def print_options(opt, if_sv=False):
"""Print and save options
It will print both current options and default values(if different).
It will save options into a text file / [checkpoints_dir] / opt.txt
"""
message = ''
message += '----------------- Options ---------------\n'
for k, v in sorted(vars(opt).items()):
comment = ''
# default = self.parser.get_default(k)
# if v != default:
# comment = '\t[default: %s]' % str(default)
message += '{:>25}: {:<30}{}\n'.format(str(k), str(v), comment)
message += '----------------- End -------------------'
print(message)
# save to the disk
if if_sv:
ut.make_folder(
opt.exp_dir
) # all option will mk dirs # saved to json file in set_env
file_name = os.path.join(opt.exp_dir,
'opts.txt'.format(opt.start_epoch)) #
with open(file_name, 'wt') as opt_file:
opt_file.write(message)
opt_file.write('\n')
def main(config):
# For fast training
cudnn.benchmark = True
# Data loader
data_loader = CelebADataset(config.supervise,
config.train,
config.dataset,
config.image_path,
config.imsize,
batchsize=config.batchsize,
shuffle=config.train)
# Create directories if not exist
check_path = os.path.join(config.sample_path, config.version)
while os.path.exists(check_path):
name, version = config.version.split("_")
config.version = f"{name}_{int(version)+1}"
check_path = os.path.join(config.sample_path, config.version)
make_folder(config.model_save_path, config.version)
make_folder(config.sample_path, config.version)
make_folder(config.log_path, config.version)
make_folder(config.attn_path, config.version)
if config.train:
# if config.model == 'sagan':
trainer = Trainer(data_loader.loader(), config)
# elif config.model == 'qgan':
# trainer = qgan_trainer(data_loader.loader(), config)
trainer.train()
else:
raise NotImplementedError
def main(config):
# For fast training
cudnn.benchmark = True
# Data loader
data_loader = Data_Loader(config.train,
config.dataset,
config.image_path,
config.imsize,
config.batch_size,
shuf=config.train)
# Create directories if not exist
make_folder(config.model_save_path, config.version)
make_folder(config.sample_path, config.version)
make_folder(config.log_path, config.version)
make_folder(config.attn_path, config.version)
if config.train:
if config.model == 'sagan':
trainer = Trainer(data_loader.loader(), config)
elif config.model == 'qgan':
trainer = qgan_trainer(data_loader.loader(), config)
trainer.train()
else:
tester = Tester(data_loader.loader(), config)
tester.test()
def set_env(opts): # to be changed accordingly for rst fd
# set sys paths
# sys.path.insert(0, 'common') # not using commong
from utils.utils import add_pypath, make_folder
add_pypath(osp.join('data')) # actually we can use ds directly
for i in range(len(opts.trainset)):
add_pypath(osp.join('data', opts.trainset[i]))
# if opts.cocoapi_dir:
# add_pypath(opts.cocoapi_dir) # add coco dir to it
add_pypath(osp.join('data', opts.testset))
# add folders
make_folder(opts.model_dir)
make_folder(opts.vis_dir)
make_folder(opts.log_dir)
make_folder(opts.rst_dir)
make_folder(opts.web_dir)
# weight softmax x convolution weight and normalize
cam = weight_softmax[class_idx].dot(feature_conv.reshape((nc, h * w)))
cam = cam.reshape(h, w)
cam = cam - np.min(cam)
cam_img = (255 * cam / np.max(cam)).astype(np.uint8)
output_cam = cv2.resize(cam_img, size_upsample)
return output_cam
if __name__ == '__main__':
args = parser.parse_args()
# make folder
today = utils.get_today() + '_cam'
save_dir = Path(args.save_dir) / today
utils.make_folder(save_dir)
# transform
transform = transforms.Compose([
transforms.Resize((128, 128)),
transforms.ToTensor(),
transforms.Normalize((0.5, 0.5, 0.5), (0.5, 0.5, 0.5))
])
#load CIFAR10 DATASETS
testset = torchvision.datasets.CIFAR10(root=args.data_dir,
train=False,
download=True,
transform=transform)
testloader = torch.utils.data.DataLoader(testset,