Browse Source

!30346 change api of dynamic shape ops

Merge pull request !30346 from lianliguang/code_docs_ms
feature/build-system-rewrite
i-robot Gitee 4 years ago
parent
commit
c5811348cd
No known key found for this signature in database GPG Key ID: 173E9B9CA92EEF8F
3 changed files with 27 additions and 33 deletions
  1. +6
    -0
      docs/api/api_python/ops/mindspore.ops.DynamicShape.rst
  2. +18
    -0
      docs/api/api_python/ops/mindspore.ops.TensorShape.rst
  3. +3
    -33
      mindspore/python/mindspore/ops/operations/array_ops.py

+ 6
- 0
docs/api/api_python/ops/mindspore.ops.DynamicShape.rst View File

@@ -0,0 +1,6 @@
mindspore.ops.DynamicShape
=================

.. py:class:: mindspore.ops.DynamicShape()

与`TensorShape`相同,`DynamicShape`将会被`TensorShape`替换,请使用`TensorShape`。

+ 18
- 0
docs/api/api_python/ops/mindspore.ops.TensorShape.rst View File

@@ -0,0 +1,18 @@
mindspore.ops.TensorShape
=================

.. py:class:: mindspore.ops.TensorShape()

返回输入Tensor的Shape。

**输入:**

- **input_x** (Tensor) - 第一个输入,是一个Tensor类型数据

**输出:**

Tensor,输入`input_x`的shape

**异常:**

- **TypeError** - `input_x` 都不是Tensor。

+ 3
- 33
mindspore/python/mindspore/ops/operations/array_ops.py View File

@@ -637,13 +637,7 @@ class Shape(Primitive):

class TensorShape(Primitive):
"""
Returns the shape of the input tensor. And it used to be dynamic shape.

Note:
Dynamic shape: After the graph is running, as the tensor flows in the graph, the specific shape of the tensor
on each node on the graph can be inferred according to the structure of the graph.
This shape is called a dynamic shape. As the input shape of the graph is different,
the dynamic shape of the tensor in the graph will change.
Returns the shape of the input tensor.

Inputs:
- **input_x** (Tensor) - The shape of tensor is :math:`(x_1, x_2, ..., x_R)`.
@@ -671,32 +665,8 @@ class TensorShape(Primitive):

class DynamicShape(Primitive):
"""
Returns the shape of the input tensor. And it used to be dynamic shape.

Note:
Dynamic shape: After the graph is running, as the tensor flows in the graph, the specific shape of the tensor
on each node on the graph can be inferred according to the structure of the graph.
This shape is called a dynamic shape. As the input shape of the graph is different,
the dynamic shape of the tensor in the graph will change.

Inputs:
- **input_x** (Tensor) - The shape of tensor is :math:`(x_1, x_2, ..., x_R)`.

Outputs:
Tensor[int], 1-dim Tensor of type int32

Raises:
TypeError: If `input_x` is not a Tensor.

Supported Platforms:
``Ascend`` ``GPU`` ``CPU``

Examples:
>>> input_x = Tensor(np.ones(shape=[3, 2, 1]), mindspore.float32)
>>> shape = ops.DynamicShape()
>>> output = shape(input_x)
>>> print(output)
[3 2 1]
Same as operator TensorShape. DynamicShape will be deprecated in the future.
Please use TensorShape instead.
"""

@deprecated("1.7", "TensorShape", True)


Loading…
Cancel
Save