init = random.sample(range(num), numInjections + offset)
model.load_weights('h5/fcn-trained.h5')
for i in init:
test_loss, test_acc = model.evaluate(test_images[i:i + 1],
test_labels[i:i + 1],
verbose=0)
if (test_acc == 1.):
ind.append(i)
ind = ind[:numInjections]
start = time.time()
for i in range(numFaults):
model.load_weights('h5/fcn-trained.h5')
tfi.inject(model=model, confFile=conf)
sdc = 0.
for i in ind:
test_loss, test_acc = model.evaluate(test_images[i:i + 1],
test_labels[i:i + 1],
verbose=0)
if (test_acc == 0.):
sdc = sdc + 1.
f.write(str(sdc / numInjections))
f.write("\n")
totsdc = totsdc + sdc
f.write("\n")
f.write(str(totsdc / (numFaults * numInjections)))
f.write("\n")
numFaults = int(sys.argv[3])
numInjections = 10
#numInjections = int(sys.argv[4])
totsdc = 0.0
start = time.time()
for i in range(numFaults):
model.load_weights('h5/vgg16-trained.h5')
sdc = 0.
for i in ind:
image = load_img(images[i], target_size=(224, 224))
image = img_to_array(image)
image = image.reshape(
(1, image.shape[0], image.shape[1], image.shape[2]))
image = preprocess_input(image)
res = tfi.inject(model=model, x_test=image, confFile=conf)
label_ = decode_predictions(res[0])
label_ = label_[0][0]
if (label_[1] != img_labels[i]):
sdc = sdc + 1.
f.write(str(sdc / numInjections))
f.write("\n")
totsdc = totsdc + sdc
f.write("\n")
f.write(str(totsdc / (numFaults * numInjections)))
f.write("\n")
f.write("Time for %d injections: %f seconds" %
(numFaults * numInjections, time.time() - start))
f.close()
num = test_images.shape[0]
totsdc = 0.0
ind = []
init = random.sample(range(num), numInjections+offset)
model.load_weights('h5/nn-trained.h5')
for i in init:
test_loss, test_acc = model.evaluate(test_images[i:i+1], test_labels[i:i+1], verbose=0)
if(test_acc == 1.):
ind.append(i)
ind = ind[:numInjections]
start = time.time()
for i in range(numFaults):
model.load_weights('h5/nn-trained.h5')
sdc = 0.
for i in ind:
res = tfi.inject(model=model, x_test=test_images[i:i+1], confFile=conf)
if(res != test_labels[i:i+1]):
sdc = sdc + 1.
f.write(str(sdc/numInjections))
f.write("\n")
totsdc = totsdc + sdc
f.write("\n")
f.write(str(totsdc/(numFaults*numInjections)))
f.write("\n")
f.write("Time for %d injections: %f seconds" % (numFaults*numInjections, time.time() - start))
f.close()
test_ = test_ds.take(2)
test_loss, test_acc = model.evaluate(test_, verbose=0)
if(test_acc == 1.):
break
for i in range(20):
test_i = test_ds.take(1)
test_loss, test_acc = model.evaluate(test_i, verbose=0)
if(test_acc == 1.):
test_ = test_.concatenate(test_i)
if(tf.data.experimental.cardinality(test_).numpy() == 10):
break
start = time.time()
print(start)
for i in range(numFaults):
model.load_weights('h5/alexnet-trained.h5')
sdc = 0.
for elem, label in test_.as_numpy_iterator():
res = tfi.inject(model=model, x_test=elem, confFile=conf)
if(res != label):
sdc = sdc + 1.
f.write(str(sdc/numInjections))
f.write("\n")
totsdc = totsdc + sdc
f.write("\n")
f.write(str(totsdc/(numFaults*numInjections)))
f.write("\n")
f.write("Time for %d injections: %f seconds" % (numFaults*numInjections, time.time() - start))
f.close()