mirrors
/
deep-person-reid
mirror of https://github.com/KaiyangZhou/deep-person-reid.git
1
0
Fork 0
Code Issues Projects Releases Wiki Activity

market_test

pull/470/head
Huy Huynh 2021-10-24 11:24:01 -05:00
parent be4383b3ca
commit 95ec05290a
5 changed files with 90 additions and 88 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

12
scripts/main_test.py
View File

@ -6,8 +6,8 @@ if __name__ == '__main__':
datamanager = torchreid.data.ImageDataManager(
root='reid-data',
sources='newdataset',
# targets='newdataset',
sources='market1501',
targets='market1501',
height=256,
width=128,
transforms=['random_flip', 'color_jitter']
@ -15,7 +15,7 @@ if __name__ == '__main__':
model = torchreid.models.build_model(
name='osnet_x1_0',
num_classes=0,
num_classes=datamanager._num_train_pids,
# loss='softmax',
pretrained=True
)
@ -39,12 +39,12 @@ if __name__ == '__main__':
)
engine.run(
save_dir='log/newdataset_visrank',
save_dir='log/market_visrank',
test_only=True,
# max_epoch=60,
# eval_freq=10,
# print_freq=10,
visrank=True,
visrank_topk=20,
# visrank=True,
# visrank_topk=20,
# rerank=True
)

152
torchreid/data/datamanager.py
View File

@ -189,84 +189,84 @@ class ImageDataManager(DataManager):
use_gpu=use_gpu
)
# print('=> Loading train (source) dataset')
# trainset = []
# for name in self.sources:
# trainset_ = init_image_dataset(
# name,
# transform=self.transform_tr,
# k_tfm=k_tfm,
# mode='train',
# combineall=combineall,
# root=root,
# split_id=split_id,
# cuhk03_labeled=cuhk03_labeled,
# cuhk03_classic_split=cuhk03_classic_split,
# market1501_500k=market1501_500k
# )
# trainset.append(trainset_)
# trainset = sum(trainset)
print('=> Loading train (source) dataset')
trainset = []
for name in self.sources:
trainset_ = init_image_dataset(
name,
transform=self.transform_tr,
k_tfm=k_tfm,
mode='train',
combineall=combineall,
root=root,
split_id=split_id,
cuhk03_labeled=cuhk03_labeled,
cuhk03_classic_split=cuhk03_classic_split,
market1501_500k=market1501_500k
)
trainset.append(trainset_)
trainset = sum(trainset)
# self._num_train_pids = trainset.num_train_pids
# self._num_train_cams = trainset.num_train_cams
self._num_train_pids = trainset.num_train_pids
self._num_train_cams = trainset.num_train_cams
# self.train_loader = torch.utils.data.DataLoader(
# trainset,
# sampler=build_train_sampler(
# trainset.train,
# train_sampler,
# batch_size=batch_size_train,
# num_instances=num_instances,
# num_cams=num_cams,
# num_datasets=num_datasets
# ),
# batch_size=batch_size_train,
# shuffle=False,
# num_workers=workers,
# pin_memory=self.use_gpu,
# drop_last=True
# )
self.train_loader = torch.utils.data.DataLoader(
trainset,
sampler=build_train_sampler(
trainset.train,
train_sampler,
batch_size=batch_size_train,
num_instances=num_instances,
num_cams=num_cams,
num_datasets=num_datasets
),
batch_size=batch_size_train,
shuffle=False,
num_workers=workers,
pin_memory=self.use_gpu,
drop_last=True
)
# self.train_loader_t = None
# if load_train_targets:
# # check if sources and targets are identical
# assert len(set(self.sources) & set(self.targets)) == 0, \
# 'sources={} and targets={} must not have overlap'.format(self.sources, self.targets)
self.train_loader_t = None
if load_train_targets:
# check if sources and targets are identical
assert len(set(self.sources) & set(self.targets)) == 0, \
'sources={} and targets={} must not have overlap'.format(self.sources, self.targets)
# print('=> Loading train (target) dataset')
# trainset_t = []
# for name in self.targets:
# trainset_t_ = init_image_dataset(
# name,
# transform=self.transform_tr,
# k_tfm=k_tfm,
# mode='train',
# combineall=False, # only use the training data
# root=root,
# split_id=split_id,
# cuhk03_labeled=cuhk03_labeled,
# cuhk03_classic_split=cuhk03_classic_split,
# market1501_500k=market1501_500k
# )
# trainset_t.append(trainset_t_)
# trainset_t = sum(trainset_t)
print('=> Loading train (target) dataset')
trainset_t = []
for name in self.targets:
trainset_t_ = init_image_dataset(
name,
transform=self.transform_tr,
k_tfm=k_tfm,
mode='train',
combineall=False, # only use the training data
root=root,
split_id=split_id,
cuhk03_labeled=cuhk03_labeled,
cuhk03_classic_split=cuhk03_classic_split,
market1501_500k=market1501_500k
)
trainset_t.append(trainset_t_)
trainset_t = sum(trainset_t)
# self.train_loader_t = torch.utils.data.DataLoader(
# trainset_t,
# sampler=build_train_sampler(
# trainset_t.train,
# train_sampler_t,
# batch_size=batch_size_train,
# num_instances=num_instances,
# num_cams=num_cams,
# num_datasets=num_datasets
# ),
# batch_size=batch_size_train,
# shuffle=False,
# num_workers=workers,
# pin_memory=self.use_gpu,
# drop_last=True
# )
self.train_loader_t = torch.utils.data.DataLoader(
trainset_t,
sampler=build_train_sampler(
trainset_t.train,
train_sampler_t,
batch_size=batch_size_train,
num_instances=num_instances,
num_cams=num_cams,
num_datasets=num_datasets
),
batch_size=batch_size_train,
shuffle=False,
num_workers=workers,
pin_memory=self.use_gpu,
drop_last=True
)
print('=> Loading test (target) dataset')
self.test_loader = {
@ -335,9 +335,9 @@ class ImageDataManager(DataManager):
print(' **************** Summary ****************')
print(' source : {}'.format(self.sources))
print(' # source datasets : {}'.format(len(self.sources)))
# print(' # source ids : {}'.format(self.num_train_pids))
# print(' # source images : {}'.format(len(trainset)))
# print(' # source cameras : {}'.format(self.num_train_cams))
print(' # source ids : {}'.format(self.num_train_pids))
print(' # source images : {}'.format(len(trainset)))
print(' # source cameras : {}'.format(self.num_train_cams))
if load_train_targets:
print(
' # target images : {} (unlabeled)'.format(len(trainset_t))

4
torchreid/data/datasets/dataset.py
View File

@ -54,8 +54,8 @@ class Dataset(object):
# extend 3-tuple (img_path(s), pid, camid) to
# 4-tuple (img_path(s), pid, camid, dsetid) by
# adding a dataset indicator "dsetid"
# if len(train[0]) == 3:
# train = [(*items, 0) for items in train]
if len(train[0]) == 3:
train = [(*items, 0) for items in train]
if len(query[0]) == 3:
query = [(*items, 0) for items in query]
if len(gallery[0]) == 3:

6
torchreid/engine/engine.py
View File

@ -27,7 +27,7 @@ class Engine(object):
def __init__(self, datamanager, use_gpu=True):
self.datamanager = datamanager
# self.train_loader = self.datamanager.train_loader
self.train_loader = self.datamanager.train_loader
self.test_loader = self.datamanager.test_loader
self.use_gpu = (torch.cuda.is_available() and use_gpu)
self.writer = None
@ -394,10 +394,10 @@ class Engine(object):
print(
'Computing distance matrix with metric={} ...'.format(dist_metric)
)
print(len(qf[0]))
# print(len(qf[0]))
distmat = metrics.compute_distance_matrix(qf, gf, dist_metric)
distmat = distmat.numpy()
print(distmat)
# print(distmat)
if rerank:
print('Applying person re-ranking ...')
distmat_qq = metrics.compute_distance_matrix(qf, qf, dist_metric)

4
torchreid/engine/image/softmax.py
View File

@ -70,7 +70,9 @@ class ImageSoftmaxEngine(Engine):
self.register_model('model', model, optimizer, scheduler)
self.criterion = CrossEntropyLoss(
num_classes=0, use_gpu=self.use_gpu, label_smooth=label_smooth
num_classes=self.datamanager.num_train_pids,
use_gpu=self.use_gpu,
label_smooth=label_smooth
)
def forward_backward(self, data):