sklearn fix and tests

python-package/xgboost/sklearn.py

@@ -67,8 +67,10 @@ def _check_rf_callback(
         )
 
 
-def _can_use_qdm(tree_method: Optional[str]) -> bool:
-    return tree_method in ("hist", "gpu_hist", None, "auto")
+def _can_use_qdm(tree_method: Optional[str],
+                 device: Optional[str]) -> bool:
+    is_sycl = device.startswith("sycl")
+    return tree_method in ("hist", "gpu_hist", None, "auto") and not is_sycl
 
 
 class _SklObjWProto(Protocol):  # pylint: disable=too-few-public-methods
@@ -1031,7 +1033,7 @@ def _duplicated(parameter: str) -> None:
 
     def _create_dmatrix(self, ref: Optional[DMatrix], **kwargs: Any) -> DMatrix:
         # Use `QuantileDMatrix` to save memory.
-        if _can_use_qdm(self.tree_method) and self.booster != "gblinear":
+        if _can_use_qdm(self.tree_method, self.device) and self.booster != "gblinear":
             try:
                 return QuantileDMatrix(
                     **kwargs, ref=ref, nthread=self.n_jobs, max_bin=self.max_bin

tests/python-sycl/test_sycl_with_sklearn.py

@@ -0,0 +1,37 @@
+import xgboost as xgb
+import pytest
+import sys
+import numpy as np
+
+from xgboost import testing as tm
+
+sys.path.append("tests/python")
+import test_with_sklearn as twskl  # noqa
+
+pytestmark = pytest.mark.skipif(**tm.no_sklearn())
+
+rng = np.random.RandomState(1994)
+
+
+def test_sycl_binary_classification():
+    from sklearn.datasets import load_digits
+    from sklearn.model_selection import KFold
+
+    digits = load_digits(n_class=2)
+    y = digits["target"]
+    X = digits["data"]
+    kf = KFold(n_splits=2, shuffle=True, random_state=rng)
+    for cls in (xgb.XGBClassifier, xgb.XGBRFClassifier):
+        for train_index, test_index in kf.split(X, y):
+            xgb_model = cls(random_state=42, device="sycl", n_estimators=4).fit(
+                X[train_index], y[train_index]
+            )
+            preds = xgb_model.predict(X[test_index])
+            labels = y[test_index]
+            err = sum(
+                1 for i in range(len(preds)) if int(preds[i] > 0.5) != labels[i]
+            ) / float(len(preds))
+            print(preds)
+            print(labels)
+            print(err)
+            assert err < 0.1