Hendrik/xgboost
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

[dask] Rework base margin test. (#6627)

Browse Source
This commit is contained in:
Jiaming Yuan 2021-01-22 17:49:13 +08:00 committed by GitHub
parent 7bc56fa0ed
commit a275f40267
No known key found for this signature in database
GPG Key ID: 4AEE18F83AFDEB23
1 changed files with 19 additions and 57 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

70
tests/python/test_with_dask.py
View File

@ -149,68 +149,30 @@ def test_dask_predict_shape_infer() -> None:
@pytest.mark.parametrize("tree_method", ["hist", "approx"]) @pytest.mark.parametrize("tree_method", ["hist", "approx"])
def test_boost_from_prediction(tree_method: str) -> None: def test_boost_from_prediction(tree_method: str, client: "Client") -> None:
if tree_method == 'approx':
pytest.xfail(reason='test_boost_from_prediction[approx] is flaky')
from sklearn.datasets import load_breast_cancer from sklearn.datasets import load_breast_cancer
X, y = load_breast_cancer(return_X_y=True) X_, y_ = load_breast_cancer(return_X_y=True)
X_ = dd.from_array(X, chunksize=100) X, y = dd.from_array(X_, chunksize=100), dd.from_array(y_, chunksize=100)
y_ = dd.from_array(y, chunksize=100)
with LocalCluster(n_workers=4) as cluster:
with Client(cluster) as _:
model_0 = xgb.dask.DaskXGBClassifier( model_0 = xgb.dask.DaskXGBClassifier(
learning_rate=0.3, learning_rate=0.3, random_state=0, n_estimators=4,
random_state=123, tree_method=tree_method)
n_estimators=4, model_0.fit(X=X, y=y)
tree_method=tree_method, margin = model_0.predict(X, output_margin=True)
)
model_0.fit(X=X_, y=y_)
margin = model_0.predict(X_, output_margin=True)
model_1 = xgb.dask.DaskXGBClassifier( model_1 = xgb.dask.DaskXGBClassifier(
learning_rate=0.3, learning_rate=0.3, random_state=0, n_estimators=4,
random_state=123, tree_method=tree_method)
n_estimators=4, model_1.fit(X=X, y=y, base_margin=margin)
tree_method=tree_method, predictions_1 = model_1.predict(X, base_margin=margin)
)
model_1.fit(X=X_, y=y_, base_margin=margin)
predictions_1 = model_1.predict(X_, base_margin=margin)
proba_1 = model_1.predict_proba(X_, base_margin=margin)
cls_2 = xgb.dask.DaskXGBClassifier( cls_2 = xgb.dask.DaskXGBClassifier(
learning_rate=0.3, learning_rate=0.3, random_state=0, n_estimators=8,
random_state=123, tree_method=tree_method)
n_estimators=8, cls_2.fit(X=X, y=y)
tree_method=tree_method, predictions_2 = cls_2.predict(X)
)
cls_2.fit(X=X_, y=y_)
predictions_2 = cls_2.predict(X_)
proba_2 = cls_2.predict_proba(X_)
cls_3 = xgb.dask.DaskXGBClassifier( assert np.all(predictions_1.compute() == predictions_2.compute())
learning_rate=0.3,
random_state=123,
n_estimators=8,
tree_method=tree_method,
)
cls_3.fit(X=X_, y=y_)
proba_3 = cls_3.predict_proba(X_)
# compute variance of probability percentages between two of the
# same model, use this to check to make sure approx is functioning
# within normal parameters
expected_variance = np.max(np.abs(proba_3 - proba_2)).compute()
if expected_variance > 0:
margin_variance = np.max(np.abs(proba_1 - proba_2)).compute()
# Ensure the margin variance is less than the expected variance + 10%
assert np.all(margin_variance <= expected_variance + .1)
else:
np.testing.assert_equal(predictions_1.compute(), predictions_2.compute())
np.testing.assert_almost_equal(proba_1.compute(), proba_2.compute())
def test_dask_missing_value_reg() -> None: def test_dask_missing_value_reg() -> None: