!8393 fix bugs of op Argmin, Assert, AssignSub, BNTrainingReduce, BasicLSTMCell and BesselI0e so on

From: @lihongkang1
Reviewed-by: @liangchenghui
Signed-off-by: @liangchenghui

mindspore/ops/operations/array_ops.py

@@ -1326,7 +1326,8 @@ class Argmin(PrimitiveWithInfer):
     Examples:
         >>> input_x = Tensor(np.array([2.0, 3.1, 1.2]), mindspore.float32)
         >>> index = P.Argmin()(input_x)
-        >>> assert index == Tensor(2, mindspore.int64)
+        >>> print(index)
+        2
     """
 
     @prim_attr_register

mindspore/ops/operations/comm_ops.py

@@ -363,6 +363,7 @@ class Broadcast(PrimitiveWithInfer):
         >>> from mindspore.communication import init
         >>> import mindspore.nn as nn
         >>> import mindspore.ops.operations as P
+        >>> import numpy as np
         >>>
         >>> init()
         >>> class Net(nn.Cell):

mindspore/ops/operations/debug_ops.py

@@ -358,12 +358,12 @@ class Assert(PrimitiveWithInfer):
         >>> class AssertDemo(nn.Cell):
         >>>     def __init__(self):
         >>>         super(AssertDemo, self).__init__()
-        >>>         self.assert = P.Assert(summarize=10)
+        >>>         self.assert1 = P.Assert(summarize=10)
         >>>         self.add = P.TensorAdd()
         >>>
         >>>     def construct(self, x, y):
         >>>         data = self.add(x, y)
-        >>>         self.assert(True, [data])
+        >>>         self.assert1(True, [data])
         >>>         return data
     """
 

mindspore/ops/operations/math_ops.py

@@ -182,7 +182,9 @@ class AssignAdd(PrimitiveWithInfer):
         >>>
         >>> net = Net()
         >>> value = Tensor(np.ones([1]).astype(np.int64)*100)
-        >>> net(value)
+        >>> output = net(value)
+        >>> print(output)
+        Parameter (name=global_step)
     """
     __mindspore_signature__ = (
         sig.make_sig('x', sig.sig_rw.RW_WRITE, dtype=sig.sig_dtype.T),
@@ -232,7 +234,9 @@ class AssignSub(PrimitiveWithInfer):
         >>>
         >>> net = Net()
         >>> value = Tensor(np.ones([1]).astype(np.int32)*100)
-        >>> net(value)
+        >>> output = net(value)
+        >>> print(output)
+        Parameter (name=global_step)
     """
 
     __mindspore_signature__ = (
@@ -3224,8 +3228,9 @@ class Atan(PrimitiveWithInfer):
         >>> tan = P.Tan()
         >>> output_y = tan(input_x)
         >>> atan = P.Atan()
-        >>> atan(output_y)
-        [[1.047, 07850001]]
+        >>> output = atan(output_y)
+        >>> print(output)
+        [[1.047, 0.7850001]]
     """
 
     @prim_attr_register
@@ -3350,7 +3355,7 @@ class BitwiseAnd(_BitwiseBinaryOp):
         - **input_x2** (Tensor) - The input tensor with same type as the `input_x1`.
 
     Outputs:
-        - **y** (Tensor) - The same type as the `input_x1`.
+        Tensor, has the same type as the `input_x1`.
 
     Examples:
          >>> input_x1 = Tensor(np.array([0, 0, 1, -1, 1, 1, 1]), mstype.int16)
@@ -3376,7 +3381,7 @@ class BitwiseOr(_BitwiseBinaryOp):
         - **input_x2** (Tensor) - The input tensor with same type as the `input_x1`.
 
     Outputs:
-        - **y** (Tensor) - The same type as the `input_x1`.
+        Tensor, has the same type as the `input_x1`.
 
     Examples:
          >>> input_x1 = Tensor(np.array([0, 0, 1, -1, 1, 1, 1]), mstype.int16)
@@ -3402,7 +3407,7 @@ class BitwiseXor(_BitwiseBinaryOp):
         - **input_x2** (Tensor) - The input tensor with same type as the `input_x1`.
 
     Outputs:
-        - **y** (Tensor) - The same type as the `input_x1`.
+        Tensor, has the same type as the `input_x1`.
 
     Examples:
          >>> input_x1 = Tensor(np.array([0, 0, 1, -1, 1, 1, 1]), mstype.int16)
@@ -3418,10 +3423,11 @@ class BesselI0e(PrimitiveWithInfer):
     Computes BesselI0e of input element-wise.
 
     Inputs:
-        - **input_x** (Tensor) - The shape of tensor is :math:`(x_1, x_2, ..., x_R)`.
+        - **input_x** (Tensor) - The shape of tensor is :math:`(x_1, x_2, ..., x_R)`. Data type must be float16 or
+          float32.
 
     Outputs:
-        Tensor, has the same shape as `input_x`. Data type must be float16 or float32.
+        Tensor, has the same shape as `input_x`.
 
     Examples:
         >>> bessel_i0e = P.BesselI0e()
@@ -3447,10 +3453,11 @@ class BesselI1e(PrimitiveWithInfer):
     Computes BesselI1e of input element-wise.
 
     Inputs:
-        - **input_x** (Tensor) - The shape of tensor is :math:`(x_1, x_2, ..., x_R)`.
+        - **input_x** (Tensor) - The shape of tensor is :math:`(x_1, x_2, ..., x_R)`. Data type must be float16 or
+          float32.
 
     Outputs:
-        Tensor, has the same shape as `input_x`. Data type must be float16 or float32.
+        Tensor, has the same shape as `input_x`.
 
     Examples:
         >>> bessel_i1e = P.BesselI1e()

mindspore/ops/operations/nn_ops.py

@@ -789,9 +789,12 @@ class BNTrainingReduce(PrimitiveWithInfer):
         - **square_sum** (Tensor) - A 1-D Tensor with float32 data type. Tensor of shape :math:`(C,)`.
 
     Examples:
-        >>> input_x = Tensor(np.ones([128, 64, 32, 64]), mindspore.float32)
+        >>> input_x = Tensor(np.ones([128, 3, 32, 3]), mindspore.float32)
         >>> bn_training_reduce = P.BNTrainingReduce()
         >>> output = bn_training_reduce(input_x)
+        >>> print(output)
+        ([1.22880000e+04, 1.22880000e+04, 1.22880000e+04],
+         [1.22880000e+04, 1.22880000e+04, 1.22880000e+04])
     """
 
     @prim_attr_register
@@ -843,15 +846,30 @@ class BNTrainingUpdate(PrimitiveWithInfer):
           Has the same shape as `variance`.
 
     Examples:
-        >>> input_x = Tensor(np.ones([128, 64, 32, 64]), mindspore.float32)
-        >>> sum = Tensor(np.ones([64]), mindspore.float32)
-        >>> square_sum = Tensor(np.ones([64]), mindspore.float32)
-        >>> scale = Tensor(np.ones([64]), mindspore.float32)
-        >>> offset = Tensor(np.ones([64]), mindspore.float32)
-        >>> mean = Tensor(np.ones([64]), mindspore.float32)
-        >>> variance = Tensor(np.ones([64]), mindspore.float32)
+        >>> input_x = Tensor(np.ones([1, 2, 2, 2]), mindspore.float32)
+        >>> sum = Tensor(np.ones([2]), mindspore.float32)
+        >>> square_sum = Tensor(np.ones([2]), mindspore.float32)
+        >>> scale = Tensor(np.ones([2]), mindspore.float32)
+        >>> offset = Tensor(np.ones([2]), mindspore.float32)
+        >>> mean = Tensor(np.ones([2]), mindspore.float32)
+        >>> variance = Tensor(np.ones([2]), mindspore.float32)
         >>> bn_training_update = P.BNTrainingUpdate()
         >>> output = bn_training_update(input_x, sum, square_sum, scale, offset, mean, variance)
+        >>> print(output)
+        ([[[[2.73200464e+00, 2.73200464e+00],
+            [2.73200464e+00, 2.73200464e+00]],
+           [[2.73200464e+00, 2.73200464e+00],
+            [2.73200464e+00, 2.73200464e+00]]]],
+         [[[[2.73200464e+00, 2.73200464e+00],
+            [2.73200464e+00, 2.73200464e+00]],
+           [[2.73200464e+00, 2.73200464e+00],
+            [2.73200464e+00, 2.73200464e+00]]]],
+         [[[[2.73200464e+00, 2.73200464e+00],
+            [2.73200464e+00, 2.73200464e+00]],
+           [[2.73200464e+00, 2.73200464e+00],
+            [2.73200464e+00, 2.73200464e+00]]]],
+            [2.50000000e-01, 2.50000000e-01],
+            [1.87500000e-01, 1.87500000e-01])
     """
 
     @prim_attr_register
@@ -928,13 +946,20 @@ class BatchNorm(PrimitiveWithInfer):
         - **reserve_space_2** (Tensor) - Tensor of shape :math:`(C,)`.
 
     Examples:
-        >>> input_x = Tensor(np.ones([32, 64]), mindspore.float32)
-        >>> scale = Tensor(np.ones([64]), mindspore.float32)
-        >>> bias = Tensor(np.ones([64]), mindspore.float32)
-        >>> mean = Tensor(np.ones([64]), mindspore.float32)
-        >>> variance = Tensor(np.ones([64]), mindspore.float32)
+        >>> input_x = Tensor(np.ones([2, 2]), mindspore.float32)
+        >>> scale = Tensor(np.ones([2]), mindspore.float32)
+        >>> bias = Tensor(np.ones([2]), mindspore.float32)
+        >>> mean = Tensor(np.ones([2]), mindspore.float32)
+        >>> variance = Tensor(np.ones([2]), mindspore.float32)
         >>> batch_norm = P.BatchNorm()
         >>> output = batch_norm(input_x, scale, bias, mean, variance)
+        >>> print(output)
+        ([[1.0, 1.0],
+          [1.0, 1.0]],
+          [1.0, 1.0],
+          [1.0, 1.0],
+          [1.0, 1.0],
+          [1.0, 1.0])
     """
 
     @prim_attr_register
@@ -1704,7 +1729,10 @@ class BiasAdd(PrimitiveWithInfer):
         >>> input_x = Tensor(np.arange(6).reshape((2, 3)), mindspore.float32)
         >>> bias = Tensor(np.random.random(3).reshape((3,)), mindspore.float32)
         >>> bias_add = P.BiasAdd()
-        >>> bias_add(input_x, bias)
+        >>> output = bias_add(input_x, bias)
+        >>> print(output)
+        [[0.4662124 1.2493685 2.3611782]
+         [3.4662123 4.2493687 5.3611784]]
     """
 
     @prim_attr_register
@@ -4817,7 +4845,7 @@ class ApplyPowerSign(PrimitiveWithInfer):
         >>>         self.beta = 0.9
         >>>     def construct(self, grad):
         >>>         out = self.apply_power_sign(self.var, self.m, self.lr, self.logbase,
-                                                self.sign_decay, self.beta, grad)
+        >>>                                        self.sign_decay, self.beta, grad)
         >>>         return out
         >>> net = Net()
         >>> grad = Tensor(np.random.rand(3, 3).astype(np.float32))
@@ -5594,7 +5622,7 @@ class BasicLSTMCell(PrimitiveWithInfer):
             LSTM layer except the last layer. Default 1.0. The range of dropout is [0.0, 1.0].
         forget_bias (float): Add forget bias to forget gate biases in order to decrease former scale. Default: 1.0.
         state_is_tuple (bool): If true, the state is a tuple of 2 tensors, containing h and c; If false, the state is
-        a tensor and it needs to be split first. Default: True.
+            a tensor and it needs to be split first. Default: True.
         activation (str): Activation. Default: "tanh". Only "tanh" is currently supported.
 
     Inputs:
@@ -5626,12 +5654,20 @@ class BasicLSTMCell(PrimitiveWithInfer):
 
     Examples:
         >>> x = Tensor(np.random.rand(1, 32).astype(np.float16))
-        >>> h = Tensor(np.random.rand(1, 64).astype(np.float16))
-        >>> c = Tensor(np.random.rand(1, 64).astype(np.float16))
-        >>> w = Tensor(np.random.rand(96, 256).astype(np.float16))
-        >>> b = Tensor(np.random.rand(256, ).astype(np.float16))
+        >>> h = Tensor(np.random.rand(1, 2).astype(np.float16))
+        >>> c = Tensor(np.random.rand(1, 2).astype(np.float16))
+        >>> w = Tensor(np.random.rand(34, 8).astype(np.float16))
+        >>> b = Tensor(np.random.rand(8, ).astype(np.float16))
         >>> lstm = P.BasicLSTMCell(keep_prob=1.0, forget_bias=1.0, state_is_tuple=True, activation='tanh')
-        >>> lstm(x, h, c, w, b)
+        >>> output = lstm(x, h, c, w, b)
+        >>> print(output)
+        ([[9.5459e-01, 9.2725e-01]],
+         [[1.0000e+00, 1.0000e+00]],
+         [[1.0000e+00, 1.0000e+00]],
+         [[1.0000e+00, 1.0000e+00]],
+         [[9.9951e-01, 1.0000e+00]],
+         [[9.5459e-01, 9.2773e-01]],
+         [[0.0000e+00, 0.0000e+00]])
     """
 
     @prim_attr_register

mindspore/ops/operations/other_ops.py

@@ -48,7 +48,9 @@ class Assign(PrimitiveWithCheck):
         >>>         return self.y
         >>> x = Tensor([2.0], mindspore.float32)
         >>> net = Net()
-        >>> net(x)
+        >>> output = net(x)
+        >>> print(output)
+        Parameter (name=y)
     """
     __mindspore_signature__ = (
         sig.make_sig('variable', sig.sig_rw.RW_WRITE, dtype=sig.sig_dtype.T),

mindspore/ops/operations/random_ops.py

@@ -134,11 +134,14 @@ class Gamma(PrimitiveWithInfer):
         The dtype is float32.
 
     Examples:
-        >>> shape = (4, 16)
+        >>> shape = (2, 2)
         >>> alpha = Tensor(1.0, mstype.float32)
         >>> beta = Tensor(1.0, mstype.float32)
         >>> gamma = P.Gamma(seed=3)
-        >>> output = Gamma(shape, alpha, beta)
+        >>> output = gamma(shape, alpha, beta)
+        >>> print(output)
+        [[0.21962446 0.33740655]
+         [1.0859369  0.25875127]]
     """
 
     @prim_attr_register