def test_binary_blocks_crf():
X, Y = toy.generate_blocks(n_samples=1)
x, y = X[0], Y[0]
w = np.array([1, 1, 0, -4, 0])
crf = GridCRF()
y_hat = crf.inference(x, w)
assert_array_equal(y, y_hat)
def test_continuous_y():
# for inference_method in ["lp", "ad3"]:
for inference_method in ["lp"]:
X, Y = toy.generate_blocks(n_samples=1)
x, y = X[0], Y[0]
w = np.array([1, 1, 0, -4, 0])
crf = GridCRF(inference_method=inference_method)
psi = crf.psi(x, y)
y_cont = np.zeros_like(x)
gx, gy = np.indices(x.shape[:-1])
y_cont[gx, gy, y] = 1
# need to generate edge marginals
vert = np.dot(y_cont[1:, :, :].reshape(-1, 2).T, y_cont[:-1, :, :].reshape(-1, 2))
# horizontal edges
horz = np.dot(y_cont[:, 1:, :].reshape(-1, 2).T, y_cont[:, :-1, :].reshape(-1, 2))
pw = vert + horz
psi_cont = crf.psi(x, (y_cont, pw))
assert_array_almost_equal(psi, psi_cont)
const = find_constraint(crf, x, y, w, relaxed=False)
const_cont = find_constraint(crf, x, y, w, relaxed=True)
# dpsi and loss are equal:
assert_array_almost_equal(const[1], const_cont[1])
assert_almost_equal(const[2], const_cont[2])
# returned y_hat is one-hot version of other
assert_array_equal(const[0], np.argmax(const_cont[0][0], axis=-1))
# test loss:
assert_equal(crf.loss(y, const[0]), crf.continuous_loss(y, const_cont[0][0]))
def test_binary_grid_unaries():
# test handling on unaries for binary grid CRFs
for ds in toy.binary:
X, Y = ds(n_samples=1)
x, y = X[0], Y[0]
crf = GridCRF(inference_method="lp")
w_unaries_only = np.zeros(5)
w_unaries_only[:2] = 1.0
# test that inference with unaries only is the
# same as argmax
inf_unaries = crf.inference(x, w_unaries_only)
pw_z = np.zeros((2, 2), dtype=np.int32)
un = np.ascontiguousarray(-1000 * x).astype(np.int32)
unaries = binary_grid(un, pw_z)
assert_array_equal(inf_unaries, unaries)
assert_array_equal(inf_unaries, np.argmax(x, axis=2))
try:
assert np.mean(inf_unaries == y) > 0.5
except:
print(ds)
# check that the right thing happens on noise-free data
X, Y = ds(n_samples=1, noise=0)
inf_unaries = crf.inference(X[0], w_unaries_only)
assert_array_equal(inf_unaries, Y[0])
def test_blocks_multinomial_crf():
X, Y = toy.generate_blocks_multinomial(n_samples=1)
x, y = X[0], Y[0]
w = np.array([1.0, 1.0, 1.0, 0.4, -0.3, 0.3, -0.5, -0.1, 0.3])
crf = GridCRF(n_states=3)
y_hat = crf.inference(x, w)
assert_array_equal(y, y_hat)
def test_binary_blocks_crf_n8_lp():
X, Y = toy.generate_blocks(n_samples=1, noise=1)
x, y = X[0], Y[0]
w = np.array([1, 1, 1, -1.4, 1])
crf = GridCRF(inference_method="lp", neighborhood=8)
y_hat = crf.inference(x, w)
assert_array_equal(y, y_hat)
def test_binary_crf_exhaustive():
# tests graph cut inference against brute force
# on random data / weights
np.random.seed(0)
for i in xrange(50):
x = np.random.uniform(-1, 1, size=(3, 3))
x = np.dstack([-x, np.zeros_like(x)]).copy()
crf = GridCRF()
w = np.random.uniform(-1, 1, size=5)
# check map inference
y_hat = crf.inference(x, w)
y_ex = exhausive_inference_binary(crf, x, w)
# print(y_hat)
# print(y_ex)
# print("++++++++++++++++++++++")
assert_array_equal(y_hat, y_ex)
def test_binary_crf_exhaustive_loss_augmented():
# tests graph cut inference against brute force
# on random data / weights
np.random.seed(0)
for i in xrange(50):
# generate data and weights
y = np.random.randint(2, size=(3, 3))
x = np.random.uniform(-1, 1, size=(3, 3))
x = np.dstack([-x, np.zeros_like(x)])
w = np.random.uniform(-1, 1, size=5)
crf = GridCRF()
# check loss augmented map inference
y_hat = crf.loss_augmented_inference(x, y, w)
y_ex = exhausive_loss_augmented_inference_binary(crf, x, y, w)
# print(y_hat)
# print(y_ex)
# print("++++++++++++++++++++++")
assert_array_equal(y_hat, y_ex)
def test_binary_ssvm_repellent_potentials():
# test non-submodular learning with and without positivity constraint
# dataset is checkerboard
X, Y = toy.generate_checker()
crf = GridCRF()
clf = StructuredSVM(problem=crf, max_iter=200, C=100,
verbose=0, check_constraints=True)
clf.fit(X, Y)
Y_pred = clf.predict(X)
# standard crf can predict perfectly
assert_array_equal(Y, Y_pred)
submodular_clf = StructuredSVM(problem=crf, max_iter=200, C=100,
verbose=0, check_constraints=True,
positive_constraint=[2, 3, 4])
submodular_clf.fit(X, Y)
Y_pred = submodular_clf.predict(X)
# submodular crf can not do better than unaries
for i, x in enumerate(X):
y_pred_unaries = crf.inference(x, np.array([1, 1, 0, 0, 0]))
assert_array_equal(y_pred_unaries, Y_pred[i])
def test_loss_augmentation():
X, Y = toy.generate_blocks(n_samples=1)
x, y = X[0], Y[0]
w = np.array([1.0, 1.0, 0.0, -4.0, 0.0])
unary_params = w[:2]
pairwise_flat = np.asarray(w[2:])
pairwise_params = np.zeros((2, 2))
pairwise_params[np.tri(2, dtype=np.bool)] = pairwise_flat
pairwise_params = pairwise_params + pairwise_params.T - np.diag(np.diag(pairwise_params))
crf = GridCRF()
x_loss_augmented = crf.loss_augment(x, y, w)
y_hat = crf.inference(x_loss_augmented, w)
# test that loss_augmented_inference does the same
y_hat_2 = crf.loss_augmented_inference(x, y, w)
assert_array_equal(y_hat_2, y_hat)
energy = compute_energy(x, y_hat, unary_params, pairwise_params)
energy_loss_augmented = compute_energy(x_loss_augmented, y_hat, unary_params, pairwise_params)
assert_almost_equal(energy + crf.loss(y, y_hat), energy_loss_augmented)
# with zero in w:
unary_params[1] = 0
assert_raises(ValueError, crf.loss_augment, x, y, w)
def test_multinomial_grid_unaries():
# test handling on unaries for multinomial grid CRFs
# on multinomial datasets
for ds in toy.multinomial:
X, Y = ds(n_samples=1)
x, y = X[0], Y[0]
n_labels = len(np.unique(Y))
crf = GridCRF(n_states=n_labels)
w_unaries_only = np.zeros(crf.size_psi)
w_unaries_only[:n_labels] = 1.0
# test that inference with unaries only is the
# same as argmax
inf_unaries = crf.inference(x, w_unaries_only)
pw_z = np.zeros((n_labels, n_labels), dtype=np.int32)
un = np.ascontiguousarray(-1000 * x).astype(np.int32)
unaries = alpha_expansion_grid(un, pw_z)
assert_array_equal(inf_unaries, unaries)
assert_array_equal(inf_unaries, np.argmax(x, axis=2))
# check that the right thing happens on noise-free data
X, Y = ds(n_samples=1, noise=0)
inf_unaries = crf.inference(X[0], w_unaries_only)
assert_array_equal(inf_unaries, Y[0])
