def load_model(model_name, activation='relu', input_shape=200):
if model_name == 'cnn-50':
from model_zoo import cnn_model
model = cnn_model.model([50, 2], activation, input_shape)
elif model_name == 'cnn-2':
from model_zoo import cnn_model
model = cnn_model.model([2, 50], activation, input_shape)
elif model_name == 'cnn-deep':
from model_zoo import cnn_deep
model = cnn_deep.model(activation, input_shape)
elif model_name == 'cnn-local':
from model_zoo import cnn_local
model = cnn_local.model(activation)
elif model_name == 'cnn-dist':
from model_zoo import cnn_dist
model = cnn_dist.model(activation)
elif model_name == 'basset':
from model_zoo import basset
model = basset.model(activation)
elif model_name == 'residualbind':
from model_zoo import residualbind
model = residualbind.model(activation)
return model
results_path = helper.make_directory('../results', args.o)
#-----------------------------------------------------------------
# load data
data_path = '../data'
filepath = os.path.join(data_path, 'synthetic_code_dataset.h5')
x_train, y_train, x_valid, y_valid, x_test, y_test, model_test = helper.load_data(
filepath)
N, L, A = x_train.shape
num_labels = y_train.shape[1]
#-----------------------------------------------------------------
# create model
model = genome_model.model(input_shape=(L, A),
num_labels=1,
activation=activation,
dropout=dropout,
bn=bn,
l2=None)
loss = keras.losses.BinaryCrossentropy(from_logits=False, label_smoothing=0.0)
optimizer = keras.optimizers.Adam(learning_rate=0.001)
#-----------------------------------------------------------------
# Fit model
attacker = tfomics.attack.PGDAttack((batch_size, L, A),
model,
loss,
learning_rate=0.01,
epsilon=epsilon,
num_steps=num_steps)
history, trainer = tfomics.fit.fit_robust(model,
loss,