diff --git a/sklearn/metrics/_regression.py b/sklearn/metrics/_regression.py
index 8806719eabdd1..62251f9b96188 100644
--- a/sklearn/metrics/_regression.py
+++ b/sklearn/metrics/_regression.py
@@ -71,6 +71,11 @@ def _check_reg_targets(y_true, y_pred, multioutput, dtype="numeric", xp=None):
     dtype : str or list, default="numeric"
         the dtype argument passed to check_array.
 
+    xp : module, default=None
+        Precomputed array namespace module. When passed, typically from a caller
+        that has already performed inspection of its own inputs, skips array
+        namespace inspection.
+
     Returns
     -------
     type_true : one of {'continuous', continuous-multioutput'}
@@ -398,7 +403,7 @@ def mean_absolute_percentage_error(
     dtype = _find_matching_floating_dtype(y_true, y_pred, sample_weight, xp=xp)
 
     y_type, y_true, y_pred, multioutput = _check_reg_targets(
-        y_true, y_pred, multioutput
+        y_true, y_pred, multioutput, dtype=dtype, xp=xp
     )
     check_consistent_length(y_true, y_pred, sample_weight)
     epsilon = xp.asarray(xp.finfo(xp.float64).eps, dtype=dtype)
@@ -1253,7 +1258,7 @@ def max_error(y_true, y_pred):
     np.int64(1)
     """
     xp, _ = get_namespace(y_true, y_pred)
-    y_type, y_true, y_pred, _ = _check_reg_targets(y_true, y_pred, None)
+    y_type, y_true, y_pred, _ = _check_reg_targets(y_true, y_pred, None, xp=xp)
     if y_type == "continuous-multioutput":
         raise ValueError("Multioutput not supported in max_error")
     return xp.max(xp.abs(y_true - y_pred))
@@ -1352,7 +1357,7 @@ def mean_tweedie_deviance(y_true, y_pred, *, sample_weight=None, power=0):
     """
     xp, _ = get_namespace(y_true, y_pred)
     y_type, y_true, y_pred, _ = _check_reg_targets(
-        y_true, y_pred, None, dtype=[xp.float64, xp.float32]
+        y_true, y_pred, None, dtype=[xp.float64, xp.float32], xp=xp
     )
     if y_type == "continuous-multioutput":
         raise ValueError("Multioutput not supported in mean_tweedie_deviance")