def main():
X_train, Y_train, X_test, Y_test, Y_test_uncat = data_setup()
# Create data shards
random_indices = np.random.choice(
len(X_train), len(X_train), replace=False)
X_train_permuted = X_train[random_indices]
Y_train_permuted = Y_train[random_indices]
X_train_shards = np.split(X_train_permuted, args.k)
Y_train_shards = np.split(Y_train_permuted, args.k)
if args.mal:
# Load malicious data
mal_data_X, mal_data_Y, true_labels = mal_data_setup(X_test, Y_test, Y_test_uncat)
if args.train:
p = Process(target=master)
p.start()
p.join()
manager = Manager()
return_dict = manager.dict()
return_dict['eval_success'] = 0.0
return_dict['eval_loss'] = 0.0
if args.mal:
return_dict['mal_suc_count'] = 0
t_final = train_fn(X_train_shards, Y_train_shards, X_test, Y_test_uncat,
return_dict, mal_data_X, mal_data_Y)
print('Malicious agent succeeded in %s of %s iterations' %
(return_dict['mal_suc_count'], t_final * args.mal_num))
else:
_ = train_fn(X_train_shards, Y_train_shards, X_test, Y_test_uncat,
return_dict)
else:
manager = Manager()
return_dict = manager.dict()
return_dict['eval_success'] = 0.0
return_dict['eval_loss'] = 0.0
if args.mal:
return_dict['mal_suc_count'] = 0
for t in range(args.T):
if not os.path.exists(gv.dir_name + 'global_weights_t%s.npy' % t):
print('No directory found for iteration %s' % t)
break
if args.mal:
p_eval = Process(target=eval_func, args=(
X_test, Y_test_uncat, t, return_dict, mal_data_X, mal_data_Y))
else:
p_eval = Process(target=eval_func, args=(
X_test, Y_test_uncat, t, return_dict))
p_eval.start()
p_eval.join()
if args.mal:
print('Malicious agent succeeded in %s of %s iterations' %
(return_dict['mal_suc_count'], (t-1) * args.mal_num))
K.set_learning_phase(0)
gv.init()
args = gv.args
weights_np = np.load(gv.dir_name + 'global_weights_t%s.npy' % 8)
X_train, Y_train, X_test, Y_test, Y_test_uncat = data_setup()
mal_analyse = True
if mal_analyse:
mal_data_X, mal_data_Y, true_labels = mal_data_setup(X_test,
Y_test,
Y_test_uncat,
gen_flag=False)
label_to_class_name = [str(i) for i in range(gv.NUM_CLASSES)]
if 'MNIST' in args.dataset:
model = model_mnist(type=args.model_num)
elif args.dataset == 'CIFAR-10':
model = cifar_10_model()
x = tf.compat.v1.placeholder(shape=(None, gv.IMAGE_ROWS, gv.IMAGE_COLS,
gv.NUM_CHANNELS),
dtype=tf.float32)
y = tf.compat.v1.placeholder(dtype=tf.int64)
logits = model(x)