def test_input_args_and_kwargs():
rng = np.random.RandomState(42)
p = 125
noise_std = 1e-1
sig = np.zeros(p)
sig[[0, 2, 13, 4, 25, 32, 80, 89, 91, 93, -1]] = 1
sig[:6] = 2
sig[-7:] = 2
sig[60:75] = 1
y = sig + noise_std * rng.randn(*sig.shape)
X = np.eye(p)
mask = np.ones((p,)).astype(np.bool)
alpha = .01
alpha_ = alpha * X.shape[0]
l1_ratio = .2
l1_weight = alpha_ * l1_ratio
f1 = lambda w: _squared_loss(X, y, w, compute_grad=False)
f1_grad = lambda w: _squared_loss(X, y, w, compute_grad=True,
compute_energy=False)
f2_prox = lambda w, l, *args, **kwargs: (_prox_l1(w, l * l1_weight),
dict(converged=True))
total_energy = lambda w: f1(w) + l1_weight * np.sum(np.abs(w))
for cb_retval in [0, 1]:
for verbose in [0, 1]:
for dgap_factor in [1., None]:
best_w, objective, init = mfista(
f1_grad, f2_prox, total_energy, 1., p,
dgap_factor=dgap_factor,
callback=lambda _: cb_retval, verbose=verbose,
max_iter=100)
assert_equal(best_w.shape, mask.shape)
assert_true(isinstance(objective, list))
assert_true(isinstance(init, dict))
for key in ["w", "t", "dgap_tol", "stepsize"]:
assert_true(key in init)
def test_same_energy_calculus_pure_lasso():
rng = check_random_state(42)
X, y, w, mask = _make_data(rng=rng, masked=True)
# check funcvals
f1 = _squared_loss(X, y, w)
f2 = _squared_loss_and_spatial_grad(X, y, w.ravel(), mask, 0.)
assert f1 == f2
# check derivatives
g1 = _squared_loss_grad(X, y, w)
g2 = _squared_loss_and_spatial_grad_derivative(X, y, w.ravel(), mask, 0.)
np.testing.assert_array_equal(g1, g2)
def test_same_energy_calculus_pure_lasso():
rng = check_random_state(42)
X, y, w, mask = _make_data(rng=rng, masked=True)
# check funcvals
f1 = _squared_loss(X, y, w)
f2 = _squared_loss_and_spatial_grad(X, y, w.ravel(), mask, 0.)
assert_equal(f1, f2)
# check derivatives
g1 = _squared_loss_grad(X, y, w)
g2 = _squared_loss_and_spatial_grad_derivative(X, y, w.ravel(), mask, 0.)
np.testing.assert_array_equal(g1, g2)
def test_input_args_and_kwargs():
rng = np.random.RandomState(42)
p = 125
noise_std = 1e-1
sig = np.zeros(p)
sig[[0, 2, 13, 4, 25, 32, 80, 89, 91, 93, -1]] = 1
sig[:6] = 2
sig[-7:] = 2
sig[60:75] = 1
y = sig + noise_std * rng.randn(*sig.shape)
X = np.eye(p)
mask = np.ones((p, )).astype(np.bool)
alpha = .01
alpha_ = alpha * X.shape[0]
l1_ratio = .2
l1_weight = alpha_ * l1_ratio
f1 = lambda w: _squared_loss(X, y, w, compute_grad=False)
f1_grad = lambda w: _squared_loss(
X, y, w, compute_grad=True, compute_energy=False)
f2_prox = lambda w, l, *args, **kwargs: (_prox_l1(w, l * l1_weight),
dict(converged=True))
total_energy = lambda w: f1(w) + l1_weight * np.sum(np.abs(w))
for cb_retval in [0, 1]:
for verbose in [0, 1]:
for dgap_factor in [1., None]:
best_w, objective, init = mfista(f1_grad,
f2_prox,
total_energy,
1.,
p,
dgap_factor=dgap_factor,
callback=lambda _: cb_retval,
verbose=verbose,
max_iter=100)
assert best_w.shape == mask.shape
assert isinstance(objective, list)
assert isinstance(init, dict)
for key in ["w", "t", "dgap_tol", "stepsize"]:
assert key in init
def test_tvl1_from_gradient(size=5, n_samples=10, random_state=42):
rng = np.random.RandomState(random_state)
shape = [size] * 3
n_voxels = np.prod(shape)
X = rng.randn(n_samples, n_voxels)
y = rng.randn(n_samples)
w = rng.randn(*shape)
mask = np.ones_like(w).astype(np.bool)
for alpha in [0., 1e-1, 1e-3]:
for l1_ratio in [0., .5, 1.]:
gradid = _gradient_id(w, l1_ratio=l1_ratio)
assert_equal(
_tvl1_objective(X, y,
w.copy().ravel(), alpha, l1_ratio, mask),
_squared_loss(X, y, w.copy().ravel(), compute_grad=False) +
alpha * _tvl1_objective_from_gradient(gradid))
def test_tvl1_from_gradient(size=5, n_samples=10, random_state=42):
rng = np.random.RandomState(random_state)
shape = [size] * 3
n_voxels = np.prod(shape)
X = rng.randn(n_samples, n_voxels)
y = rng.randn(n_samples)
w = rng.randn(*shape)
mask = np.ones_like(w).astype(np.bool)
for alpha in [0., 1e-1, 1e-3]:
for l1_ratio in [0., .5, 1.]:
gradid = _gradient_id(w, l1_ratio=l1_ratio)
assert_equal(_tvl1_objective(
X, y, w.copy().ravel(), alpha, l1_ratio, mask),
_squared_loss(X, y, w.copy().ravel(),
compute_grad=False
) + alpha * _tvl1_objective_from_gradient(
gradid))
