biases3 = np.genfromtxt('../mnist/para/biases3.csv', delimiter=',')
B1 = [ RealVal(biases1[i]) for i in range(l1_n) ]
B2 = [ RealVal(biases2[i]) for i in range(l2_n) ]
B3 = [ RealVal(biases3[i]) for i in range(l3_n) ]
print "Creating Assertions"
if verify_mode == "specific":
# The MNIST data set has 10,000 images for testing
inputs = np.genfromtxt('../mnist/para/mnist_test_images_100.csv', delimiter=',')
InX = [ RealVal(inputs[0][i]) for i in range(l0_n) ]
else:
InX = [ Real('inX-%s' % i) for i in range(l0_n) ]
InY= [ Real('inY-%s' % i) for i in range(l0_n) ]
L1X = core.vmmul(InX, W1, B1, l0_n, l1_n, act_func)
L2X = core.vmmul(L1X, W2, B2, l1_n, l2_n, act_func)
OutX = core.vmmul(L2X, W3, B3, l2_n, l3_n, act_func)
L1Y = core.vmmul(InY, W1, B1, l0_n, l1_n, act_func)
L2Y = core.vmmul(L1Y, W2, B2, l1_n, l2_n, act_func)
OutY = core.vmmul(L2Y, W3, B3, l2_n, l3_n, act_func)
#TODO: The input pertubation constriants have to be more general
input_cond = [ And(InX[i] - InY[i] < input_bound, InY[i] - InX[i] < input_bound, 0 <= InY[i], InY[i] <= 1, 0 <= InX[i], InX[i] <= 1) for i in range(l0_n) ]
output_cond = [ Not( core.full_robust(OutX, OutY, l3_n, robust_cons, bound=output_bound) ) ]
s = Solver()
s.add(input_cond +
output_cond)
else:
InX = [ Real('inX-%s' % i) for i in range(l0_n) ]
InY= [ Real('inY-%s' % i) for i in range(l0_n) ]
else: #unscaled
scale = RealVal(255)
if verify_mode == "specific":
inputs = np.genfromtxt('../mnist/para/mnist_test_images_unscaled_100.csv', delimiter=',')
X = [ IntVal(inputs[input_id][i]) for i in range(l0_n) ]
InX = core.scaleInput(X, scale, l0_n)
else:
X = [ Int('x%s' % i) for i in range(l0_n) ]
InX = core.scaleInput(X, scale, l0_n)
Y = [ Int('y%s' % i) for i in range(l0_n) ]
InY = core.scaleInput(Y, scale, l0_n)
OutX = core.vmmul(InX, W, B, l0_n, l1_n, "none")
OutY = core.vmmul(InY, W, B, l0_n, l1_n, "none")
if input_scaled == "scaled":
input_cond = [ And(InX[i] - InY[i] < input_var, InY[i] - InX[i] < input_var, 0 <= InY[i], InY[i] <= 1, 0 <= InX[i], InX[i] <= 1) for i in range(l0_n) ]
else:
input_cond = [ And(X[i] - Y[i] < input_var, Y[i] - X[i] < input_var, 0 <= Y[i], Y[i] <= 255, 0 <= X[i], X[i] <= 255) for i in range(l0_n) ]
output_cond = [ Not( core.full_robust(OutX, OutY, l1_n, robust_cons, bound=output_bound) ) ]
s = Solver()
s.add(input_cond +
output_cond)
result = core.solveIt(s)
if (result == sat):
m = s.model()
outx = [m.evaluate(OutX[i]) for i in range(l1_n)]
outy = [m.evaluate(OutY[i]) for i in range(l1_n)]