Revert "Intermediate values use default dtypes in statistical_functio…

…ns.py."

This reverts commit 6a2f60f.

cubed/array_api/statistical_functions.py

@@ -27,15 +27,15 @@ def max(x, /, *, axis=None, keepdims=False, split_every=None):
     )
 
 
-def mean(x, /, *, axis=None, keepdims=False, split_every=None, device=None):
+def mean(x, /, *, axis=None, keepdims=False, split_every=None):
     if x.dtype not in _real_floating_dtypes:
         raise TypeError("Only real floating-point dtypes are allowed in mean")
     # This implementation uses a Zarr group of two arrays to store a
     # pair of fields needed to keep per-chunk counts and totals for computing
     # the mean.
     dtype = x.dtype
-    dtypes = __array_namespace_info__().default_dtypes(device=device)
-    intermediate_dtype = [("n", dtypes['integral']), ("total", dtypes['real floating'])]
+    #TODO(#658): Should these be default dtypes?
+    intermediate_dtype = [("n", nxp.int64), ("total", nxp.float64)]
     extra_func_kwargs = dict(dtype=intermediate_dtype)
     return reduction(
         x,
@@ -161,15 +161,14 @@ def var(
     correction=0.0,
     keepdims=False,
     split_every=None,
-    device=None,
 ):
     # This implementation follows https://en.wikipedia.org/wiki/Algorithms_for_calculating_variance#Parallel_algorithm
 
     if x.dtype not in _real_floating_dtypes:
         raise TypeError("Only real floating-point dtypes are allowed in var")
     dtype = x.dtype
-    dtypes = __array_namespace_info__().default_dtypes(device=device)
-    intermediate_dtype = [("n", dtypes['integral']), ("mu", dtypes['real floating']), ("M2", dtypes['real floating'])]
+    #TODO(#658): Should these be default dtypes?
+    intermediate_dtype = [("n", nxp.int64), ("mu", nxp.float64), ("M2", nxp.float64)]
     extra_func_kwargs = dict(dtype=intermediate_dtype, correction=correction)
     return reduction(
         x,