fix numpy bugs

mindspore/ccsrc/backend/kernel_compiler/gpu/math/broadcast_gpu_kernel.cc

@@ -199,6 +199,9 @@ MS_REG_GPU_KERNEL_ONE(
 MS_REG_GPU_KERNEL_ONE(
   Div, KernelAttr().AddInputAttr(kNumberTypeInt32).AddInputAttr(kNumberTypeInt32).AddOutputAttr(kNumberTypeInt32),
   BroadcastOpGpuKernel, int)
+MS_REG_GPU_KERNEL_ONE(
+  RealDiv, KernelAttr().AddInputAttr(kNumberTypeInt32).AddInputAttr(kNumberTypeInt32).AddOutputAttr(kNumberTypeInt32),
+  BroadcastOpGpuKernel, int)
 MS_REG_GPU_KERNEL_ONE(
   DivNoNan, KernelAttr().AddInputAttr(kNumberTypeInt32).AddInputAttr(kNumberTypeInt32).AddOutputAttr(kNumberTypeInt32),
   BroadcastOpGpuKernel, int)

mindspore/numpy/array_creations.py

@@ -516,7 +516,8 @@ def linspace(start, stop, num=50, endpoint=True, retstep=False, dtype=None, axis
     iota_shape = _tuple_slice(iota_shape, None, axis) + (num,) + _tuple_slice(iota_shape, axis+1, None)
     num_tensor = _type_convert(Tensor, num).astype(mstype.float32)
     div = (num_tensor - 1) if endpoint else num_tensor
-
+    out = None
+    delta = None
     if num > 1:
         delta = (stop - start) / div
         # This is similar to how numpy and jax compute linspace
@@ -530,8 +531,7 @@ def linspace(start, stop, num=50, endpoint=True, retstep=False, dtype=None, axis
         delta = nan if endpoint else stop - start
         out = reshape(start, bounds_shape)
     else:  # num == 0
-        delta = nan
-        out = _type_convert(Tensor, []).astype(dtype)
+        _raise_value_error("cannot support Tensor with num=0.")
     if retstep:
         return out.astype(dtype), delta
     return out.astype(dtype)
@@ -631,7 +631,7 @@ def geomspace(start, stop, num=50, endpoint=True, dtype=None, axis=0):
     root = num
     if endpoint:
         root -= 1
-    bases = F.tensor_pow(F.tensor_div(stop, start), asarray_const(1/(root)))
+    bases = F.tensor_pow(F.tensor_div(stop, start), asarray_const(1./(root)))
     exponents = linspace(zeros(F.shape(bases)), F.fill(F.dtype(bases), F.shape(bases), root),
                          num, endpoint=endpoint, dtype=dtype, axis=axis)
     shape = F.shape(bases)

mindspore/numpy/array_ops.py

@@ -1422,6 +1422,8 @@ def _split(x, indices_or_sections, opname, axis=0):
     arr_shape = x.shape
     length_along_dim = arr_shape[axis]
     if isinstance(indices_or_sections, int):
+        if indices_or_sections > length_along_dim:
+            _raise_value_error("empty tensor encountered.")
         if opname == "split" or length_along_dim % indices_or_sections == 0:
             res = P.Split(axis, indices_or_sections)(x)
         else:
@@ -1461,6 +1463,8 @@ def _split_sub_tensors(x, indices, axis):
     for i, idx in enumerate(indices):
         begin[axis] = 0 if i == 0 else indices[i-1]
         end[axis] = idx
+        if end[axis] <= begin[axis]:
+            _raise_value_error("empty sub-tensor encountered.")
         sliced_tensor = F.strided_slice(x, _type_convert(tuple, begin), _type_convert(tuple, end), strides)
         sub_tensors.append(sliced_tensor)
     return sub_tensors
@@ -2136,19 +2140,19 @@ def choose(a, choices, mode='clip'):
     with ``shape Ba.shape`` is created as follows:
 
     - if ``mode='raise'`` (the default), then, first of all, each element of `a`
-    (and thus `Ba`) must be in the range `[0, n-1]`; now, suppose that `i`
-    (in that range) is the value at the `(j0, j1, ..., jm)` position in
-    `Ba` - then the value at the same position in the new array is the
-    value in ``Bchoices[i]`` at that same position;
+      (and thus `Ba`) must be in the range `[0, n-1]`; now, suppose that `i`
+      (in that range) is the value at the `(j0, j1, ..., jm)` position in
+      `Ba` - then the value at the same position in the new array is the
+      value in ``Bchoices[i]`` at that same position;
 
     - if ``mode='wrap'``, values in `a` (and thus `Ba`) may be any (signed)
-    integer; modular arithmetic is used to map integers outside the
-    range ``[0, n-1]`` back into that range; and then the new array is
-    constructed as above;
+      integer; modular arithmetic is used to map integers outside the
+      range ``[0, n-1]`` back into that range; and then the new array is
+      constructed as above;
 
     - if ``mode='clip'``, values in `a` (and thus `Ba`) may be any (signed) integer;
-    negative integers are mapped to 0; values greater than `n-1` are mapped to
-    `n-1`; and then the new array is constructed as above.
+      negative integers are mapped to 0; values greater than `n-1` are mapped to
+      `n-1`; and then the new array is constructed as above.
 
     Note:
         Numpy argument `out` is not supported.

mindspore/numpy/math_ops.py

@@ -5645,10 +5645,10 @@ def bitwise_and(x1, x2, dtype=None):
 
 
 def bitwise_or(x1, x2, dtype=None):
-    """
+    r"""
     Computes the bit-wise OR of two arrays element-wise.
     Computes the bit-wise OR of the underlying binary representation of the integers in
-    the input arrays. This ufunc implements the C/Python operator |.
+    the input arrays. This ufunc implements the C/Python operator \|.
 
     Note:
         Numpy arguments `out`, `where`, `casting`, `order`, `subok`, `signature`, and `extobj` are