'alexnet': 224,
'densenet': 224,
'resnet': 224,
'inception': 299,
'squeezenet': 224,
'vgg': 224
}
if args.mode == "train":
flickr = load_flickr(args.img_dir, args.annot_path)
data = DataHandler(flickr)
if os.path.isfile('./data/process.pkl.gz'):
data.load_data('./data/process.pkl.gz')
else:
data.build_vocab()
data.resize_images('./data/resized', img_size=input_size['resnet'])
data.pad_data()
if not os.path.isfile('./data/process.pkl.gz'):
data.save_data('./data/process.pkl.gz')
transform = transforms.Compose([
transforms.RandomHorizontalFlip(),
transforms.ToTensor(),
transforms.Normalize((0.485, 0.456, 0.406), (0.229, 0.224, 0.225))
])
data.set_augment(transform)
