disable on GPU since it's not implemented there

sklearnex/linear_model/linear.py

@@ -16,6 +16,7 @@
 
 import logging
 from abc import ABC
+from functools import partialmethod
 
 import numpy as np
 from sklearn.exceptions import NotFittedError
@@ -140,7 +141,7 @@ def score(self, X, y, sample_weight=None):
             sample_weight=sample_weight,
         )
 
-    def _onedal_fit_supported(self, method_name, *data):
+    def _onedal_fit_supported(self, is_gpu, method_name, *data):
         assert method_name == "fit"
         assert len(data) == 3
         X, y, sample_weight = data
@@ -179,7 +180,7 @@ def _onedal_fit_supported(self, method_name, *data):
                     "Forced positive coefficients are not supported.",
                 ),
                 (
-                    not (is_underdetermined and not supports_all_variants),
+                    not (is_underdetermined and (not supports_all_variants or is_gpu)),
                     "The shape of X (fitting) does not satisfy oneDAL requirements:"
                     "Number of features + 1 >= number of samples.",
                 ),
@@ -212,15 +213,15 @@ def _onedal_predict_supported(self, method_name, *data):
 
         return patching_status
 
-    def _onedal_supported(self, method_name, *data):
+    def _onedal_supported(self, is_gpu, method_name, *data):
         if method_name == "fit":
-            return self._onedal_fit_supported(method_name, *data)
+            return self._onedal_fit_supported(is_gpu, method_name, *data)
         if method_name in ["predict", "score"]:
             return self._onedal_predict_supported(method_name, *data)
         raise RuntimeError(f"Unknown method {method_name} in {self.__class__.__name__}")
 
-    _onedal_gpu_supported = _onedal_supported
-    _onedal_cpu_supported = _onedal_supported
+    _onedal_gpu_supported = partialmethod(_onedal_supported, True)
+    _onedal_cpu_supported = partialmethod(_onedal_supported, False)
 
     def _initialize_onedal_estimator(self):
         onedal_params = {"fit_intercept": self.fit_intercept, "copy_X": self.copy_X}