Huawei_Technology
/
mindspore
Not watched
1
0
Fork 0
Code
Releases 13 Wiki evaluate Activity
Issues 0 Pull Requests 0 Datasets Model Cloudbrain HPC
Browse Source

perfect annotation of ops and context and support bool equal

tags/v1.0.0
buxue 5 years ago
parent
952135362f
commit
c3c06514d7
10 changed files with 224 additions and 197 deletions
Split View
Diff Options
Show Stats Download Patch File Download Diff File
  1. +36
    -35
      mindspore/context.py
  2. +15
    -0
      mindspore/ops/composite/multitype_ops/equal_impl.py
  3. +1
    -0
      mindspore/ops/functional.py
  4. +92
    -88
      mindspore/ops/operations/array_ops.py
  5. +1
    -1
      mindspore/ops/operations/debug_ops.py
  6. +26
    -26
      mindspore/ops/operations/math_ops.py
  7. +37
    -37
      mindspore/ops/operations/nn_ops.py
  8. +7
    -7
      mindspore/ops/operations/random_ops.py
  9. +2
    -2
      mindspore/ops/operations/sparse_ops.py
  10. +7
    -1
      tests/ut/python/runtest.sh

+ 36
- 35
mindspore/context.py View File

@@ -14,7 +14,7 @@
# ============================================================================
"""
The context of mindspore, used to configure the current execution environment,
including execution mode, execution backend and other feature switches.
includes the execution mode, execution backend and other feature switches.
"""
import os
import time
@@ -338,40 +338,40 @@ def set_auto_parallel_context(**kwargs):

Note:
Attribute name is required for setting attributes.
If a program has tasks with different parallel modes, then before setting new parallel mode for
If a program has tasks with different parallel modes, then before setting new parallel mode for the
next task, interface mindspore.context.reset_auto_parallel_context() needs to be called to reset
the configuration.
Setting or changing parallel modes must be called before any Initializer created, or RuntimeError
may be raised when compile network.
Setting or changing parallel modes must be called before any creating Initializer, otherwise,
RuntimeError may be raised when compiling the network.

Args:
device_num (int): Available device number, the value must be in [1, 4096]. Default: 1.
global_rank (int): Global rank id, the value must be in [0, 4095]. Default: 0.
gradients_mean (bool): Whether to perform mean operator after all-reduce of mirror.
"stand_alone" do not support gradients_mean. Default: False.
gradient_fp32_sync (bool): Gradients allreduce by fp32 even though gradients is fp16 if this flag is True..
"stand_alone" does not support `gradients_mean`. Default: False.
gradient_fp32_sync (bool): Gradients allreduce by fp32, even though gradients is fp16 if this flag is True..
"stand_alone", "data_parallel" and "hybrid_parallel" do not support
gradient_fp32_sync. Default: True.
parallel_mode (str): There are five kinds of parallel modes, "stand_alone", "data_parallel",
"hybrid_parallel", "semi_auto_parallel" and "auto_parallel". Default: "stand_alone".

- stand_alone: Only one processor working.
- stand_alone: Only one processor is working.

- data_parallel: Distributing the data across different processors.
- data_parallel: Distributes the data across different processors.

- hybrid_parallel: Achieving data parallelism and model parallelism manually.
- hybrid_parallel: Achieves data parallelism and model parallelism manually.

- semi_auto_parallel: Achieving data parallelism and model parallelism by
- semi_auto_parallel: Achieves data parallelism and model parallelism by
setting parallel strategies.

- auto_parallel: Achieving parallelism automatically.
- auto_parallel: Achieves parallelism automatically.
auto_parallel_search_mode (str): There are two kinds of search modes, "recursive_programming"
and "dynamic_programming". Default: "dynamic_programming".

- recursive_programming: Recursive programming search mode.

- dynamic_programming: Dynamic programming search mode.
parameter_broadcast (bool): Indicating whether to broadcast parameters before training.
parameter_broadcast (bool): Whether to broadcast parameters before training.
"stand_alone", "semi_auto_parallel" and "auto_parallel" do not support parameter
broadcast. Default: False.
strategy_ckpt_load_file (str): The path to load parallel strategy checkpoint. Default: ''
@@ -468,7 +468,7 @@ def set_context(**kwargs):

When the `save_graphs` attribute is set to True, attribute of `save_graphs_path` is used to set the
intermediate compilation graph storage path. By default, the graphs are saved in the current directory.
As for other configurations and arguments, please refer to the corresponding module
For other configurations and arguments, please refer to the corresponding module
description, the configuration is optional and can be enabled when needed.

Note:
@@ -498,9 +498,9 @@ def set_context(**kwargs):

Args:
mode (int): Running in GRAPH_MODE(0) or PYNATIVE_MODE(1). Default: PYNATIVE_MODE(1).
device_target (str): The target device to run, support "Ascend", "GPU", "CPU". Default: "Ascend".
device_id (int): Id of target device, the value must be in [0, device_num_per_host-1],
while device_num_per_host should no more than 4096. Default: 0.
device_target (str): The target device to run, support "Ascend", "GPU", and "CPU". Default: "Ascend".
device_id (int): ID of the target device, the value must be in [0, device_num_per_host-1],
while device_num_per_host should be no more than 4096. Default: 0.
save_graphs (bool): Whether to save graphs. Default: False.
save_graphs_path (str): Path to save graphs. Default: "."
enable_auto_mixed_precision (bool): Whether to enable auto mixed precision. Default: False.
@@ -509,33 +509,34 @@ def set_context(**kwargs):
reserve_class_name_in_scope (bool) : Whether to save the network class name in the scope. Default: True.
enable_reduce_precision (bool): Whether to enable precision reduction. Default: True.
enable_dump (bool): Whether to enable dump. Default: False.
save_dump_path (str): When the program is executed on Ascend, operators can dump data here.
save_dump_path (str): When the program is executed on Ascend, operators can dump data in this path.
The root dump path is configured in /home/HwHiAiUser/ide_daemon/ide_daemon.cfg.
So the real dump path is "{configured root dump path}/{`save_dump_path`}". Default: ".".
variable_memory_max_size (str): Sets variable memory max size. Default: "0GB".
variable_memory_max_size (str): Set the maximum size of the variable memory max size. Default: "0GB".
enable_profiling (bool): Whether to open profiling. Default: False.
profiling_options (str): Sets profiling collection options, operators can profiling data here.
Profiling collection options, the values are as follows, supporting the collection of multiple data.
profiling_options (str): Set profiling collection options, operators can profiling data here.
The values of profiling collection options are as follows, supporting the collection of multiple data.

- training_trace: collect iterative trajectory data, that is, the training task and software information of
the AI software stack, to achieve performance analysis of the training task, focusing on data
enhancement, forward and backward calculation, gradient aggregation update and other related data.

- task_trace: collect task trajectory data, that is, the hardware information of the HWTS/AICore of
the Ascend 910 processor, and analyze the information of start and end of the task.
the Ascend 910 processor, and analyze the information of beginning and ending of the task.

- op_trace: collect single operator performance data.
The profiling can choose training_trace, task_trace, training_trace and task_trace combination and
separated by colons; single operator can choose op_trace, op_trace cannot be combined with
training_trace and task_trace. Default: "training_trace".
The profiling can choose the combination of `training_trace`, `task_trace`,
`training_trace` and `task_trace` combination, and eparated by colons;
a single operator can choose `op_trace`, `op_trace` cannot be combined with
`training_trace` and `task_trace`. Default: "training_trace".
check_bprop (bool): Whether to check bprop. Default: False.
max_device_memory (str): Sets the maximum memory available for device, currently only supported on GPU.
The format is "xxGB". Default: "1024GB".
print_file_path (str): The path of print data to save. If this parameter is set, print data is saved to
a file by default, and turn off printing to the screen. If the file already exists, add a timestamp
max_device_memory (str): Sets the maximum memory available for devices.
Currently, it is only supported on GPU. The format is "xxGB". Default: "1024GB".
print_file_path (str): The path of saving print data. If this parameter is set, print data is saved to
a file by default, and turns off printing to the screen. If the file already exists, add a timestamp
suffix to the file. Default: ''.
enable_sparse (bool): Whether to enable sparsity feature. Default: False.
max_call_depth(int): Specify the function call depth limit. Default: 1000.
max_call_depth(int): Specify the maximum depth of function call. Default: 1000.

Raises:
ValueError: If input key is not an attribute in context.
@@ -614,13 +615,13 @@ class ParallelMode:
There are five kinds of parallel modes, "STAND_ALONE", "DATA_PARALLEL",
"HYBRID_PARALLEL", "SEMI_AUTO_PARALLEL" and "AUTO_PARALLEL". Default: "STAND_ALONE".

- STAND_ALONE: Only one processor working.
- DATA_PARALLEL: Distributing the data across different processors.
- HYBRID_PARALLEL: Achieving data parallelism and model parallelism manually.
- SEMI_AUTO_PARALLEL: Achieving data parallelism and model parallelism by setting parallel strategies.
- AUTO_PARALLEL: Achieving parallelism automatically.
- STAND_ALONE: Only one processor is working.
- DATA_PARALLEL: Distributes the data across different processors.
- HYBRID_PARALLEL: Achieves data parallelism and model parallelism manually.
- SEMI_AUTO_PARALLEL: Achieves data parallelism and model parallelism by setting parallel strategies.
- AUTO_PARALLEL: Achieves parallelism automatically.

MODE_LIST: The list for all supported parallel modes.
MODE_LIST: The list of all supported parallel modes.
"""

STAND_ALONE = "stand_alone"


+ 15
- 0
mindspore/ops/composite/multitype_ops/equal_impl.py View File

@@ -26,6 +26,21 @@ using ".register" decorator
"""


@equal.register("Bool", "Bool")
def _equal_bool(x, y):
"""
Determine if two bool objects are equal.

Args:
x (bool): first input bool object.
y (bool): second input bool object.

Returns:
bool, if x == y return true, x != y return false.
"""
return F.bool_eq(x, y)


@equal.register("Number", "Number")
def _equal_scalar(x, y):
"""


+ 1
- 0
mindspore/ops/functional.py View File

@@ -123,6 +123,7 @@ string_concat = Primitive('string_concat')
bool_not = Primitive("bool_not")
bool_or = Primitive("bool_or")
bool_and = Primitive("bool_and")
bool_eq = Primitive("bool_eq")
logical_and = P.LogicalAnd()
logical_or = P.LogicalOr()
logical_not = P.LogicalNot()


+ 92
- 88
mindspore/ops/operations/array_ops.py View File

@@ -289,7 +289,7 @@ class Cast(PrimitiveWithInfer):

class IsSubClass(PrimitiveWithInfer):
"""
Check whether one type is sub class of another type.
Check whether one type is subtraction class of another type.

Inputs:
- **sub_type** (mindspore.dtype) - The type to be checked. Only constant value is allowed.
@@ -478,7 +478,7 @@ class DynamicShape(Primitive):

class Squeeze(PrimitiveWithInfer):
"""
Returns a tensor with the same type but dimensions of 1 being removed based on axis.
Returns a tensor with the same type but dimensions of 1 are removed based on `axis`.

Note:
The dimension index starts at 0 and must be in the range `[-input.dim(), input.dim())`.
@@ -536,7 +536,7 @@ class Squeeze(PrimitiveWithInfer):

class Transpose(PrimitiveWithInfer):
"""
Permutes the dimensions of input tensor according to input perm.
Permutes the dimensions of input tensor according to input permutation.

Inputs:
- **input_x** (Tensor) - The shape of tensor is :math:`(x_1, x_2, ..., x_R)`.
@@ -545,7 +545,7 @@ class Transpose(PrimitiveWithInfer):
allowed.

Outputs:
Tensor, the type of output tensor is same as `input_x` and the shape of output tensor is decided by the
Tensor, the type of output tensor is the same as `input_x` and the shape of output tensor is decided by the
shape of `input_x` and the value of `input_perm`.

Examples:
@@ -654,7 +654,7 @@ class SparseGatherV2(GatherV2):
- **input_params** (Tensor) - The shape of tensor is :math:`(x_1, x_2, ..., x_R)`.
The original Tensor.
- **input_indices** (Tensor) - The shape of tensor is :math:`(y_1, y_2, ..., y_S)`.
Specifies the indices of elements of the original Tensor. Must be in the range
Specifies the indices of elements of the original Tensor, must be in the range
`[0, input_param.shape[axis])`.
- **axis** (int) - Specifies the dimension index to gather indices.

@@ -718,15 +718,15 @@ class Split(PrimitiveWithInfer):
output_num (int): The number of output tensors. Default: 1.

Raises:
ValueError: If axis is out of the range [-len(input_x.shape), len(input_x.shape)),
or if the output_num is less than or equal to 0, or if the
dimension which to split cannot be evenly divided by output_num.
ValueError: If `axis` is out of the range [-len(`input_x.shape`), len(`input_x.shape`)),
or if the `output_num` is less than or equal to 0, or if the
dimension which to split cannot be evenly divided by `output_num`.

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

Outputs:
tuple[Tensor], the shape of each output tensor is same, which is
tuple[Tensor], the shape of each output tensor is the same, which is
:math:`(y_1, y_2, ..., y_S)`.

Examples:
@@ -806,14 +806,14 @@ class TruncatedNormal(PrimitiveWithInfer):
The generated values follow a normal distribution.

Args:
seed (int): A int number used to create random seed. Default: 0.
seed (int): A integer number used to create random seed. Default: 0.
dtype (:class:`mindspore.dtype`): Data type. Default: mindspore.float32.

Inputs:
- **shape** (tuple[int]) - Shape of output tensor, is a tuple of positive int.
- **shape** (tuple[int]) - The shape of the output tensor, is a tuple of positive integer.

Outputs:
Tensor, type of output tensor is same as attribute `dtype`.
Tensor, the dat type of output tensor is the same as attribute `dtype`.

Examples:
>>> shape = (1, 2, 3)
@@ -953,13 +953,13 @@ class ZerosLike(PrimitiveWithInfer):
"""
Creates a new tensor. All elements value are 0.

Returns a tensor of zeros with the same shape and type as the input tensor.
Returns a tensor of zeros with the same shape and data type as the input tensor.

Inputs:
- **input_x** (Tensor) - Input tensor.

Outputs:
Tensor, has the same shape and type as `input_x` but filled with zeros.
Tensor, has the same shape and data type as `input_x` but filled with zeros.

Examples:
>>> zeroslike = P.ZerosLike()
@@ -982,15 +982,16 @@ class ZerosLike(PrimitiveWithInfer):

class TupleToArray(PrimitiveWithInfer):
"""
Converts a tuple to tensor.
Convert a tuple to a tensor.

If the first number type of tuple is int, the output tensor type is int. Else, the output tensor type is float.
If the type of the first number in the tuple is integer, the data type of the output tensor is int.
Otherwise, the data type of the output tensor is float.

Inputs:
- **input_x** (tuple) - A tuple of numbers. These numbers have the same type. Only constant value is allowed.

Outputs:
Tensor, if the input tuple contain `N` numbers, then the shape of the output tensor is (N,).
Tensor, if the input tuple contains `N` numbers, then the shape of the output tensor is (N,).

Examples:
>>> type = P.TupleToArray()((1,2,3))
@@ -1355,7 +1356,7 @@ class Tile(PrimitiveWithInfer):
Replicates a tensor with given multiples times.

Creates a new tensor by replicating input multiples times. The dimension of
output tensor is the larger of the dimension length of input and the length of multiples.
output tensor is the larger of the input tensor dimension and the length of `multiples`.

Inputs:
- **input_x** (Tensor) - 1-D or higher Tensor. Set the shape of input tensor as
@@ -1363,16 +1364,17 @@ class Tile(PrimitiveWithInfer):

- **multiples** (tuple[int]) - The input tuple is constructed by multiple
integers, i.e., :math:`(y_1, y_2, ..., y_S)`. The length of `multiples`
can't be smaller than the length of shape in `input_x`. Only constant value is allowed.
cannot be smaller than the length of the shape of `input_x`.
Only constant value is allowed.

Outputs:
Tensor, has the same type as the `input_x`.
Tensor, has the same data type as the `input_x`.

- If the length of `multiples` is the same as the length of shape in `input_x`,
- If the length of `multiples` is the same as the length of shape of `input_x`,
then the shape of their corresponding positions can be multiplied, and
the shape of Outputs is :math:`(x_1*y_1, x_2*y_2, ..., x_S*y_R)`.
- If the length of `multiples` is larger than the length of shape in `input_x`,
fill in multiple 1 in front of the shape in `input_x` until their lengths are consistent.
- If the length of `multiples` is larger than the length of shape of `input_x`,
fill in multiple 1 in the length of the shape of `input_x` until their lengths are consistent.
Such as set the shape of `input_x` as :math:`(1, ..., x_1, x_2, ..., x_S)`,
then the shape of their corresponding positions can be multiplied, and
the shape of Outputs is :math:`(1*y_1, ..., x_S*y_R)`.
@@ -1501,7 +1503,7 @@ class UnsortedSegmentMin(PrimitiveWithInfer):
- **num_segments** (int) - The value spcifies the number of distinct `segment_ids`.

Outputs:
Tensor, Set the number of `num_segments` as `N`, the shape is :math:`(N, x_2, ..., x_R)`.
Tensor. Set the number of `num_segments` as `N`, the shape is :math:`(N, x_2, ..., x_R)`.

Examples:
>>> input_x = Tensor(np.array([[1, 2, 3], [4, 5, 6], [4, 2, 1]]).astype(np.float32))
@@ -1552,7 +1554,7 @@ class UnsortedSegmentProd(PrimitiveWithInfer):
should be greater than 0.

Outputs:
Tensor, Set the number of `num_segments` as `N`, the shape is :math:`(N, x_2, ..., x_R)`.
Tensor, set the number of `num_segments` as `N`, the shape is :math:`(N, x_2, ..., x_R)`.

Examples:
>>> input_x = Tensor(np.array([[1, 2, 3], [4, 5, 6], [4, 2, 1]]).astype(np.float32))
@@ -1783,10 +1785,10 @@ class Unpack(PrimitiveWithInfer):

Inputs:
- **input_x** (Tensor) - The shape is :math:`(x_1, x_2, ..., x_R)`.
A rank R > 0 Tensor to be unpacked.
A tensor to be unpacked and the rank of the tensor must be greater than 0.

Outputs:
A tuple of Tensors, the shape of each objects is same.
A tuple of tensors, the shape of each objects is the same.

Raises:
ValueError: If axis is out of the range [-len(input_x.shape), len(input_x.shape)).
@@ -1834,7 +1836,7 @@ class Unpack(PrimitiveWithInfer):

class Slice(PrimitiveWithInfer):
"""
Slice a tensor in specified shape.
Slice a tensor in the specified shape.

Args:
x (Tensor): The target tensor.
@@ -1955,16 +1957,16 @@ class Select(PrimitiveWithInfer):

Given a tensor as input, this operation inserts a dimension of 1 at the dimension,
if both :math:`x` and :math:`y` are none, the operation returns the coordinates of the true
element in the condition, the coordinates are returned as a two-dimensional
element in the `condition`, the coordinates are returned as a two-dimensional
tensor, where the first dimension (row) represents the number of true elements
and the second dimension (columns) represents the coordinates of the true
elements. Keep in mind that the shape of the output tensor can vary depending
on how much of the true value is in the input. Indexes are output in row-first
on how many true values are in the input. Indexes are output in row-first
order.

If neither is None, :math:`x` and :math:`y` must have the same shape. If :math:`x` and :math:`y` are
scalars, the conditional tensor must be a scalar. If :math:`x` and :math:`y` are
higher-demensional vectors, the condition must be a vector whose size matches the
higher-demensional vectors, the `condition` must be a vector whose size matches the
first dimension of :math:`x`, or must have the same shape as :math:`y`.

The conditional tensor acts as an optional compensation (mask), which
@@ -1979,14 +1981,14 @@ class Select(PrimitiveWithInfer):

Inputs:
- **input_x** (Tensor[bool]) - The shape is :math:`(x_1, x_2, ..., x_N)`.
The condition tensor, decides whose element is chosen.
The condition tensor, decides which element is chosen.
- **input_y** (Tensor) - The shape is :math:`(x_1, x_2, ..., x_N, ..., x_R)`.
The first input tensor.
- **input_z** (Tensor) - The shape is :math:`(x_1, x_2, ..., x_N, ..., x_R)`.
The second input tensor.

Outputs:
Tensor, has the same shape as input_y. The shape is :math:`(x_1, x_2, ..., x_N, ..., x_R)`.
Tensor, has the same shape as `input_y`. The shape is :math:`(x_1, x_2, ..., x_N, ..., x_R)`.

Examples:
>>> select = P.Select()
@@ -2080,12 +2082,12 @@ def _compute_slicing_length(begin, end, stride, x_shape, i):
class StridedSlice(PrimitiveWithInfer):
r"""

Extracts a strided slice of a tensor.
Extract a strided slice of a tensor.

Given an input tensor, this operation inserts a dimension of length 1 at the dimension.
This operation extracts a fragment of size (end-begin)/stride from the given
'input_tensor'. Starting from the position specified by the begin, the fragment
continues adding stride to the index until all dimensions are not less than end.
This operation extracts a fragment of size (end-begin)/stride from the given 'input_tensor'.
Starting from the begining position, the fragment continues adding stride to the index until
all dimensions are not less than the ending position.

Note:
The stride may be negative value, which causes reverse slicing.
@@ -2102,22 +2104,22 @@ class StridedSlice(PrimitiveWithInfer):
- **input_x** (Tensor) - The input Tensor.
- **begin** (tuple[int]) - A tuple which represents the location where to start. Only
constant value is allowed.
- **end** (tuple[int]) - A tuple or which represents the maximum location where to stop.
- **end** (tuple[int]) - A tuple or which represents the maximum location where to end.
Only constant value is allowed.
- **strides** (tuple[int]) - A tuple which represents the stride continuously added
before reach the maximum location. Only constant value is allowed.
- **strides** (tuple[int]) - A tuple which represents the stride is continuously added
before reaching the maximum location. Only constant value is allowed.

Outputs:
Tensor.
Explain with the following example.
- In the 0th dim, begin is 1, end is 2, and strides is 1,
The output is explained by following example.
- In the 0th dimension, begin is 1, end is 2, and strides is 1,
because :math:`1+1=2\geq2`, the interval is :math:`[1,2)`.
Thus, return the element with :math:`index = 1` in 0th dim, i.e., [[3, 3, 3], [4, 4, 4]].
- In the 1st dim, similarly, the interval is :math:`[0,1)`.
Based on the return value of the 0th dim, return the element with :math:`index = 0`,
Thus, return the element with :math:`index = 1` in 0th dimension, i.e., [[3, 3, 3], [4, 4, 4]].
- In the 1st dimension, similarly, the interval is :math:`[0,1)`.
Based on the return value of the 0th dimension, return the element with :math:`index = 0`,
i.e., [3, 3, 3].
- In the 2nd dim, similarly, the interval is :math:`[0,3)`.
Based on the return value of the 1st dim, return the element with :math:`index = 0,1,2`,
- In the 2nd dimension, similarly, the interval is :math:`[0,3)`.
Based on the return value of the 1st dimension, return the element with :math:`index = 0,1,2`,
i.e., [3, 3, 3].
- Finally, the output is [3, 3, 3].

@@ -2358,7 +2360,7 @@ class DiagPart(PrimitiveWithInfer):
class Eye(PrimitiveWithInfer):
"""

Creates a tensor with ones on the diagonal and zeros the rest.
Create a tensor with ones on the diagonal and zeros the rest.

Inputs:
- **n** (int) - The number of rows of returned tensor
@@ -2391,12 +2393,12 @@ class Eye(PrimitiveWithInfer):

class ScatterNd(PrimitiveWithInfer):
"""
Scatters a tensor into a new tensor depending on the specified indices.
Scatter a tensor into a new tensor depending on the specified indices.

Creates an empty tensor, and set values by scattering the update tensor depending on indices.
Create an empty tensor, and set values by scattering the update tensor depending on indices.

Inputs:
- **indices** (Tensor) - The index of scattering in the new tensor. With int32 data type.
- **indices** (Tensor) - The index of scattering in the new tensor with int32 data type.
- **update** (Tensor) - The source Tensor to be scattered.
- **shape** (tuple[int]) - Define the shape of the output tensor, has the same type as indices.

@@ -2480,7 +2482,7 @@ class ResizeNearestNeighbor(PrimitiveWithInfer):

class GatherNd(PrimitiveWithInfer):
"""
Gathers slices from a tensor by indices.
Gather slices from a tensor by indices.

Using given indices to gather slices from a tensor with a specified shape.

@@ -2516,9 +2518,7 @@ class GatherNd(PrimitiveWithInfer):

class TensorScatterUpdate(PrimitiveWithInfer):
"""
Update tensor value by using input indices and value.

Using given values to update tensor value, along with the input indices.
Update tensor value using given values, along with the input indices.

Inputs:
- **input_x** (Tensor) - The target tensor.
@@ -2619,7 +2619,7 @@ class ScatterNdUpdate(_ScatterNdOp):
Inputs:
- **input_x** (Parameter) - The target tensor, with data type of Parameter.
- **indices** (Tensor) - The index of input tensor, with int32 data type.
- **update** (Tensor) - The tensor to add to the input tensor, has the same type as input.
- **update** (Tensor) - The tensor to be updated to the input tensor, has the same type as input.

Outputs:
Tensor, has the same shape and type as `input_x`.
@@ -2757,9 +2757,9 @@ class ScatterAdd(_ScatterOp):

class ScatterSub(_ScatterOp):
"""
Update the value of the input tensor through the sub operation.
Update the value of the input tensor through the subtraction operation.

Using given values to update tensor value through the sub operation, along with the input indices.
Using given values to update tensor value through the subtraction operation, along with the input indices.
This operation outputs the `input_x` after the update is done, which makes it convenient to use the updated value.

Inputs of `input_x` and `updates` comply with the implicit type conversion rules to make the data types consistent.
@@ -2772,9 +2772,10 @@ class ScatterSub(_ScatterOp):

Inputs:
- **input_x** (Parameter) - The target parameter.
- **indices** (Tensor) - The index to do sub operation whose data type should be mindspore.int32.
- **updates** (Tensor) - The tensor doing the sub operation with `input_x`,
the data type is same as `input_x`, the shape is `indices_shape + x_shape[1:]`.
- **indices** (Tensor) - The index to perform the subtraction operation
whose data type should be mindspore.int32.
- **updates** (Tensor) - The tensor that performs the subtraction operation with `input_x`,
the data type is the same as `input_x`, the shape is `indices_shape + x_shape[1:]`.

Outputs:
Parameter, the updated `input_x`.
@@ -2895,7 +2896,7 @@ class ScatterNdSub(_ScatterNdOp):
"""
Applies sparse subtraction to individual values or slices in a Tensor.

Using given values to update tensor value through the sub operation, along with the input indices.
Using given values to update tensor value through the subtraction operation, along with the input indices.
This operation outputs the `input_x` after the update is done, which makes it convenient to use the updated value.

Inputs of `input_x` and `updates` comply with the implicit type conversion rules to make the data types consistent.
@@ -2909,8 +2910,8 @@ class ScatterNdSub(_ScatterNdOp):
Inputs:
- **input_x** (Parameter) - The target parameter.
- **indices** (Tensor) - The index to do add operation whose data type should be mindspore.int32.
- **updates** (Tensor) - The tensor doing the sub operation with `input_x`,
the data type is same as `input_x`, the shape is `indices_shape[:-1] + x_shape[indices_shape[-1]:]`.
- **updates** (Tensor) - The tensor that performs the subtraction operation with `input_x`,
the data type is the same as `input_x`, the shape is `indices_shape[:-1] + x_shape[indices_shape[-1]:]`.

Outputs:
Parameter, the updated `input_x`.
@@ -2939,9 +2940,9 @@ class ScatterNonAliasingAdd(_ScatterNdOp):

Inputs:
- **input_x** (Parameter) - The target parameter. The data type should be float16, float32 or int32.
- **indices** (Tensor) - The index to do add operation whose data type should be mindspore.int32.
- **updates** (Tensor) - The tensor doing the add operation with `input_x`,
the data type is same as `input_x`, the shape is `indices_shape[:-1] + x_shape[indices_shape[-1]:]`.
- **indices** (Tensor) - The index to perform the addition operation whose data type should be mindspore.int32.
- **updates** (Tensor) - The tensor that performs the addition operation with `input_x`,
the data type is the same as `input_x`, the shape is `indices_shape[:-1] + x_shape[indices_shape[-1]:]`.

Outputs:
Parameter, the updated `input_x`.
@@ -2987,7 +2988,7 @@ class SpaceToDepth(PrimitiveWithInfer):
- **x** (Tensor) - The target tensor.

Outputs:
Tensor, the same type as `x`. It must be a 4-D tensor.
Tensor, the same data type as `x`. It must be a 4-D tensor.

Examples:
>>> x = Tensor(np.random.rand(1,3,2,2), mindspore.float32)
@@ -3083,24 +3084,26 @@ class SpaceToBatch(PrimitiveWithInfer):
r"""
Divide spatial dimensions into blocks and combine the block size with the original batch.

This operation will divide spatial dimensions (H, W) into blocks with block_size, the output tensor's H and W
This operation will divide spatial dimensions (H, W) into blocks with `block_size`, the output tensor's H and W
dimension is the corresponding number of blocks after division. The output tensor's batch dimension is the
product of the original batch and the square of block_size. Prior to division into blocks, the spatial dimensions
product of the original batch and the square of block_size. Before division, the spatial dimensions
of the input are zero padded according to paddings if necessary.

Args:
block_size (int): The block size of division, has the value not less than 2.
paddings (list): The padding value for H and W dimension, containing 2 sub list, each containing 2 int value.
All values must be >= 0. paddings[i] specifies the paddings for spatial dimension i, which corresponds to
input dimension i+2. It is required that input_shape[i+2]+paddings[i][0]+paddings[i][1] is divisible
by block_size.
block_size (int): The block size of dividing blocks with value greater than 2.
paddings (list): The padding values for H and W dimension, containing 2 subtraction lists.
Each subtraction list contains 2 integer value. All values must be greater than 0.
paddings[i] specifies the paddings for the spatial dimension i, which corresponds to the
input dimension i+2. It is required that input_shape[i+2]+paddings[i][0]+paddings[i][1]
is divisible by block_size.

Inputs:
- **input_x** (Tensor) - The input tensor. It must be a 4-D tensor.

Outputs:
Tensor, the output tensor with the same type as input. Assume input shape is :math:`(n, c, h, w)` with
:math:`block\_size` and :math:`paddings`. The shape of the output tensor will be :math:`(n', c', h', w')`, where
Tensor, the output tensor with the same data type as input. Assume input shape is :math:`(n, c, h, w)` with
:math:`block\_size` and :math:`paddings`. The shape of the output tensor will be :math:`(n', c', h', w')`,
where

:math:`n' = n*(block\_size*block\_size)`

@@ -3159,7 +3162,7 @@ class BatchToSpace(PrimitiveWithInfer):

Args:
block_size (int): The block size of division, has the value not less than 2.
crops (Union[list(int), tuple(int)]): The crop value for H and W dimension, containing 2 sub lists.
crops (Union[list(int), tuple(int)]): The crop value for H and W dimension, containing 2 subtraction lists.
Each list contains 2 integers.
All values must be not less than 0. crops[i] specifies the crop values for the spatial dimension i, which
corresponds to the input dimension i+2. It is required that
@@ -3230,21 +3233,22 @@ class SpaceToBatchND(PrimitiveWithInfer):

This operation will divide spatial dimensions (H, W) into blocks with block_shape, the output tensor's H and W
dimension is the corresponding number of blocks after division. The output tensor's batch dimension is the
product of the original batch and the product of block_shape. Prior to division into blocks, the spatial dimensions
of the input are zero padded according to paddings if necessary.
product of the original batch and the product of `block_shape`. Before division,
the spatial dimensions of the input are zero padded according to paddings if necessary.

Args:
block_shape (Union[list(int), tuple(int)]): The block shape of dividing block with all value >= 1.
The length of block_shape is M correspoding to the number of spatial dimensions.
paddings (list): The padding value for H and W dimension, containing M sub list, each containing 2 int value.
All values must be >= 0. paddings[i] specifies the paddings for spatial dimension i, which corresponds to
input dimension i+2. It is required that input_shape[i+2]+paddings[i][0]+paddings[i][1] is divisible
by block_shape[i].
block_shape (Union[list(int), tuple(int)]): The block shape of dividing block with all value greater than 1.
The length of `block_shape` is M correspoding to the number of spatial dimensions.
paddings (list): The padding values for H and W dimension, containing M subtraction list.
Each contains 2 integer value. All values must be greater than 0.
`paddings[i]` specifies the paddings for the spatial dimension i,
which corresponds to the input dimension i+2.
It is required that input_shape[i+2]+paddings[i][0]+paddings[i][1] is divisible by block_shape[i].

Inputs:
- **input_x** (Tensor) - The input tensor. It must be a 4-D tensor.
Outputs:
Tensor, the output tensor with the same type as input. Assume input shape is :math:`(n, c, h, w)` with
Tensor, the output tensor with the same data type as input. Assume input shape is :math:`(n, c, h, w)` with
:math:`block\_shape` and :math:`padddings`. The shape of the output tensor will be :math:`(n', c', h', w')`,
where

@@ -3321,7 +3325,7 @@ class BatchToSpaceND(PrimitiveWithInfer):
Args:
block_shape (Union[list(int), tuple(int)]): The block shape of dividing block with all value >= 1.
The length of block_shape is M correspoding to the number of spatial dimensions.
crops (Union[list(int), tuple(int)]): The crop value for H and W dimension, containing 2 sub list,
crops (Union[list(int), tuple(int)]): The crop value for H and W dimension, containing 2 subtraction list,
each containing 2 int value.
All values must be >= 0. crops[i] specifies the crop values for spatial dimension i, which corresponds to
input dimension i+2. It is required that input_shape[i+2]*block_shape[i] > crops[i][0]+crops[i][1].


+ 1
- 1
mindspore/ops/operations/debug_ops.py View File

@@ -117,7 +117,7 @@ class ImageSummary(PrimitiveWithInfer):

class TensorSummary(PrimitiveWithInfer):
"""
Output tensor to protocol buffer through tensor summary operator.
Output a tensor to a protocol buffer through a tensor summary operator.

Inputs:
- **name** (str) - The name of the input variable.


+ 26
- 26
mindspore/ops/operations/math_ops.py View File

@@ -125,10 +125,10 @@ class TensorAdd(_MathBinaryOp):
the scalar could only be a constant.

Inputs:
- **input_x** (Union[Tensor, Number, bool]) - The first input is a number or
a bool or a tensor whose data type is number or bool.
- **input_y** (Union[Tensor, Number, bool]) - The second input is a number or
a bool when the first input is a tensor or a tensor whose data type is number or bool.
- **input_x** (Union[Tensor, Number, bool]) - The first input is a number, or a bool,
or a tensor whose data type is number or bool.
- **input_y** (Union[Tensor, Number, bool]) - The second input is a number,
or a bool when the first input is a tensor, or a tensor whose data type is number or bool.

Outputs:
Tensor, the shape is the same as the one after broadcasting,
@@ -1079,10 +1079,10 @@ class Sub(_MathBinaryOp):
the scalar could only be a constant.

Inputs:
- **input_x** (Union[Tensor, Number, bool]) - The first input is a number or
a bool or a tensor whose data type is number or bool.
- **input_y** (Union[Tensor, Number, bool]) - The second input is a number or
a bool when the first input is a tensor or a tensor whose data type is number or bool.
- **input_x** (Union[Tensor, Number, bool]) - The first input is a number, or a bool,
or a tensor whose data type is number or bool.
- **input_y** (Union[Tensor, Number, bool]) - The second input is a number,
or a bool when the first input is a tensor, or a tensor whose data type is number or bool.

Outputs:
Tensor, the shape is the same as the one after broadcasting,
@@ -1157,10 +1157,10 @@ class SquaredDifference(_MathBinaryOp):
the scalar could only be a constant.

Inputs:
- **input_x** (Union[Tensor, Number, bool]) - The first input is a number or
a bool or a tensor whose data type is float16, float32, int32 or bool.
- **input_y** (Union[Tensor, Number, bool]) - The second input is a number or
a bool when the first input is a tensor or a tensor whose data type is
- **input_x** (Union[Tensor, Number, bool]) - The first input is a number, or a bool,
or a tensor whose data type is float16, float32, int32 or bool.
- **input_y** (Union[Tensor, Number, bool]) - The second input is a number,
or a bool when the first input is a tensor or a tensor whose data type is
float16, float32, int32 or bool.

Outputs:
@@ -1863,10 +1863,10 @@ class TruncateDiv(_MathBinaryOp):
the scalar could only be a constant.

Inputs:
- **input_x** (Union[Tensor, Number, bool]) - The first input is a number or
a bool or a tensor whose data type is number or bool.
- **input_y** (Union[Tensor, Number, bool]) - The second input is a number or
a bool when the first input is a tensor or a tensor whose data type is number or bool.
- **input_x** (Union[Tensor, Number, bool]) - The first input is a number, or a bool,
or a tensor whose data type is number or bool.
- **input_y** (Union[Tensor, Number, bool]) - The second input is a number,
or a bool when the first input is a tensor, or a tensor whose data type is number or bool.

Outputs:
Tensor, the shape is the same as the one after broadcasting,
@@ -1893,10 +1893,10 @@ class TruncateMod(_MathBinaryOp):
the scalar could only be a constant.

Inputs:
- **input_x** (Union[Tensor, Number, bool]) - The first input is a number or
a bool or a tensor whose data type is number or bool.
- **input_y** (Union[Tensor, Number, bool]) - The second input is a number or
a bool when the first input is a tensor or a tensor whose data type is number or bool.
- **input_x** (Union[Tensor, Number, bool]) - The first input is a number, or a bool,
or a tensor whose data type is number or bool.
- **input_y** (Union[Tensor, Number, bool]) - The second input is a number,
or a bool when the first input is a tensor, or a tensor whose data type is number or bool.

Outputs:
Tensor, the shape is the same as the one after broadcasting,
@@ -2048,10 +2048,10 @@ class Xdivy(_MathBinaryOp):
the scalar could only be a constant.

Inputs:
- **input_x** (Union[Tensor, Number, bool]) - The first input is a number or
a bool or a tensor whose data type is float16, float32 or bool.
- **input_y** (Union[Tensor, Number, bool]) - The second input is a number or
a bool when the first input is a tensor or a tensor whose data type is float16, float32 or bool.
- **input_x** (Union[Tensor, Number, bool]) - The first input is a number, or a bool,
or a tensor whose data type is float16, float32 or bool.
- **input_y** (Union[Tensor, Number, bool]) - The second input is a number,
or a bool when the first input is a tensor, or a tensor whose data type is float16, float32 or bool.

Outputs:
Tensor, the shape is the same as the one after broadcasting,
@@ -3069,7 +3069,7 @@ class Sign(PrimitiveWithInfer):
Note:
.. math::
sign(x) = \begin{cases} -1, &if\ x < 0 \cr
0, &if\ x == 0 \cr
0, &if\ x = 0 \cr
1, &if\ x > 0\end{cases}

Inputs:
@@ -3251,7 +3251,7 @@ class SquareSumAll(PrimitiveWithInfer):

Inputs:
- **input_x1** (Tensor) - The input tensor. The data type must be float16 or float32.
- **input_x2** (Tensor) - The input tensor same type and shape as the `input_x1`.
- **input_x2** (Tensor) - The input tensor has the same type and shape as the `input_x1`.

Note:
SquareSumAll only supports float16 and float32 data type.


+ 37
- 37
mindspore/ops/operations/nn_ops.py View File

@@ -98,7 +98,7 @@ class Softmax(PrimitiveWithInfer):
Softmax operation.

Applies the Softmax operation to the input tensor on the specified axis.
Suppose a slice in the given aixs :math:`x` then for each element :math:`x_i`
Suppose a slice in the given aixs :math:`x`, then for each element :math:`x_i`,
the Softmax function is shown as follows:

.. math::
@@ -107,7 +107,7 @@ class Softmax(PrimitiveWithInfer):
where :math:`N` is the length of the tensor.

Args:
axis (Union[int, tuple]): The axis to do the Softmax operation. Default: -1.
axis (Union[int, tuple]): The axis to perform the Softmax operation. Default: -1.

Inputs:
- **logits** (Tensor) - The input of Softmax, with float16 or float32 data type.
@@ -1549,17 +1549,17 @@ class TopK(PrimitiveWithInfer):
Finds values and indices of the `k` largest entries along the last dimension.

Args:
sorted (bool): If true, the resulting elements will
sorted (bool): If True, the obtained elements will
be sorted by the values in descending order. Default: False.

Inputs:
- **input_x** (Tensor) - Input to be computed, data type should be float16, float32 or int32.
- **k** (int) - Number of top elements to be computed along the last dimension, constant input is needed.
- **k** (int) - The number of top elements to be computed along the last dimension, constant input is needed.

Outputs:
Tuple of 2 Tensors, the values and the indices.
Tuple of 2 tensors, the values and the indices.

- **values** (Tensor) - The `k` largest elements along each last dimensional slice.
- **values** (Tensor) - The `k` largest elements in each slice of the last dimensional.
- **indices** (Tensor) - The indices of values within the last dimension of input.

Examples:
@@ -1593,7 +1593,7 @@ class TopK(PrimitiveWithInfer):

class SoftmaxCrossEntropyWithLogits(PrimitiveWithInfer):
r"""
Gets the softmax cross-entropy value between logits and labels which shoule be one-hot encoding.
Gets the softmax cross-entropy value between logits and labels with one-hot encoding.

Note:
Sets input logits as `X`, input label as `Y`, output as `loss`. Then,
@@ -1609,7 +1609,7 @@ class SoftmaxCrossEntropyWithLogits(PrimitiveWithInfer):
- **labels** (Tensor) - Ground truth labels, with shape :math:`(N, C)`, has the same data type with `logits`.

Outputs:
Tuple of 2 Tensors, the loss shape is `(N,)`, and the dlogits with the same shape as `logits`.
Tuple of 2 tensors, the `loss` shape is `(N,)`, and the `dlogits` with the same shape as `logits`.

Examples:
>>> logits = Tensor([[2, 4, 1, 4, 5], [2, 1, 2, 4, 3]], mindspore.float32)
@@ -1653,7 +1653,7 @@ class SparseSoftmaxCrossEntropyWithLogits(PrimitiveWithInfer):
loss = \sum_{ij} loss_{ij}

Args:
is_grad (bool): If it's true, this operation returns the computed gradient. Default: False.
is_grad (bool): If true, this operation returns the computed gradient. Default: False.

Inputs:
- **logits** (Tensor) - Input logits, with shape :math:`(N, C)`. Data type should be float16 or float32.
@@ -4084,19 +4084,19 @@ class SparseApplyAdagrad(PrimitiveWithInfer):
Args:
lr (float): Learning rate.
update_slots (bool): If `True`, `accum` will be updated. Default: True.
use_locking (bool): If true, the var and accumulation tensors will be protected from being updated.
use_locking (bool): If true, the `var` and `accumulation` tensors will be protected from being updated.
Default: False.

Inputs:
- **var** (Parameter) - Variable to be updated. The data type must be float16 or float32.
- **accum** (Parameter) - Accumulation to be updated. The shape and dtype should be the same as `var`.
- **grad** (Tensor) - Gradient. The shape must be the same as `var`'s shape except first dimension.
Has the same data type as `var`.
- **accum** (Parameter) - Accumulation to be updated. The shape and data type should be the same as `var`.
- **grad** (Tensor) - Gradient. The shape must be the same as `var`'s shape except the first dimension.
Gradients has the same data type as `var`.
- **indices** (Tensor) - A vector of indices into the first dimension of `var` and `accum`.
The shape of `indices` must be the same as `grad` in first dimension, the type must be int32.

Outputs:
Tuple of 2 Tensors, the updated parameters.
Tuple of 2 tensors, the updated parameters.

- **var** (Tensor) - The same shape and data type as `var`.
- **accum** (Tensor) - The same shape and data type as `accum`.
@@ -4170,20 +4170,20 @@ class SparseApplyAdagradV2(PrimitiveWithInfer):
Args:
lr (float): Learning rate.
epsilon (float): A small value added for numerical stability.
use_locking (bool): If `True`, the var and accumulation tensors will be protected from being updated.
use_locking (bool): If `True`, the `var` and `accum` tensors will be protected from being updated.
Default: False.
update_slots (bool): If `True`, the computation logic will be different to `False`. Default: True.

Inputs:
- **var** (Parameter) - Variable to be updated. The data type must be float16 or float32.
- **accum** (Parameter) - Accumulation to be updated. The shape and dtype should be the same as `var`.
- **grad** (Tensor) - Gradient. The shape must be the same as `var`'s shape except first dimension.
Has the same data type as `var`.
- **accum** (Parameter) - Accumulation to be updated. The shape and data type should be the same as `var`.
- **grad** (Tensor) - Gradient. The shape must be the same as `var`'s shape except the first dimension.
Gradients has the same data type as `var`.
- **indices** (Tensor) - A vector of indices into the first dimension of `var` and `accum`.
The shape of `indices` must be the same as `grad` in first dimension, the type must be int32.

Outputs:
Tuple of 2 Tensors, the updated parameters.
Tuple of 2 tensors, the updated parameters.

- **var** (Tensor) - The same shape and data type as `var`.
- **accum** (Tensor) - The same shape and data type as `accum`.
@@ -4361,23 +4361,23 @@ class SparseApplyProximalAdagrad(PrimitiveWithCheck):
RuntimeError exception will be thrown when the data type conversion of Parameter is required.

Args:
use_locking (bool): If true, the var and accumulation tensors will be protected from being updated.
use_locking (bool): If true, the `var` and `accum` tensors will be protected from being updated.
Default: False.

Inputs:
- **var** (Parameter) - Variable tensor to be updated. The data type must be float16 or float32.
- **accum** (Parameter) - Variable tensor to be updated, has the same dtype as `var`.
- **lr** (Union[Number, Tensor]) - The learning rate value. Tshould be a float number or
- **lr** (Union[Number, Tensor]) - The learning rate value. should be a float number or
a scalar tensor with float16 or float32 data type.
- **l1** (Union[Number, Tensor]) - l1 regularization strength. should be a float number or
a scalar tensor with float16 or float32 data type.
- **l2** (Union[Number, Tensor]) - l2 regularization strength. should be a float number or
a scalar tensor with float16 or float32 data type..
- **grad** (Tensor) - A tensor of the same type as `var`, for the gradient.
- **indices** (Tensor) - A vector of indices into the first dimension of `var` and `accum`.
- **indices** (Tensor) - A vector of indices in the first dimension of `var` and `accum`.

Outputs:
Tuple of 2 Tensors, the updated parameters.
Tuple of 2 tensors, the updated parameters.

- **var** (Tensor) - The same shape and data type as `var`.
- **accum** (Tensor) - The same shape and data type as `accum`.
@@ -4982,16 +4982,16 @@ class SparseApplyFtrl(PrimitiveWithCheck):

Inputs:
- **var** (Parameter) - The variable to be updated. The data type must be float16 or float32.
- **accum** (Parameter) - The accumulation to be updated, must be same type and shape as `var`.
- **linear** (Parameter) - the linear coefficient to be updated, must be same type and shape as `var`.
- **accum** (Parameter) - The accumulation to be updated, must be same data type and shape as `var`.
- **linear** (Parameter) - the linear coefficient to be updated, must be the same data type and shape as `var`.
- **grad** (Tensor) - A tensor of the same type as `var`, for the gradient.
- **indices** (Tensor) - A vector of indices into the first dimension of `var` and `accum`.
The shape of `indices` must be the same as `grad` in first dimension. The type must be int32.
- **indices** (Tensor) - A vector of indices in the first dimension of `var` and `accum`.
The shape of `indices` must be the same as `grad` in the first dimension. The type must be int32.

Outputs:
- **var** (Tensor) - Tensor, has the same shape and type as `var`.
- **accum** (Tensor) - Tensor, has the same shape and type as `accum`.
- **linear** (Tensor) - Tensor, has the same shape and type as `linear`.
- **var** (Tensor) - Tensor, has the same shape and data type as `var`.
- **accum** (Tensor) - Tensor, has the same shape and data type as `accum`.
- **linear** (Tensor) - Tensor, has the same shape and data type as `linear`.

Examples:
>>> import mindspore
@@ -5074,18 +5074,18 @@ class SparseApplyFtrlV2(PrimitiveWithInfer):

Inputs:
- **var** (Parameter) - The variable to be updated. The data type must be float16 or float32.
- **accum** (Parameter) - The accumulation to be updated, must be same type and shape as `var`.
- **linear** (Parameter) - the linear coefficient to be updated, must be same type and shape as `var`.
- **accum** (Parameter) - The accumulation to be updated, must be same data type and shape as `var`.
- **linear** (Parameter) - the linear coefficient to be updated, must be same data type and shape as `var`.
- **grad** (Tensor) - A tensor of the same type as `var`, for the gradient.
- **indices** (Tensor) - A vector of indices into the first dimension of `var` and `accum`.
The shape of `indices` must be the same as `grad` in first dimension. The type must be int32.
- **indices** (Tensor) - A vector of indices in the first dimension of `var` and `accum`.
The shape of `indices` must be the same as `grad` in the first dimension. The type must be int32.

Outputs:
Tuple of 3 Tensor, the updated parameters.

- **var** (Tensor) - Tensor, has the same shape and type as `var`.
- **accum** (Tensor) - Tensor, has the same shape and type as `accum`.
- **linear** (Tensor) - Tensor, has the same shape and type as `linear`.
- **var** (Tensor) - Tensor, has the same shape and data type as `var`.
- **accum** (Tensor) - Tensor, has the same shape and data type as `accum`.
- **linear** (Tensor) - Tensor, has the same shape and data type as `linear`.

Examples:
>>> import mindspore


+ 7
- 7
mindspore/ops/operations/random_ops.py View File

@@ -34,7 +34,7 @@ class StandardNormal(PrimitiveWithInfer):
- **shape** (tuple) - The shape of random tensor to be generated. Only constant value is allowed.

Outputs:
Tensor. The shape that the input 'shape' denotes. The dtype is float32.
Tensor. The shape is the same as the input `shape`. The dtype is float32.

Examples:
>>> shape = (4, 16)
@@ -239,13 +239,13 @@ class UniformInt(PrimitiveWithInfer):

Inputs:
- **shape** (tuple) - The shape of random tensor to be generated. Only constant value is allowed.
- **minval** (Tensor) - The a distribution parameter.
It defines the minimum possibly generated value. With int32 data type. Only one number is supported.
- **maxval** (Tensor) - The b distribution parameter.
It defines the maximum possibly generated value. With int32 data type. Only one number is supported.
- **minval** (Tensor) - The distribution parameter, a.
It defines the minimum possibly generated value, with int32 data type. Only one number is supported.
- **maxval** (Tensor) - The distribution parameter, b.
It defines the maximum possibly generated value, with int32 data type. Only one number is supported.

Outputs:
Tensor. The shape that the input 'shape' denotes. The dtype is int32.
Tensor. The shape is the same as the input 'shape', and the data type is int32.

Examples:
>>> shape = (4, 16)
@@ -284,7 +284,7 @@ class UniformInt(PrimitiveWithInfer):

class UniformReal(PrimitiveWithInfer):
r"""
Produces random floating-point values i, uniformly distributed on the interval [0, 1).
Produces random floating-point values i, uniformly distributed to the interval [0, 1).

Args:
seed (int): Random seed. Must be non-negative. Default: 0.


+ 2
- 2
mindspore/ops/operations/sparse_ops.py View File

@@ -29,10 +29,10 @@ class SparseToDense(PrimitiveWithInfer):
Inputs:
- **indices** (Tensor) - The indices of sparse representation.
- **values** (Tensor) - Values corresponding to each row of indices.
- **dense_shape** (tuple) - A int tuple which specifies the shape of dense tensor.
- **dense_shape** (tuple) - An int tuple which specifies the shape of dense tensor.

Returns:
Tensor, the shape of tensor is dense_shape.
Tensor, the shape of tensor is `dense_shape`.

Examples:
>>> indices = Tensor([[0, 1], [1, 2]])


+ 7
- 1
tests/ut/python/runtest.sh View File

@@ -62,7 +62,13 @@ else
exit ${RET}
fi

pytest -n 4 --dist=loadfile -v $CURRPATH/parallel $CURRPATH/train $CURRPATH/ops
pytest -n 4 --dist=loadfile -v $CURRPATH/parallel $CURRPATH/train
RET=$?
if [ ${RET} -ne 0 ]; then
exit ${RET}
fi

pytest -n 2 --dist=loadfile -v $CURRPATH/ops
RET=$?
if [ ${RET} -ne 0 ]; then
exit ${RET}


Write Preview
Loading…
Cancel
Save