fix(ivy): Extends ivy.pow to work for all input and exponent cases

ivy/functional/backends/numpy/elementwise.py

@@ -611,7 +611,9 @@ def pow(
     out: Optional[np.ndarray] = None,
 ) -> np.ndarray:
     x1, x2 = ivy.promote_types_of_inputs(x1, x2)
-    return np.power(x1, x2, out=out)
+    if ivy.is_int_dtype(x1) and ivy.any(x2 < 0):
+        return np.float_power(x1, x2, casting='unsafe').astype(x1.dtype)
+    return np.power(x1, x2)
 
 
 pow.support_native_out = True

ivy/functional/backends/tensorflow/elementwise.py

@@ -620,7 +620,14 @@ def pow(
             if x2.dtype.is_unsigned:
                 x2 = tf.cast(x2, tf.float64)
             return tf.cast(tf.experimental.numpy.power(x1, x2), promoted_type)
-    return tf.experimental.numpy.power(x1, x2)
+    orig_x1_dtype = None
+    if ivy.is_int_dtype(x1) and ivy.any(x2 < 0):
+        orig_x1_dtype = x1.dtype
+        x1 = tf.cast(x1, tf.float32)
+    ret = tf.experimental.numpy.power(x1, x2)
+    if orig_x1_dtype is not None:
+        return tf.cast(ret, orig_x1_dtype)
+    return ret
 
 
 @with_unsupported_dtypes({"2.13.0 and below": ("bfloat16", "complex")}, backend_version)

ivy/functional/ivy/elementwise.py

@@ -5201,13 +5201,6 @@ def pow(
     (the exponent), where ``x2_i`` is the corresponding element of the input array
     ``x2``.
 
-    .. note::
-       If both ``x1`` and ``x2`` have integer data types, the result of ``pow`` when
-       ``x2_i`` is negative (i.e., less than zero) is unspecified and thus
-       implementation-dependent. If ``x1`` has an integer data type and ``x2`` has a
-       floating-point data type, behavior is implementation-dependent (type promotion
-       between data type "kinds" (integer versus floating-point) is unspecified).
-
     **Special cases**
 
     For floating-point operands,

ivy_tests/test_ivy/test_functional/test_core/test_elementwise.py

@@ -136,10 +136,6 @@ def cast_filter(dtype1_x1_dtype2):
         )
     )
     dtype2 = dtype2[0]
-    if "int" in dtype2:
-        x2 = ivy.nested_map(
-            x2[0], lambda x: abs(x), include_derived={"list": True}, shallow=False
-        )
     return [dtype1, dtype2], [x1, x2]
 
 
@@ -1593,19 +1589,8 @@ def test_positive(*, dtype_and_x, test_flags, backend_fw, fn_name, on_device):
 )
 def test_pow(*, dtype_and_x, test_flags, backend_fw, fn_name, on_device):
     input_dtype, x = dtype_and_x
-
     # bfloat16 is not supported by numpy
-    assume("bfloat16" not in input_dtype)
-
-    # Make sure x2 isn't a float when x1 is integer
-    assume(
-        not (ivy.is_int_dtype(input_dtype[0] and ivy.is_float_dtype(input_dtype[1])))
-    )
-
-    # Make sure x2 is non-negative when both is integer
-    if ivy.is_int_dtype(input_dtype[1]) and ivy.is_int_dtype(input_dtype[0]):
-        x[1] = np.abs(x[1])
-
+    assume(not ("bfloat16" in input_dtype))
     x[0] = not_too_close_to_zero(x[0])
     x[1] = not_too_close_to_zero(x[1])
     helpers.test_function(