if opt.PCB:
fnorm = torch.norm(ff, p=2, dim=1, keepdim=True) * np.sqrt(6)
ff = ff.div(fnorm.expand_as(ff))
ff = ff.view(ff.size(0), -1)
else:
fnorm = torch.norm(ff, p=2, dim=1, keepdim=True)
ff = ff.div(fnorm.expand_as(ff))
features = torch.cat((features, ff), 0)
return features
if opt.use_dense:
model_structure = ft_net_dense(751)
else:
model_structure = ft_net(751)
model = load_network(model_structure)
# Remove the final fc layer and classifier layer
if not opt.PCB:
model.model.fc = nn.Sequential()
model.classifier = nn.Sequential()
else:
model = PCB_test(model)
# Change to test mode
model = model.eval()
model = model.cuda()
file_path = '../friends2/'
gallery_path = image_datasets['gallery'].imgs
query_path = image_datasets['query'].imgs
gallery_cam, gallery_label, gallery_frames = get_id(gallery_path)
query_cam, query_label, query_frames = get_id(query_path)
######################################################################
# Load Collected data Trained model
class_num = 100
print('-------test-----------')
if opt.use_dense:
model_structure = ft_net_dense(class_num)
else:
model_structure = ft_net(class_num)
if opt.PCB:
model_structure = PCB(class_num)
model = load_network(model_structure)
# Remove the final fc layer and classifier layer
if not opt.PCB:
model.model.fc = nn.Sequential()
model.classifier = nn.Sequential()
else:
model = PCB_test(model)
# Change to test mode
model = model.eval()
def main(mode):
ImageFile.LOAD_TRUNCATED_IMAGES = True
######################################################################
# Options
# --------
gpu_ids = '0'
which_epoch = 'last'
name = 'ft_ResNet50'
batchsize = 32
ms = '1'
if mode == 'all': # 전체 공고 대상
test_dir = 'C:/Users/kdan/BigJob12/main_project/_db/data'
mode_path = 'preprocessed_data'
else: # 특정 공고 대상
test_dir = 'C:/Users/kdan/BigJob12/main_project/_db/data/model_data'
mode_path = 'gallery'
# load the training config
config_path = os.path.join(
'C:/Users/kdan/BigJob12/main_project/_src/data_analysis/re_id/code/model',
name, 'opts.yaml')
with open(config_path, 'r') as stream:
config = yaml.load(stream)
stride = config['stride']
if 'nclasses' in config: # tp compatible with old config files
nclasses = config['nclasses']
######################################################################
# GPU Options
# --------
str_ids = gpu_ids.split(',')
gpu_ids = []
for str_id in str_ids:
id = int(str_id)
if id >= 0:
gpu_ids.append(id)
print('We use the scale: %s' % ms)
str_ms = ms.split(',')
ms = []
for s in str_ms:
s_f = float(s)
ms.append(math.sqrt(s_f))
# set gpu ids
if len(gpu_ids) > 0:
torch.cuda.set_device(gpu_ids[0])
cudnn.benchmark = True
######################################################################
# Data transform
data_transforms = transforms.Compose([
transforms.Resize((256, 128), interpolation=3),
transforms.ToTensor(),
transforms.Normalize([0.485, 0.456, 0.406], [0.229, 0.224, 0.225])
])
######################################################################
# Load data
data_dir = test_dir
image_datasets = {
x: datasets.ImageFolder(os.path.join(data_dir, x), data_transforms)
for x in [mode_path]
}
dataloaders = {
x: torch.utils.data.DataLoader(image_datasets[x],
batch_size=batchsize,
shuffle=False,
num_workers=0)
for x in [mode_path]
}
use_gpu = torch.cuda.is_available()
# Load Collected data Trained model
print('-------test-----------')
model_structure = ft_net(nclasses, stride=stride)
model = load_network(model_structure, name, which_epoch)
# Remove the final fc layer and classifier layer
model.classifier.classifier = nn.Sequential()
print('model')
# Change to test mode
model = model.eval()
if use_gpu:
model = model.cuda()
# Extract feature
with torch.no_grad():
gallery_feature = extract_feature(model, dataloaders[mode_path], ms)
# Save to Matlab for check
gallery_result = {'gallery_f': gallery_feature.numpy()}
if mode == 'all':
scipy.io.savemat(
'C:/Users/kdan/BigJob12/main_project/_src/data_analysis/re_id/code/gallery_all_result.mat',
gallery_result)
pd.DataFrame(gallery_result['gallery_f']).to_csv(
('gallery_all_result.csv'))
else:
scipy.io.savemat(
'C:/Users/kdan/BigJob12/main_project/_src/data_analysis/re_id/code/gallery_result.mat',
gallery_result)
print(name)
######################################################################
# Finetuning the convnet
# ----------------------
#
# Load a pretrainied model and reset final fully connected layer.
#
if opt.use_dense:
model = ft_net_dense(len(class_names), opt.droprate, circle=opt.circle)
elif opt.use_NAS:
model = ft_net_NAS(len(class_names), opt.droprate)
else:
model = ft_net(len(class_names),
opt.droprate,
opt.stride,
circle=opt.circle)
if opt.PCB:
model = PCB(len(class_names))
opt.nclasses = len(class_names)
print(model)
if not opt.PCB:
ignored_params = list(map(id, model.classifier.parameters()))
base_params = filter(lambda p: id(p) not in ignored_params,
model.parameters())
optimizer_ft = optim.SGD([{
'params': base_params,
network.cuda(gpu_ids[0])
#nn.DataParallel(network, device_ids=[2,3]).cuda()
######################################################################
# Finetuning the convnet
# ----------------------
#
# Load a pretrainied model and reset final fully connected layer.
#
def load_network_path(network, save_path):
network.load_state_dict(torch.load(save_path))
return network
model_structure = ft_net(len(class_names))
target = 'experiment_Result/p1'
path = target + '/ft_ResNet50/net_39.pth'
print('origin_with: ' + target)
model = load_network_path(model_structure, path)
model_from = ft_net(len(class_names))
nw.dup_net6(model_from, model, 'model.conv1', 'model.bn1', 'xx', 'xx',
'model.fc', 'classifier') #C1+FC remain
if use_gpu:
model = model.cuda()
#nn.DataParallel(model, device_ids=[2,3]).cuda()
criterion = nn.CrossEntropyLoss()
save_path = os.path.join('./model', name, 'net_best.pth')
else:
save_path = model_name
network.load_state_dict(torch.load(save_path))
return network
# print('------------'+str(len(clas_names))+'--------------')
if opt.use_dense:
# print(len(class_names['train']))
model = ft_net_dense(
751) # 751 class for training data in market 1501 in total
model_pred = ft_net_dense(
751) # 751 class for training data in market 1501 in total
else:
model = ft_net(751)
model_pred = ft_net(751)
if use_gpu:
model = model.cuda()
model_pred = model_pred.cuda()
# load_network(model_pred)
# model_pred.eval()
# read dcgan data
class dcganDataset(Dataset):
def __init__(self, root, transform=None, targte_transform=None):
super(dcganDataset, self).__init__()
self.image_dir = os.path.join(opt.data_dir, root)
opt.num_gpus = num_gpus
num_classes, data_loader = get_data_loader(opt.data_path, opt.batch_size)
if opt.num_classes < num_classes:
opt.num_classes = num_classes
class_split = None
if opt.model_parallel:
# If using model parallel, split the number of classes
# accroding to the number of GPUs
class_split = get_class_split(opt.num_classes, num_gpus)
model = ft_net(
feature_dim=256,
num_classes=opt.num_classes,
num_gpus=num_gpus,
am=opt.am,
model_parallel=opt.model_parallel,
class_split=class_split
)
optimizer_ft = optim.SGD(
model.parameters(),
lr=opt.lr,
weight_decay=5e-4,
momentum=0.9,
nesterov=True
)
if opt.model_parallel:
# When using model parallel, we wrap all the model except classifier in DataParallel
model.backbone = nn.DataParallel(model.backbone).cuda()
model.features = nn.DataParallel(model.features).cuda()
search = target + '/ft_ResNet50'
file_list = os.listdir(search)
for file_name in file_list:
#if file_name[-5:] == 't.pth':
if file_name[-6:] == '39.pth':
#if file_name[-6:-5] == '4':
#if file_name[-3:] == 'pth':
path = search + '/' + file_name
else:
continue
print(path)
#model_structure = ft_net(702) #duke
model_structure = ft_net(num_class) #market
model = load_network_path(model_structure, path)
# Remove the final fc layer and classifier layer
model.model.fc = nn.Sequential()
model.classifier = nn.Sequential()
# Change to test mode
model = model.eval()
if use_gpu:
model = model.cuda()
# Extract feature
gallery_feature = extract_feature(model, dataloaders['gallery'])
query_feature = extract_feature(model, dataloaders['query'])
# Save to Matlab for check
class_num = {}
class_num['train'] = len(os.listdir(dataset_train_dir))
class_num['val'] = len(os.listdir(dataset_val_dir))
print('class_num train = %d test = %d' %
(class_num['train'], class_num['val']))
if opt.use_dense:
# print(len(class_names['train']))
model = ft_net_dense(
class_num['train']
) # 751 class for training data in market 1501 in total
model_pred = ft_net_dense(
class_num['train']
) # 751 class for training data in market 1501 in total
else:
model = ft_net(class_num['train'])
model_pred = ft_net(class_num['train'])
if use_gpu:
model = model.cuda()
model_pred = model_pred.cuda()
######################################################################
# Training the model
# ------------------
y_loss = {} # loss history
y_loss['train'] = []
y_loss['val'] = []
y_err = {}
y_err['train'] = []
if opt.re_rank:
newD = re_ranking(q_g_dist, q_q_dist, g_g_dist)
newD = np.transpose(newD)
res += _cmc_core(newD, gallery_labels_set, query_label, k)
res /= Times
return res
if __name__ == '__main__':
# Load Collected data Trained model
print('-------test-----------')
if opt.use_dense:
model_structure = ft_net_dense(316)
else:
model_structure = ft_net(316)
model = load_network(model_structure)
# Remove the final fc layer and classifier layer
model.model.fc = nn.Sequential()
model.classifier = nn.Sequential()
# Change to test mode
model = model.eval()
if use_gpu:
model = model.cuda()
# Extract feature
gallery_feature, gallery_label = extract_feature(model,
dataloaders['gallery'])
query_feature, query_label = extract_feature(model, dataloaders['query'])
def test():
print('-------test-----------')
time_start = time.time()
if opt.use_dense:
model_structure = ft_net_dense(opt.nclasses)
#elif opt.use_NAS:
# model_structure = ft_net_NAS(opt.nclasses)
else:
model_structure = ft_net(opt.nclasses, stride=opt.stride)
if opt.PCB:
model_structure = PCB(opt.nclasses)
#if opt.fp16:
# model_structure = network_to_half(model_structure)
model = load_network(model_structure)
# Remove the final fc layer and classifier layer
if opt.PCB:
#if opt.fp16:
# model = PCB_test(model[1])
#else:
model = PCB_test(model)
else:
#if opt.fp16:
#model[1].model.fc = nn.Sequential()
#model[1].classifier = nn.Sequential()
#else:
model.classifier.classifier = nn.Sequential()
# Change to test mode
model = model.eval()
if use_gpu:
model = model.cuda()
# Extract feature
with torch.no_grad():
gallery_feature = extract_feature(model, dataloaders['gallery'])
query_feature = extract_feature(model, dataloaders['query'])
if opt.multi:
mquery_feature = extract_feature(model, dataloaders['multi-query'])
# Save to Matlab for check
result = {
'gallery_f': gallery_feature.numpy(),
'gallery_label': gallery_label,
'gallery_cam': gallery_cam,
'query_f': query_feature.numpy(),
'query_label': query_label,
'query_cam': query_cam
}
scipy.io.savemat('pytorch_result.mat', result)
print(opt.name)
os.system('python evaluate_gpu.py')
if opt.multi:
result = {
'mquery_f': mquery_feature.numpy(),
'mquery_label': mquery_label,
'mquery_cam': mquery_cam
}
scipy.io.savemat('multi_query.mat', result)
time_end = time.time()
print('totally cost', time_end - time_start)
opt.nclasses = len(class_names)
if opt.use_dense:
model = ft_net_dense(len(class_names), opt.droprate, opt.stride, None,
opt.pool)
elif opt.use_NAS:
model = ft_net_NAS(len(class_names), opt.droprate, opt.stride)
elif opt.use_SE:
model = ft_net_SE(len(class_names), opt.droprate, opt.stride, opt.pool,
None)
elif opt.use_DSE:
model = ft_net_DSE(len(class_names), opt.droprate, opt.stride,
opt.pool)
elif opt.use_IR:
model = ft_net_IR(len(class_names), opt.droprate, opt.stride)
else:
model = ft_net(len(class_names), opt.droprate, opt.stride, None,
opt.pool)
if opt.PCB:
model = PCB(len(class_names))
if opt.CPB:
model = CPB(len(class_names))
if opt.angle:
model = ft_net_angle(len(class_names), opt.droprate, opt.stride)
elif opt.arc:
model = ft_net_arc(len(class_names), opt.droprate, opt.stride)
if opt.init_name != 'imagenet':
old_opt = parser.parse_args()
init_model, old_opt, _ = load_network(opt.init_name, old_opt)
print('Training complete in {:.0f}m {:.0f}s'.format(time_elapsed // 60, time_elapsed % 60))
model.load_state_dict(last_model_wts)
save_network(model)
return model
# 保存pth文件
def save_network(network):
save_path = './net.pth' #生成pth文件
torch.save(network.cpu().state_dict(), save_path)
if torch.cuda.is_available():
network.cuda(0)
#加载网络，丢弃全连接层
model = ft_net(len(class_names), droprate, stride)
ignored_params = list(map(id, model.classifier.parameters() ))
base_params = filter(lambda p: id(p) not in ignored_params, model.parameters())
optimizer_ft = optim.SGD([
{'params': base_params, 'lr': 0.1*lr},
{'params': model.classifier.parameters(), 'lr': lr}
], weight_decay=5e-4, momentum=0.9, nesterov=True)
#学习率参数
exp_lr_scheduler = lr_scheduler.StepLR(optimizer_ft, step_size=40, gamma=0.1)
if use_gpu:
model = model.cuda()
criterion = nn.CrossEntropyLoss()
model = train_model(model, criterion, optimizer_ft, exp_lr_scheduler,num_epochs=60)#训练的epoch数量设置
query_path = image_datasets['query'].imgs
gallery_cam,gallery_label = get_id(gallery_path)
query_cam,query_label = get_id(query_path)
if multi_query:
mquery_path = image_datasets['multi-query'].imgs
mquery_cam,mquery_label = get_id(mquery_path)
#%%
train_all = '_all'
image_datasets_train = {}
image_datasets_train['train'] = datasets.ImageFolder(os.path.join(data_dir, 'train' + train_all), data_transforms)
class_names_train = image_datasets_train['train'].classes
model = ft_net(len(class_names_train), 0.2, 2)
model.load_state_dict(torch.load('net_last.pth'))
temp = nn.Sequential(*list(model.children())[:-1])
res50_conv = nn.Sequential(*list(temp[0].children())[:-1])
#for param in res50_conv.parameters():
# param.requires_grad = False
res50_conv.eval()
model = res50_conv
if use_gpu:
model = model.cuda()
# Extract feature
with torch.no_grad():
gallery_feature = extract_feature(model,dataloaders['gallery'])
######################################################################
# Load Collected data Trained model
print('-------test-----------')
search = target + '/ft_ResNet50'
file_list = os.listdir(search)
print(search)
for tepoch in range(euntil + 1):
if tepoch < efrom:
continue
file_name = 'net_%d.pth' % tepoch
path = search + '/' + file_name
model = ft_net(num_class) #market
model = load_network_path(model, path)
# Change to test mode
model = model.eval()
if use_gpu:
model = model.cuda()
# Extract feature
query_feature = extract_feature(model, dataloaders['query'])
gallery_feature = extract_feature(model, dataloaders['gallery'])
# Save to Matlab for check
query_label = np.asarray(query_label)
gallery_label = np.asarray(gallery_label)
transforms.Resize((256, 128), interpolation=3),
transforms.ToTensor(),
transforms.Normalize([0.485, 0.456, 0.406], [0.229, 0.224, 0.225])
])
# load training dataDatasetFolder
print('Starting loading training data: ', args.dataset)
train_dataset = loader.DatasetFolder(os.path.join(data_dir, 'train'),
transform=train_transform)
class_names = train_dataset.classes
dataset_sizes_train = len(train_dataset)
use_gpu = torch.cuda.is_available()
# Define a model
model1 = ft_net(len(class_names))
model2 = ft_net(len(class_names))
if args.pattern:
model_tmp = ft_net(len(class_names))
if use_gpu:
model1 = model1.cuda()
model2 = model2.cuda()
if args.pattern:
model_tmp = model_tmp.cuda()
# Load a pretrainied model
if args.resume:
# two pretrained SELF models (ignore when you start)
model_name1 = 'market_noise_0.2_batch_32_refine_lr_1.0e-02_model1.t'
print('Initilizaing weights with ', args.resume)
model_path1 = checkpoint_path + model_name1
transforms.Resize((256, 128), interpolation=3),
transforms.ToTensor(),
transforms.Normalize([0.485, 0.456, 0.406], [0.229, 0.224, 0.225]),
])
data_transforms_visualize = transforms.Compose(
[transforms.Resize((256, 128), interpolation=3),
transforms.ToTensor()])
image_datasets = {
x: datasets.ImageFolder(os.path.join(opt.data_dir, x), data_transforms)
for x in ['gallery', 'query']
}
query_path, _ = image_datasets['query'].imgs[qi]
query_img = Image.open(query_path)
query = data_transforms(query_img)
query_vis = data_transforms_visualize(query_img)
query.unsqueeze_(0)
model = ft_net(class_num=575)
model.load_state_dict(torch.load('./model/ft_ResNet/net_99.pth'))
model.classifier.classifier = nn.Sequential()
model = model.eval().cuda()
heatmap = generate_heatmap(model, query)
heatmap2d(query_vis, heatmap)
model,
os.path.join('./model', name,
'whole_net_best' + '_stage_' + str(stage)) + '.pth')
return model
# print('------------'+str(len(clas_names))+'--------------')
if True: # opt.use_dense:
model = ft_net_dense(
id_num, cam_num,
istrain=True) # 751 class for training data in all dataset in total
# model = ft_net_dense(751, 6, istrain=True) # 751 class for training data in market 1501 in total
# model = ft_net_dense(702, 8, istrain=True) # 751 class for training data in duke in total
else:
model = ft_net(751)
# print(model)
if use_gpu:
model = model.cuda()
criterion = nn.CrossEntropyLoss()
# Decay LR by a factor of 0.1 every 40 epochs
dir_name = os.path.join('./model', name)
if not os.path.isdir(dir_name):
os.mkdir(dir_name)
# save opts
torch.save(network.cpu().state_dict(), save_path)
if torch.cuda.is_available():
network.cuda(gpu_ids[0])
######################################################################
# Finetuning the convnet
# ----------------------
#
# Load a pretrainied model and reset final fully connected layer.
#
if opt.use_dense:
model = ft_net_dense(len(class_names), opt.droprate)
else:
model = ft_net(len(class_names), opt.droprate)
if opt.PCB:
model = PCB(len(class_names))
print(model)
if use_gpu:
model = model.cuda()
criterion = nn.CrossEntropyLoss()
if not opt.PCB:
ignored_params = list(map(id, model.model.fc.parameters() )) + list(map(id, model.classifier.parameters() ))
base_params = filter(lambda p: id(p) not in ignored_params, model.parameters())
optimizer_ft = optim.SGD([
save_network(model, 'last')
return model
# Save model
def save_network(network, epoch_label):
save_filename = 'net_%s.pth' % epoch_label
save_path = os.path.join('./model', name, save_filename)
torch.save(network.cpu().state_dict(), save_path)
if torch.cuda.is_available():
network.cuda(gpu_ids[0])
# Finetuning the convnet
# Load a pretrainied model and reset final fully connected layer.
model = ft_net(len(image_datasets.classes))
print(model)
if torch.cuda.is_available():
model = model.cuda()
criterion = nn.CrossEntropyLoss()
ignored_params = list(map(id, model.model.fc.parameters())) + list(
map(id, model.classifier.parameters()))
base_params = filter(lambda p: id(p) not in ignored_params, model.parameters())
optimizer_ft = optim.SGD([{
'params': base_params,
'lr': 0.01
}, {
'params': model.model.fc.parameters(),
'lr': 0.1
gallery_path = image_datasets['gallery'].imgs
query_path = image_datasets['query'].imgs
gallery_cam, gallery_label = get_id(gallery_path)
query_cam, query_label = get_id(query_path)
######################################################################
# Load Collected data Trained model
nnn = opt.training_set_classes
# duke-market 702
print('-------test-----------')
if opt.use_two_stream_resnet:
model_structure = two_stream_resnet(nnn, True)
else:
model_structure = ft_net(nnn)
model = load_network(model_structure)
#model.model.avgpool = nn.AdaptiveMaxPool2d((7,1))
# Remove the final fc layer and classifier layer
if not opt.use_two_stream_resnet:
model.classifier.fc = nn.Sequential()
model.classifier = nn.Sequential()
# Change to test mode
model = model.eval()
if use_gpu:
model = model.cuda()
# Extract feature
gallery_path = image_datasets['gallery'].imgs
query_path = image_datasets['query'].imgs
mquery_path = image_datasets['multi-query'].imgs
gallery_cam,gallery_label = get_id(gallery_path)
query_cam,query_label = get_id(query_path)
mquery_cam,mquery_label = get_id(mquery_path)
######################################################################
# Load Collected data Trained model
print('-------test-----------')
if opt.use_dense:
model_structure = ft_net_dense(751)
else:
model_structure = ft_net(751)
if opt.PCB:
model_structure = PCB(751)
model = load_network(model_structure)
#model = model_structure
# Remove the final fc layer and classifier layer
#if not opt.PCB:
# model.model.fc = nn.Sequential()
# model.classifier = nn.Sequential()
#else:
# model = PCB_test(model)
# Change to test mode
return model
######################################################################
# Train and evaluate
# ^^^^^^^^^^^^^^^^^^
#
# It should take around 1-2 hours on GPU.
#
dir_name = os.path.join('./model', name)
if not os.path.exists('model'):
os.mkdir('model')
# class_num *= 3
print('class_num = %d' % (class_num))
model = ft_net(class_num)
if use_gpu:
model.cuda()
if False:
model = load_whole_network(model, name, 'pretrain')
opt.lr *= 0.5
# print('model structure')
# print(model)
criterion = nn.CrossEntropyLoss()
criterion_soft = SoftLabelLoss()
classifier_id = list(map(id, model.classifier.parameters())) \
+ list(map(id, model.model.fc.parameters()))
def __init__(self, gpuID, model_path):
####################################################
# Options
# ------------------
self.gpu_ids = 1
self.which_epoch = '59'
self.batch_size = 1
self.use_dense = False
self.use_PCB = False
self.model_path = model_path
self.class_num = 751
self.score_threshold = 0.9
self.confidence_threshold = 0.6
####################################################
# Set gpu
# ------------------
use_gpu = torch.cuda.is_available()
if not use_gpu:
print('can not user gpu')
exit()
torch.cuda.set_device(self.gpu_ids)
####################################################
# Load model
# ------------------
print('load model...')
if self.use_dense:
model_structure = ft_net_dense(self.class_num)
else:
model_structure = ft_net(self.class_num)
if self.use_PCB:
model_structure = PCB(self.class_num)
model = utils.load_network(model_structure, self.model_path,
self.which_epoch)
# Remove the final fc layer and classifier layer
if not self.use_PCB:
model.model.fc = nn.Sequential()
model.classifier = nn.Sequential()
else:
model = PCB_test(model)
model = model.eval()
#print(model)
if use_gpu:
model = model.cuda()
self.model = model
####################################################
# Set Transform
# ------------------
print('set transform...')
self.data_transforms = transforms.Compose([
transforms.Resize((256, 128), interpolation=3),
transforms.ToTensor(),
transforms.Normalize([0.485, 0.456, 0.406], [0.229, 0.224, 0.225])
])
if self.use_PCB:
self.data_transforms = transforms.Compose([
transforms.Resize((384, 192), interpolation=3),
transforms.ToTensor(),
transforms.Normalize([0.485, 0.456, 0.406],
[0.229, 0.224, 0.225])
])
torch.save(network.cpu().state_dict(), save_path)
if torch.cuda.is_available:
network.cuda(gpu_ids[0])
######################################################################
# Finetuning the convnet
# ----------------------
#
# Load a pretrainied model and reset final fully connected layer.
#
if opt.use_three_stream_resnet:
model = three_stream_resnet(len(class_names))
else:
model = ft_net(len(class_names))
print(model)
if use_gpu:
model = model.cuda()
criterion = nn.CrossEntropyLoss()
if opt.use_three_stream_resnet:
ignored_params = list(map(id, model.model.fc.parameters()))
ignored_params += (list(map(id, model.classifier_original.parameters())) +
list(map(id, model.classifier_upper.parameters())) +
list(map(id, model.classifier_lower.parameters())))
base_params = filter(lambda p: id(p) not in ignored_params,
model.parameters())
optimizer_ft = optim.SGD([
import shutil
import re
from PIL import Image
from torchvision import datasets, transforms, models
from model import ft_net
data_dir='/home/luke/'
target=4
ip='127.0.0.2' #ip address of the server
port_num=1234 #port to connect
sc=True #True: panda acts as server; False: panda acts as client;
test_transforms = transforms.Compose([transforms.Resize(size=(256,128),interpolation=3),transforms.ToTensor(),transforms.Normalize([0.485, 0.456, 0.406], [0.229, 0.224, 0.225])])
device=torch.device('cpu')
imsize = 256
model1=ft_net(6)
#model.load_state_dict(torch.load(file_path))
model1.load_state_dict(torch.load('net_9.pth'))
model1.eval()
def predict_image(image):
#image=Image.open(image_p)
image_tensor = test_transforms(image).float()
image_tensor = image_tensor.unsqueeze_(0)
input = Variable(image_tensor)
input = input.to(device)
print (input.shape)
output = model1(input)
index = output.data.cpu().numpy().argmax()
return index
camera_id.append(int(camera[0]))
return camera_id, labels
gallery_path = image_datasets['gallery'].imgs
query_path = image_datasets['query'].imgs
gallery_cam,gallery_label = get_id(gallery_path)
query_cam,query_label = get_id(query_path)
######################################################################
# Load Collected data Trained model
print('-------test-----------')
if opt.use_dense:
model_structure = ft_net_dense(767)
else:
model_structure = ft_net(767)
if opt.use_resnext:
model_structure = resnext50(num_classes=767)
model = load_network(model_structure)
# Remove the final fc layer and classifier layer
#model.model.fc = nn.Sequential()
#model.classifier = nn.Sequential()
# Change to test mode
model = model.eval()
if use_gpu:
model = model.cuda()
# Extract feature
gallery_feature = extract_feature(model,dataloaders['gallery'])
query_path = image_datasets['query'].imgs
gallery_cam, gallery_label = get_id(gallery_path)
query_cam, query_label = get_id(query_path)
if opt.multi:
mquery_path = image_datasets['multi-query'].imgs
mquery_cam, mquery_label = get_id(mquery_path)
######################################################################
# Load Collected data Trained model
print('-------test-----------')
if opt.use_dense:
model_structure = ft_net_dense(opt.nclasses)
else:
model_structure = ft_net(opt.nclasses)
if opt.PCB:
model_structure = PCB(opt.nclasses)
#if opt.fp16:
# model_structure = network_to_half(model_structure)
model = load_network(model_structure)
# Remove the final fc layer and classifier layer
if opt.PCB:
#if opt.fp16:
# model = PCB_test(model[1])
#else:
model = PCB_test(model)
save_path = os.path.join('./model2/', 'net_%s.pth' % opt.which_epoch)
# os.path.join('./model',name,'net_%s.pth'%opt.which_epoch)
network.load_state_dict(torch.load(save_path))
return network
use_gpu = torch.cuda.is_available()
# Load Collected data Trained model
print('-------test-----------')
if opt.use_dense:
model_structure = ft_net_dense(opt.nclasses)
elif opt.use_NAS:
model_structure = ft_net_NAS(opt.nclasses)
else:
model_structure = ft_net(opt.nclasses, stride=opt.stride)
model = load_network(model_structure)
model.classifier.classifier = nn.Sequential()
model = model.eval()
if use_gpu:
model = model.cuda()
# encoder = create_box_encoder('./mars-small128.pb', batch_size = 32)
######################################################################
# Base Functions
######################################################################
class DetectorAPI:
def train(opt):
version = torch.__version__
fp16 = opt.fp16
data_dir = opt.data_dir
name = opt.name
str_ids = opt.gpu_ids.split(',')
gpu_ids = []
for str_id in str_ids:
gid = int(str_id)
if gid >= 0:
gpu_ids.append(gid)
# set gpu ids
if len(gpu_ids) > 0:
torch.cuda.set_device(gpu_ids[0])
cudnn.benchmark = True
######################################################################
# Load Data
# ---------
#
transform_train_list = [
# transforms.RandomResizedCrop(size=128, scale=(0.75,1.0), ratio=(0.75,1.3333), interpolation=3), #Image.BICUBIC)
transforms.Resize((256, 128), interpolation=3),
transforms.Pad(10),
transforms.RandomCrop((256, 128)),
transforms.RandomHorizontalFlip(),
transforms.ToTensor(),
transforms.Normalize([0.485, 0.456, 0.406], [0.229, 0.224, 0.225])
]
transform_val_list = [
transforms.Resize(size=(256, 128), interpolation=3), # Image.BICUBIC
transforms.ToTensor(),
transforms.Normalize([0.485, 0.456, 0.406], [0.229, 0.224, 0.225])
]
if opt.PCB:
transform_train_list = [
transforms.Resize((384, 192), interpolation=3),
transforms.RandomHorizontalFlip(),
transforms.ToTensor(),
transforms.Normalize([0.485, 0.456, 0.406], [0.229, 0.224, 0.225])
]
transform_val_list = [
transforms.Resize(size=(384, 192),
interpolation=3), # Image.BICUBIC
transforms.ToTensor(),
transforms.Normalize([0.485, 0.456, 0.406], [0.229, 0.224, 0.225])
]
if opt.erasing_p > 0:
transform_train_list = transform_train_list + [
RandomErasing(probability=opt.erasing_p, mean=[0.0, 0.0, 0.0])
]
if opt.color_jitter:
transform_train_list = [
transforms.ColorJitter(
brightness=0.1, contrast=0.1, saturation=0.1, hue=0)
] + transform_train_list
# print(transform_train_list)
data_transforms = {
'train': transforms.Compose(transform_train_list),
'val': transforms.Compose(transform_val_list),
}
train_all = ''
if opt.train_all:
train_all = '_all'
image_datasets = {}
image_datasets['train'] = datasets.ImageFolder(
os.path.join(data_dir, 'train' + train_all), data_transforms['train'])
image_datasets['val'] = datasets.ImageFolder(os.path.join(data_dir, 'val'),
data_transforms['val'])
dataloaders = {
x: torch.utils.data.DataLoader(image_datasets[x],
batch_size=opt.batchsize,
shuffle=True,
num_workers=8,
pin_memory=True)
# 8 workers may work faster
for x in ['train', 'val']
}
dataset_sizes = {x: len(image_datasets[x]) for x in ['train', 'val']}
class_names = image_datasets['train'].classes
use_gpu = torch.cuda.is_available()
#since = time.time()
#inputs, classes = next(iter(dataloaders['train']))
#print('time used for loading data: %ds' %(time.time() - since))
######################################################################
# Training the model
# ------------------
#
# Now, let's write a general function to train a model. Here, we will
# illustrate:
#
# - Scheduling the learning rate
# - Saving the best model
#
# In the following, parameter ``scheduler`` is an LR scheduler object from
# ``torch.optim.lr_scheduler``.
y_loss = {} # loss history
y_loss['train'] = []
y_loss['val'] = []
y_err = {}
y_err['train'] = []
y_err['val'] = []
def train_model(model, criterion, optimizer, scheduler, num_epochs=25):
since = time.time()
results = []
for epoch in range(num_epochs):
print('Epoch {}/{}'.format(epoch, num_epochs - 1))
# Each epoch has a training and validation phase
for phase in ['train', 'val']:
if phase == 'train':
scheduler.step()
model.train(True) # Set model to training mode
else:
model.train(False) # Set model to evaluate mode
running_loss = 0.0
running_corrects = 0.0
# Iterate over data.
pbar = tqdm(dataloaders[phase])
for inputs, labels in pbar:
# get the inputs
now_batch_size, c, h, w = inputs.shape
if now_batch_size < opt.batchsize: # skip the last batch
continue
# print(inputs.shape)
# wrap them in Variable
if use_gpu:
inputs = Variable(inputs.cuda().detach())
labels = Variable(labels.cuda().detach())
else:
inputs, labels = Variable(inputs), Variable(labels)
# if we use low precision, input also need to be fp16
# if fp16:
# inputs = inputs.half()
# zero the parameter gradients
optimizer.zero_grad()
# forward
if phase == 'val':
with torch.no_grad():
outputs = model(inputs)
else:
outputs = model(inputs)
if not opt.PCB:
_, preds = torch.max(outputs.data, 1)
loss = criterion(outputs, labels)
else:
part = {}
sm = nn.Softmax(dim=1)
num_part = 6
for i in range(num_part):
part[i] = outputs[i]
score = sm(part[0]) + sm(part[1]) + sm(part[2]) + sm(
part[3]) + sm(part[4]) + sm(part[5])
_, preds = torch.max(score.data, 1)
loss = criterion(part[0], labels)
for i in range(num_part - 1):
loss += criterion(part[i + 1], labels)
# backward + optimize only if in training phase
if phase == 'train':
if fp16: # we use optimier to backward loss
with amp.scale_loss(loss,
optimizer) as scaled_loss:
scaled_loss.backward()
else:
loss.backward()
optimizer.step()
# statistics
if int(version[0]) > 0 or int(
version[2]
) > 3: # for the new version like 0.4.0, 0.5.0 and 1.0.0
running_loss += loss.item() * now_batch_size
else: # for the old version like 0.3.0 and 0.3.1
running_loss += loss.data[0] * now_batch_size
running_corrects += float(torch.sum(preds == labels.data))
pbar.set_description(
desc='loss: {:.4f}'.format(loss.item()))
epoch_loss = running_loss / dataset_sizes[phase]
epoch_acc = running_corrects / dataset_sizes[phase]
print('\r\n{} Loss: {:.4f} Acc: {:.4f}'.format(
phase, epoch_loss, epoch_acc))
logging.info('epoch: {}, {} Loss: {:.4f} Acc: {:.4f}'.format(
epoch, phase, epoch_loss, epoch_acc))
y_loss[phase].append(epoch_loss)
y_err[phase].append(1.0 - epoch_acc)
# deep copy the model
if phase == 'val':
results.append({
'epoch': epoch,
'trainLoss': y_loss['train'][-1],
'trainError': y_err['train'][-1],
'valLoss': y_loss['val'][-1],
'valError': y_err['val'][-1]
})
last_model_wts = model.state_dict()
if epoch % 10 == 9:
save_network(model, epoch)
draw_curve(epoch)
write_to_csv(results)
time_elapsed = time.time() - since
print('\r\nTraining complete in {:.0f}m {:.0f}s'.format(
time_elapsed // 60, time_elapsed % 60))
print()
time_elapsed = time.time() - since
print('\r\nTraining complete in {:.0f}m {:.0f}s'.format(
time_elapsed // 60, time_elapsed % 60))
# print('Best val Acc: {:4f}'.format(best_acc))
# load best model weights
model.load_state_dict(last_model_wts)
save_network(model, 'last')
return model
######################################################################
# Draw Curve
# ---------------------------
x_epoch = []
fig = plt.figure()
ax0 = fig.add_subplot(121, title="loss")
ax1 = fig.add_subplot(122, title="top1err")
def draw_curve(current_epoch):
x_epoch.append(current_epoch)
ax0.plot(x_epoch, y_loss['train'], 'bo-', label='train')
ax0.plot(x_epoch, y_loss['val'], 'ro-', label='val')
ax1.plot(x_epoch, y_err['train'], 'bo-', label='train')
ax1.plot(x_epoch, y_err['val'], 'ro-', label='val')
if current_epoch == 0:
ax0.legend()
ax1.legend()
fig.savefig(os.path.join('./model', name, 'train.jpg'))
def write_to_csv(results):
path = os.path.join('./model', name, 'result.csv')
with open(path, 'w', newline='') as csvfile:
writer = csv.DictWriter(csvfile,
fieldnames=list(results[0].keys()))
writer.writeheader()
writer.writerows(results)
######################################################################
# Save model
# ---------------------------
def save_network(network, epoch_label):
save_filename = 'net_%s.pth' % epoch_label
rpth = os.path.join('./model', name, 'Model Files')
if not os.path.exists(rpth):
os.makedirs(rpth)
save_path = os.path.join(rpth, save_filename)
torch.save(network.cpu().state_dict(), save_path)
if torch.cuda.is_available():
network.cuda(gpu_ids[0])
######################################################################
# Finetuning the convnet
# ----------------------
#
# Load a pretrainied model and reset final fully connected layer.
#
if opt.use_dense:
model = ft_net_dense(len(class_names), opt.droprate)
else:
model = ft_net(len(class_names), opt.droprate, opt.stride)
if opt.PCB:
model = PCB(len(class_names))
opt.nclasses = len(class_names)
print(model)
print('model loaded')
if not opt.PCB:
ignored_params = list(map(id, model.model.fc.parameters())) + list(
map(id, model.classifier.parameters()))
base_params = filter(lambda p: id(p) not in ignored_params,
model.parameters())
optimizer_ft = optim.SGD([{
'params': base_params,
'lr': 0.1 * opt.lr
}, {
'params': model.model.fc.parameters(),
'lr': opt.lr
}, {
'params': model.classifier.parameters(),
'lr': opt.lr
}],
weight_decay=5e-4,
momentum=0.9,
nesterov=True)
else:
ignored_params = list(map(id, model.model.fc.parameters()))
ignored_params += (
list(map(id, model.classifier0.parameters())) +
list(map(id, model.classifier1.parameters())) +
list(map(id, model.classifier2.parameters())) +
list(map(id, model.classifier3.parameters())) +
list(map(id, model.classifier4.parameters())) +
list(map(id, model.classifier5.parameters()))
# +list(map(id, model.classifier6.parameters() ))
# +list(map(id, model.classifier7.parameters() ))
)
base_params = filter(lambda p: id(p) not in ignored_params,
model.parameters())
optimizer_ft = optim.SGD(
[
{
'params': base_params,
'lr': 0.1 * opt.lr
},
{
'params': model.model.fc.parameters(),
'lr': opt.lr
},
{
'params': model.classifier0.parameters(),
'lr': opt.lr
},
{
'params': model.classifier1.parameters(),
'lr': opt.lr
},
{
'params': model.classifier2.parameters(),
'lr': opt.lr
},
{
'params': model.classifier3.parameters(),
'lr': opt.lr
},
{
'params': model.classifier4.parameters(),
'lr': opt.lr
},
{
'params': model.classifier5.parameters(),
'lr': opt.lr
},
# {'params': model.classifier6.parameters(), 'lr': 0.01},
# {'params': model.classifier7.parameters(), 'lr': 0.01}
],
weight_decay=5e-4,
momentum=0.9,
nesterov=True)
# Decay LR by a factor of 0.1 every 40 epochs
exp_lr_scheduler = lr_scheduler.StepLR(optimizer_ft,
step_size=40,
gamma=0.1)
######################################################################
# Train and evaluate
# ^^^^^^^^^^^^^^^^^^
#
# It should take around 1-2 hours on GPU.
#
dir_name = os.path.join('./model', name)
if not os.path.isdir(dir_name):
os.mkdir(dir_name)
# record every run
copyfile('./train.py', dir_name + '/train.py')
copyfile('./model.py', dir_name + '/model.py')
# save opts
with open('%s/opts.yaml' % dir_name, 'w') as fp:
yaml.dump(vars(opt), fp, default_flow_style=False)
# model to gpu
model = model.cuda()
if fp16:
# model = network_to_half(model)
# optimizer_ft = FP16_Optimizer(optimizer_ft, static_loss_scale = 128.0)
model, optimizer_ft = amp.initialize(model,
optimizer_ft,
opt_level="O1")
criterion = losses.DualLoss()
model = train_model(model,
criterion,
optimizer_ft,
exp_lr_scheduler,
num_epochs=60)
#
# if __name__ == "__main__":
# train(opt)
torch.save(network.cpu().state_dict(), save_path)
if torch.cuda.is_available:
network.cuda(gpu_ids[0])
######################################################################
# Finetuning the convnet
# ----------------------
#
# Load a pretrainied model and reset final fully connected layer.
#
if opt.use_dense:
model = ft_net_dense(len(class_names))
else:
model = ft_net(len(class_names))
print(model)
if use_gpu:
model = model.cuda()
criterion = nn.CrossEntropyLoss()
ignored_params = list(map(id, model.model.fc.parameters() )) + list(map(id, model.classifier.parameters() ))
base_params = filter(lambda p: id(p) not in ignored_params, model.parameters())
# Observe that all parameters are being optimized
optimizer_ft = optim.SGD([
{'params': base_params, 'lr': 0.01},
{'params': model.model.fc.parameters(), 'lr': 0.1},
{'params': model.classifier.parameters(), 'lr': 0.1}
def main(first=True):
ImageFile.LOAD_TRUNCATED_IMAGES = True
######################################################################
# Options
# --------
# gpu_ids='0' # use gpu
which_epoch = 'last'
test_dir = 'C:/Users/kdan/BigJob12/main_project/_db/data/model_data'
name = 'ft_ResNet50'
batchsize = 32
ms = '1'
# load the training config
config_path = os.path.join(
'C:/Users/kdan/BigJob12/main_project/_src/data_analysis/re_id/code/model',
name, 'opts.yaml')
with open(config_path, 'r') as stream:
config = yaml.load(stream)
stride = config['stride']
if 'nclasses' in config: # tp compatible with old config files
nclasses = config['nclasses']
######################################################################
# gpu Options
# --------
# str_ids = gpu_ids.split(',')
# gpu_ids = []
# for str_id in str_ids:
# id = int(str_id)
# if id >= 0:
# gpu_ids.append(id)
print('We use the scale: %s' % ms)
str_ms = ms.split(',')
ms = []
for s in str_ms:
s_f = float(s)
ms.append(math.sqrt(s_f))
# set gpu ids
# if len(gpu_ids) > 0:
# torch.cuda.set_device(gpu_ids[0])
# cudnn.benchmark = True
######################################################################
# Data transform
data_transforms = transforms.Compose([
transforms.Resize((256, 128), interpolation=3),
transforms.ToTensor(),
transforms.Normalize([0.485, 0.456, 0.406], [0.229, 0.224, 0.225])
])
######################################################################
# Load data
data_dir = test_dir
image_datasets = {
x: datasets.ImageFolder(os.path.join(data_dir, x), data_transforms)
for x in ['query']
}
dataloaders = {
x: torch.utils.data.DataLoader(image_datasets[x],
batch_size=batchsize,
shuffle=False,
num_workers=0)
for x in ['query']
}
class_names = image_datasets['query'].classes
use_gpu = torch.cuda.is_available()
# Load Collected data Trained model
print('-------test-----------')
model_structure = ft_net(nclasses, stride=stride)
model = load_network(model_structure, name, which_epoch)
# Remove the final fc layer and classifier layer
model.classifier.classifier = nn.Sequential()
print('model')
# Change to test mode
model = model.eval()
# if use_gpu:
# model = model.cuda()
start_load = time.time()
# Extract feature
with torch.no_grad():
query_feature = extract_feature(model, dataloaders['query'], ms)
print(time.time() - start_load)
# Save to Matlab for check
query_result = {'query_f': query_feature.numpy()}
reid_sort.main(query_result, image_datasets, 'all', first)
scipy.io.savemat(
'C:/Users/kdan/BigJob12/main_project/_src/data_analysis/re_id/code/query_result.mat',
query_result)
print(name)
