continue
for k in xrange(dets.shape[0]):
f.write('{:f} {:f} {:f} {:f} {:.32f}\n'.format(\
dets[k, 0], dets[k, 1], dets[k, 2], dets[k, 3], dets[k, 4]))
def evaluate_proposals_msr(self, all_boxes, output_dir):
# for each image
for im_ind, index in enumerate(self.image_index):
filename = os.path.join(output_dir, index + '.txt')
print 'Writing PASCAL results to file ' + filename
with open(filename, 'wt') as f:
dets = all_boxes[im_ind]
if dets == []:
continue
for k in xrange(dets.shape[0]):
f.write('{:f} {:f} {:f} {:f} {:.32f}\n'.format(dets[k, 0], dets[k, 1], dets[k, 2], dets[k, 3], dets[k, 4]))
def competition_mode(self, on):
if on:
self.config['use_salt'] = False
self.config['cleanup'] = False
else:
self.config['use_salt'] = True
self.config['cleanup'] = True
if __name__ == '__main__':
d = datasets.pascal3d('train')
res = d.roidb
from IPython import embed; embed()
dets[k, 0], dets[k, 1], dets[k, 2], dets[k, 3], dets[k, 4]))
def evaluate_proposals_msr(self, all_boxes, output_dir):
# for each image
for im_ind, index in enumerate(self.image_index):
filename = os.path.join(output_dir, index + '.txt')
print 'Writing PASCAL results to file ' + filename
with open(filename, 'wt') as f:
dets = all_boxes[im_ind]
if dets == []:
continue
for k in xrange(dets.shape[0]):
f.write('{:f} {:f} {:f} {:f} {:.32f}\n'.format(
dets[k, 0], dets[k, 1], dets[k, 2], dets[k, 3],
dets[k, 4]))
def competition_mode(self, on):
if on:
self.config['use_salt'] = False
self.config['cleanup'] = False
else:
self.config['use_salt'] = True
self.config['cleanup'] = True
if __name__ == '__main__':
d = datasets.pascal3d('train')
res = d.roidb
from IPython import embed
embed()
def get_data_loaders(dataset, batch_size, num_workers, model, num_classes=12):
image_size = 227
train_transform = transforms.Compose([transforms.ToTensor(),
transforms.Normalize(mean=(0., 0., 0.),
std=(1. / 255., 1. / 255., 1. / 255.)
),
transforms.Normalize(mean=(104, 116.668, 122.678),
std=(1., 1., 1.)
)
])
test_transform = transforms.Compose([transforms.ToTensor(),
transforms.Normalize(mean=(0., 0., 0.),
std=(1. / 255., 1. / 255., 1. / 255.)
),
transforms.Normalize(mean=(104, 116.668, 122.678),
std=(1., 1., 1.)
)
])
# # The New transform for ImageNet Stuff
# new_transform = transforms.Compose([
# transforms.ToTensor(),
# transforms.Normalize(mean=(0.485, 0.456, 0.406),
# std=(0.229, 0.224, 0.225))])
if dataset == "pascal":
csv_train = os.path.join(root_dir, 'data/pascal3d_train.csv')
csv_test = os.path.join(root_dir, 'data/pascal3d_valid.csv')
# 生成训练数据集，测试数据集
train_set = pascal3d(csv_train, dataset_root=dataset_root, transform=train_transform, im_size=image_size)
test_set = pascal3d(csv_test, dataset_root=dataset_root, transform=test_transform, im_size=image_size)
elif dataset == "pascalEasy":
csv_train = os.path.join(root_dir, 'data/pascal3d_train_easy.csv')
csv_test = os.path.join(root_dir, 'data/pascal3d_valid_easy.csv')
train_set = pascal3d(csv_train, dataset_root=dataset_root,
transform=train_transform, im_size=image_size)
test_set = pascal3d(csv_test, dataset_root=dataset_root,
transform=test_transform, im_size=image_size)
elif dataset == "pascalFull":
csv_train = os.path.join(root_dir, 'data/train.csv')
csv_test = os.path.join(root_dir, 'data/val.csv')
train_set = pascal3d(csv_train, dataset_root=dataset_root, transform=train_transform, im_size=image_size)
test_set = pascal3d(csv_test, dataset_root=dataset_root, transform=test_transform, im_size=image_size)
else:
print("Error in load_datasets: Dataset name not defined.")
# Generate data loaders
train_loader = torch.utils.data.DataLoader(dataset=train_set,
batch_size=batch_size,
shuffle=True,
num_workers=num_workers,
drop_last=True)
test_loader = torch.utils.data.DataLoader(dataset=test_set,
batch_size=batch_size,
shuffle=False,
num_workers=num_workers,
drop_last=False)
return train_loader, test_loader
