def plot_gradient(V, loc):
g = grad(V)
gx = generate(g.shape[0], lambda i, j: g[i][j][0])
gy = generate(g.shape[0], lambda i, j: g[i][j][1])
fig = plt.imshow(V, interpolation='none')
plt.savefig(loc + "-original.png")
fig = plt.imshow(gx, interpolation='none')
plt.savefig(loc + "-ddx.png")
fig = plt.imshow(gy, interpolation='none')
plt.savefig(loc + "-ddy.png")
def test_more_complex(self):
v = np.array([[1, 2], [0, 3]])
vx = np.array([[[-1, 1], [1, 1]], [[-1, 3], [1, 3]]])
assertArrayEq(self, vx, grad(v))
def test_basic(self):
v = np.array([[1, 2], [1, 2]])
vx = np.array([[[0, 1], [0, 1]], [[0, 1], [0, 1]]])
assertArrayEq(self, vx, grad(v))
def test_constant_function(self):
x = np.array([[1, 1], [1, 1]])
assertArrayEq(self, np.zeros((2, 2, 2)), grad(x))
x = np.array([[1, 1, 1, 1, 1], [1, 1, 1, 1, 1], [1, 1, 1, 1, 1],
[1, 1, 1, 1, 1], [1, 1, 1, 1, 1]])
assertArrayEq(self, np.zeros((5, 5, 2)), grad(x))
def test_more_complex(self):
v = np.array([[1,2],[0,3]])
vx = np.array([[[-1,1], [1,1]], [[-1,3], [1,3]]])
assertArrayEq(self, vx, grad(v))
def test_basic(self):
v = np.array([[1,2],[1,2]])
vx = np.array([[[0,1], [0,1]], [[0,1], [0,1]]])
assertArrayEq(self, vx, grad(v))
def test_constant_function(self):
x = np.array([[1,1],[1,1]])
assertArrayEq(self, np.zeros((2,2,2)), grad(x))
x = np.array([[1,1,1,1,1],[1,1,1,1,1],[1,1,1,1,1],[1,1,1,1,1],[1,1,1,1,1]])
assertArrayEq(self, np.zeros((5,5,2)), grad(x))
