train_files: ' + str(len(train_dataset.filepaths)))
print('num_val_files: ' + str(len(val_dataset.filepaths)))
trainer = ModelNetTrainer(cnet,
train_loader,
val_loader,
optimizer_model,
optimizer_centerloss,
softmax_loss,
center_loss,
'svcnn',
log_dir,
num_views=1)
trainer.train(20)
# STAGE 2
log_dir = 'smtcloss_wonor_' + args.name + '_stage_2_' + localtime
create_folder(log_dir)
cnet_2 = MVCNN(args.name,
cnet,
nclasses=40,
cnn_name=args.cnn_name,
num_views=args.num_views)
del cnet
# cnet_2.module.load_state_dict(torch.load("Vgg11_Seg_white_stage_2/Vgg11_Seg_white/model-00027.pth"))
# cnet_2.module.eval()
# else:
# cnet_2.load_state_dict(torch.load("Vgg11_Seg_white_stage_2/Vgg11_Seg_white/model-00027.pth"))
# cnet_2.eval()
### -------------------------------------------------------------------------------------------------------------
if use_dataparallel:
cnet_2 = nn.DataParallel(cnet_2)
cnet_2.to(device)
optimizer = optim.Adam(cnet_2.parameters(), lr=args.lr, weight_decay=args.weight_decay, betas=(0.9, 0.999))
train_dataset = MultiviewImgDataset(args.train_path, scale_aug=False, rot_aug=False, num_models=n_models_train, num_views=args.num_views,KNU_data=args.KNU_Data,pixel_augmentation=args.pixel_augmentation)
train_loader = torch.utils.data.DataLoader(train_dataset, batch_size=args.batchSize, shuffle=False, num_workers=0)# shuffle needs to be false! it's done within the trainer
val_dataset = MultiviewImgDataset(args.val_path, scale_aug=False, rot_aug=False, num_views=args.num_views,KNU_data=args.KNU_Data,pixel_augmentation=args.pixel_augmentation)
val_loader = torch.utils.data.DataLoader(val_dataset, batch_size=args.batchSize, shuffle=False, num_workers=0)
print('num_train_files: '+str(len(train_dataset.filepaths)))
print('num_val_files: '+str(len(val_dataset.filepaths)))
if(args.loss_type == 'focal_loss'):
focal_loss = FocalLoss(gamma=2, alpha=0.25)
trainer = ModelNetTrainer(cnet_2, train_loader, val_loader, optimizer, focal_loss, 'mvcnn', log_dir, num_views=args.num_views, nClasses=nclasses)
else:
trainer = ModelNetTrainer(cnet_2, train_loader, val_loader, optimizer, nn.CrossEntropyLoss(), 'mvcnn', log_dir, num_views=args.num_views, nClasses=nclasses)
trainer.train(30, use_dataparallel)