def test_embedding_no_layers():
# Make sure the embedding works with no layers.
data = load_diabetes()
X, y = data['data'], data['target']
clf = MLPRegressor(n_epochs=1, hidden_units=[])
clf.fit(X, y)
assert clf.transform(X).shape[1] == 1
def test_embedding_default():
# Make sure the embedding works by default.
data = load_diabetes()
X, y = data['data'], data['target']
clf = MLPRegressor(n_epochs=1)
clf.fit(X, y)
assert clf.transform(X).shape[1] == 256
def test_partial_fit():
data = load_diabetes()
clf = MLPRegressor(n_epochs=1)
X, y = data['data'], data['target']
for _ in range(30):
clf.partial_fit(X, y)
y_pred = clf.predict(X)
assert pearsonr(y_pred, y)[0] > 0.5
def test_partial_fit():
data = load_diabetes()
clf = MLPRegressor(n_epochs=1)
X, y = data['data'], data['target']
for _ in range(30):
clf.partial_fit(X, y)
y_pred = clf.predict(X)
assert pearsonr(y_pred, y)[0] > 0.5
def test_embedding_specific_layer():
# Make sure the embedding works with no layers.
data = load_diabetes()
X, y = data['data'], data['target']
clf = MLPRegressor(n_epochs=1,
hidden_units=(256, 8, 256),
transform_layer_index=1)
clf.fit(X, y)
assert clf.transform(X).shape[1] == 8
def test_replicability():
"""Make sure running fit twice in a row finds the same parameters."""
diabetes = load_diabetes()
X_diabetes, y_diabetes = diabetes.data, diabetes.target
ind = np.arange(X_diabetes.shape[0])
rng = np.random.RandomState(0)
rng.shuffle(ind)
X_diabetes, y_diabetes = X_diabetes[ind], y_diabetes[ind]
clf = MLPRegressor(keep_prob=0.9, random_state=42, n_epochs=100)
target = y_diabetes
# Just predict on the training set, for simplicity.
pred1 = clf.fit(X_diabetes, target).predict(X_diabetes)
pred2 = clf.fit(X_diabetes, target).predict(X_diabetes)
assert_array_almost_equal(pred1, pred2)
def test_replicability():
"""Make sure running fit twice in a row finds the same parameters."""
diabetes = load_diabetes()
X_diabetes, y_diabetes = diabetes.data, diabetes.target
ind = np.arange(X_diabetes.shape[0])
rng = np.random.RandomState(0)
rng.shuffle(ind)
X_diabetes, y_diabetes = X_diabetes[ind], y_diabetes[ind]
clf = MLPRegressor(keep_prob=0.9, random_state=42, n_epochs=100)
target = y_diabetes
# Just predict on the training set, for simplicity.
pred1 = clf.fit(X_diabetes, target).predict(X_diabetes)
pred2 = clf.fit(X_diabetes, target).predict(X_diabetes)
assert_array_almost_equal(pred1, pred2)
def test_predict():
"""Test binary classification."""
check_predictions(MLPRegressor(**KWARGS), X, Y)
check_predictions(MLPRegressor(**KWARGS), X_sp, Y)
def test_sample_weight():
"""Ensure we handle sample weights for regression problems."""
assert_sample_weights_work(
make_regression, {'n_samples': 3000}, lambda: MLPRegressor(
n_epochs=30, random_state=42, keep_prob=0.8, hidden_units=(128, )))