parser.add_argument('-t', '--testing', action='store_true',
help="Test the trained model on testing dataset")
parser.add_argument('--digit', default=5, type=int,
help="Digit to manipulate")
parser.add_argument('-w', '--weights', default=None,
help="The path of the saved weights. Should be specified when testing")
parser.add_argument('--lr', default=0.001, type=float,
help="Initial learning rate")
parser.add_argument('--gpus', default=4, type=int)
args = parser.parse_args()
print(args)
if not os.path.exists(args.save_dir):
os.makedirs(args.save_dir)
# load data
(x_train, y_train), (x_test, y_test) = load_captcha() if DATA_SOURCE == 'captcha' else load_mnist()
# define model
with tf.device('/cpu:0'):
model, eval_model, manipulate_model = CapsNet(input_shape=x_train.shape[1:],
n_class=len(np.unique(np.argmax(y_train, 1))),
routings=args.routings)
model.summary()
plot_model(model, to_file=args.save_dir+'/model.png', show_shapes=True)
# train or test
if args.weights is not None: # init the model weights with provided one
model.load_weights(args.weights)
if not args.testing:
# define muti-gpu model
multi_model = multi_gpu_model(model, gpus=args.gpus)
'-w',
'--weights',
default=None,
help="The path of the saved weights. Should be specified when testing")
parser.add_argument('--lr',
default=0.001,
type=float,
help="Initial learning rate")
parser.add_argument('--gpus', default=2, type=int)
args = parser.parse_args()
print(args)
if not os.path.exists(args.save_dir):
os.makedirs(args.save_dir)
# load data
(x_train, y_train), (x_test, y_test) = load_mnist()
# define model
with tf.device('/cpu:0'):
model, eval_model, manipulate_model = CapsNet(
input_shape=x_train.shape[1:],
n_class=len(np.unique(np.argmax(y_train, 1))),
routings=args.routings)
model.summary()
plot_model(model, to_file=args.save_dir + '/model.png', show_shapes=True)
# train or test
if args.weights is not None: # init the model weights with provided one
model.load_weights(args.weights)
if not args.testing:
# define muti-gpu model
parser.add_argument('--lam_recon', default=0.392, type=float) # 784 * 0.0005, paper uses sum of SE, here uses MSE
parser.add_argument('--num_routing', default=3, type=int) # num_routing should > 0
parser.add_argument('--shift_fraction', default=0.1, type=float)
parser.add_argument('--debug', default=0, type=int) # debug>0 will save weights by TensorBoard
parser.add_argument('--save_dir', default='./result')
parser.add_argument('--is_training', default=1, type=int)
parser.add_argument('--weights', default=None)
parser.add_argument('--lr', default=0.001, type=float)
parser.add_argument('--gpus', default=2, type=int)
args = parser.parse_args()
print(args)
if not os.path.exists(args.save_dir):
os.makedirs(args.save_dir)
# load data
(x_train, y_train), (x_test, y_test) = load_mnist()
# define model
with tf.device('/cpu:0'):
model, eval_model = CapsNet(input_shape=x_train.shape[1:],
n_class=len(np.unique(np.argmax(y_train, 1))),
num_routing=args.num_routing)
model.summary()
plot_model(model, to_file=args.save_dir+'/model.png', show_shapes=True)
# define muti-gpu model
multi_model = multi_gpu_model(model, gpus=args.gpus)
# train or test
if args.weights is not None: # init the model weights with provided one
model.load_weights(args.weights)
if args.is_training: