diff --git a/mindspore/common/initializer.py b/mindspore/common/initializer.py index 4261621272..d55e03314d 100644 --- a/mindspore/common/initializer.py +++ b/mindspore/common/initializer.py @@ -276,7 +276,7 @@ def initializer(init, shape=None, dtype=mstype.float32): shape (Union[tuple, list, int]): A list of integers, a tuple of integers or an integer as the shape of output. Default: None. - dtype (:class:`mindspore.dtype`): The type of data in initialized tensor. Default: mstype.float32. + dtype (:class:`mindspore.dtype`): The type of data in initialized tensor. Default: mindspore.float32. Returns: Tensor, initialized tensor. diff --git a/mindspore/ops/operations/array_ops.py b/mindspore/ops/operations/array_ops.py index 59d3083c5d..dda490566f 100644 --- a/mindspore/ops/operations/array_ops.py +++ b/mindspore/ops/operations/array_ops.py @@ -62,7 +62,7 @@ class ExpandDims(PrimitiveWithInfer): Examples: >>> input_tensor = Tensor(np.array([[2, 2], [2, 2]]), mindspore.float32) - >>> expand_dims = ExpandDims() + >>> expand_dims = P.ExpandDims() >>> output = expand_dims(input_tensor, 0) """ @@ -101,7 +101,7 @@ class DType(PrimitiveWithInfer): Examples: >>> input_tensor = Tensor(np.array([[2, 2], [2, 2]]), mindspore.float32) - >>> type = DType()(input_tensor) + >>> type = P.DType()(input_tensor) """ @prim_attr_register @@ -134,7 +134,7 @@ class SameTypeShape(PrimitiveWithInfer): Examples: >>> input_x = Tensor(np.array([[2, 2], [2, 2]]), mindspore.float32) >>> input_y = Tensor(np.array([[2, 2], [2, 2]]), mindspore.float32) - >>> out = SameTypeShape()(input_x, input_y) + >>> out = P.SameTypeShape()(input_x, input_y) """ @prim_attr_register @@ -175,7 +175,7 @@ class Cast(PrimitiveWithInfer): >>> input_np = np.random.randn(2, 3, 4, 5).astype(np.float32) >>> input_x = Tensor(input_np) >>> type_dst = mindspore.int32 - >>> cast = Cast() + >>> cast = P.Cast() >>> result = cast(input_x, type_dst) >>> expect = input_np.astype(type_dst) """ @@ -227,7 +227,7 @@ class IsSubClass(PrimitiveWithInfer): bool, the check result. Examples: - >>> result = IsSubClass()(mindspore.int32, mindspore.intc) + >>> result = P.IsSubClass()(mindspore.int32, mindspore.intc) """ @prim_attr_register @@ -262,7 +262,7 @@ class IsInstance(PrimitiveWithInfer): Examples: >>> a = 1 - >>> result = IsInstance()(a, mindspore.int32) + >>> result = P.IsInstance()(a, mindspore.int32) """ @prim_attr_register @@ -303,7 +303,7 @@ class Reshape(PrimitiveWithInfer): Examples: >>> input_tensor = Tensor(np.array([[-0.1, 0.3, 3.6], [0.4, 0.5, -3.2]]), mindspore.float32) - >>> reshape = Reshape() + >>> reshape = P.Reshape() >>> output = reshape(input_tensor, (3, 2)) """ @@ -366,7 +366,7 @@ class Shape(Primitive): Examples: >>> input_tensor = Tensor(np.ones(shape=[3, 2, 1]), mindspore.float32) - >>> shape = Shape() + >>> shape = P.Shape() >>> output = shape(input_tensor) """ @@ -398,7 +398,7 @@ class Squeeze(PrimitiveWithInfer): Examples: >>> input_tensor = Tensor(np.ones(shape=[3, 2, 1]), mindspore.float32) - >>> squeeze = Squeeze(2) + >>> squeeze = P.Squeeze(2) >>> output = squeeze(input_tensor) """ @@ -450,7 +450,7 @@ class Transpose(PrimitiveWithInfer): Examples: >>> input_tensor = Tensor(np.array([[[1, 2, 3], [4, 5, 6]], [[7, 8, 9], [10, 11, 12]]]), mindspore.float32) >>> perm = (0, 2, 1) - >>> transpose = Transpose() + >>> transpose = P.Transpose() >>> output = transpose(input_tensor, perm) """ @@ -504,10 +504,10 @@ class GatherV2(PrimitiveWithInfer): Tensor, the shape of tensor is :math:`(z_1, z_2, ..., z_N)`. Examples: - >>> params = Tensor(np.array([[1, 2, 7, 42], [3, 4, 54, 22], [2, 2, 55, 3]]), mindspore.float32) - >>> indices = Tensor(np.array([1, 2]), mindspore.int32) + >>> input_params = Tensor(np.array([[1, 2, 7, 42], [3, 4, 54, 22], [2, 2, 55, 3]]), mindspore.float32) + >>> input_indices = Tensor(np.array([1, 2]), mindspore.int32) >>> axis = 1 - >>> out = GatherV2()(params, indices, axis) + >>> out = P.GatherV2()(input_params, input_indices, axis) """ @prim_attr_register @@ -556,7 +556,7 @@ class Split(PrimitiveWithInfer): :math:`(y_1, y_2, ..., y_S)`. Examples: - >>> split = Split(1, 2) + >>> split = P.Split(1, 2) >>> x = Tensor(np.array([[1, 1, 1, 1], [2, 2, 2, 2]])) >>> output = split(x) """ @@ -606,7 +606,7 @@ class Rank(PrimitiveWithInfer): Examples: >>> input_tensor = Tensor(np.array([[2, 2], [2, 2]]), mindspore.float32) - >>> rank = Rank() + >>> rank = P.Rank() >>> rank(input_tensor) """ @@ -640,7 +640,7 @@ class TruncatedNormal(PrimitiveWithInfer): Examples: >>> input_shape = Tensor(np.array([1, 2, 3])) - >>> truncated_normal = TruncatedNormal() + >>> truncated_normal = P.TruncatedNormal() >>> output = truncated_normal(input_shape) """ @@ -681,7 +681,7 @@ class Size(PrimitiveWithInfer): Examples: >>> input_tensor = Tensor(np.array([[2, 2], [2, 2]]), mindspore.float32) - >>> size = Size() + >>> size = P.Size() >>> output = size(input_tensor) """ @@ -826,7 +826,7 @@ class TupleToArray(PrimitiveWithInfer): Tensor, if the input tuple contain `N` numbers, then the output tensor shape is (N,). Examples: - >>> type = TupleToArray()((1,2,3)) + >>> type = P.TupleToArray()((1,2,3)) """ @prim_attr_register @@ -861,7 +861,7 @@ class ScalarToArray(PrimitiveWithInfer): Tensor. 0-D Tensor and the content is the input. Examples: - >>> op = ScalarToArray() + >>> op = P.ScalarToArray() >>> data = 1.0 >>> output = op(data) """ @@ -893,7 +893,7 @@ class ScalarToTensor(PrimitiveWithInfer): Tensor. 0-D Tensor and the content is the input. Examples: - >>> op = ScalarToTensor() + >>> op = P.ScalarToTensor() >>> data = 1 >>> output = op(data, mindspore.float32) """ @@ -934,7 +934,7 @@ class InvertPermutation(PrimitiveWithInfer): tuple[int]. the lenth is same as input. Examples: - >>> invert = InvertPermutation() + >>> invert = P.InvertPermutation() >>> input_data = (3, 4, 0, 2, 1) >>> output = invert(input_data) >>> output == (2, 4, 3, 0, 1) @@ -982,8 +982,8 @@ class Argmax(PrimitiveWithInfer): Tensor, indices of the max value of input tensor across the axis. Examples: - >>> input = Tensor(np.array([2.0, 3.1, 1.2])) - >>> index = Argmax()(input) + >>> input_x = Tensor(np.array([2.0, 3.1, 1.2])) + >>> index = P.Argmax()(input_x) >>> assert index == Tensor(1, mindspore.int64) """ @@ -1030,8 +1030,8 @@ class Argmin(PrimitiveWithInfer): Tensor, indices of the min value of input tensor across the axis. Examples: - >>> input = Tensor(np.array([2.0, 3.1, 1.2])) - >>> index = Argmin()(input) + >>> input_x = Tensor(np.array([2.0, 3.1, 1.2])) + >>> index = P.Argmin()(input_x) >>> assert index == Tensor(2, mindspore.int64) """ @@ -1082,8 +1082,8 @@ class ArgMaxWithValue(PrimitiveWithInfer): :math:`(x_1, x_2, ..., x_{axis-1}, x_{axis+1}, ..., x_N)`. Examples: - >>> input = Tensor(np.random.rand(5)) - >>> index, output = ArgMaxWithValue()(input) + >>> input_x = Tensor(np.random.rand(5)) + >>> index, output = P.ArgMaxWithValue()(input_x) """ @prim_attr_register @@ -1129,8 +1129,8 @@ class ArgMinWithValue(PrimitiveWithInfer): :math:`(x_1, x_2, ..., x_{axis-1}, x_{axis+1}, ..., x_N)`. Examples: - >>> input = Tensor(np.random.rand(5)) - >>> index, output = ArgMinWithValue()(input) + >>> input_x = Tensor(np.random.rand(5)) + >>> index, output = P.ArgMinWithValue()(input_x) """ @prim_attr_register def __init__(self, axis=0, keep_dims=False): @@ -1325,7 +1325,7 @@ class Concat(PrimitiveWithInfer): Examples: >>> data1 = Tensor(np.array([[0, 1], [2, 1]]).astype(np.int32)) >>> data2 = Tensor(np.array([[0, 1], [2, 1]]).astype(np.int32)) - >>> op = Concat() + >>> op = P.Concat() >>> output = op((data1, data2)) """ @@ -1607,7 +1607,7 @@ class Select(PrimitiveWithInfer): Tensor, has the same shape as input_y. The shape is :math:`(x_1, x_2, ..., x_N, ..., x_R)`. Examples: - >>> select = Select() + >>> select = P.Select() >>> input_x = Tensor([True, False]) >>> input_y = Tensor([2,3], mindspore.float32) >>> input_z = Tensor([1,2], mindspore.float32) @@ -1681,7 +1681,7 @@ class StridedSlice(PrimitiveWithInfer): Examples >>> input_x = Tensor([[[1, 1, 1], [2, 2, 2]], [[3, 3, 3], [4, 4, 4]], >>> [[5, 5, 5], [6, 6, 6]]], mindspore.float32) - >>> slice = StridedSlice() + >>> slice = P.StridedSlice() >>> output = slice(input_x, (1, 0, 0), (2, 1, 3), (1, 1, 1)) >>> output.shape() (1, 1, 3) @@ -1913,9 +1913,9 @@ class ScatterNd(PrimitiveWithInfer): Tensor, the new tensor, has the same type as `update` and the same shape as `shape`. Examples: - >>> op = ScatterNd() - >>> update = Tensor(np.array([3.2, 1.1]), mindspore.float32) + >>> op = P.ScatterNd() >>> indices = Tensor(np.array([[0, 1], [1, 1]]), mindspore.int32) + >>> update = Tensor(np.array([3.2, 1.1]), mindspore.float32) >>> shape = (3, 3) >>> output = op(indices, update, shape) """ @@ -1964,7 +1964,7 @@ class ResizeNearestNeighbor(PrimitiveWithInfer): Examples: >>> input_tensor = Tensor(np.array([[-0.1, 0.3, 3.6], [0.4, 0.5, -3.2]]), mindspore.float32) - >>> resize = ResizeNearestNeighbor((2, 2)) + >>> resize = P.ResizeNearestNeighbor((2, 2)) >>> output = resize(input_tensor) """ @@ -1997,7 +1997,7 @@ class GatherNd(PrimitiveWithInfer): Examples: >>> input_x = Tensor(np.array([[-0.1, 0.3, 3.6], [0.4, 0.5, -3.2]]), mindspore.float32) >>> indices = Tensor(np.array([[0, 0], [1, 1]]), mindspore.int32) - >>> op = GatherNd() + >>> op = P.GatherNd() >>> output = op(input_x, indices) """ @@ -2039,7 +2039,7 @@ class ScatterNdUpdate(PrimitiveWithInfer): >>> input_x = Tensor(np.array([[-0.1, 0.3, 3.6], [0.4, 0.5, -3.2]]), mindspore.float32) >>> indices = Tensor(np.array([[0, 0], [1, 1]]), mindspore.int32) >>> update = Tensor(np.array([1.0, 2.2]), mindspore.float32) - >>> op = ScatterNdUpdate() + >>> op = P.ScatterNdUpdate() >>> output = op(input_x, indices, update) """ @@ -2090,7 +2090,7 @@ class SpaceToDepth(PrimitiveWithInfer): Examples: >>> x = Tensor(np.random.rand(1,3,2,2), mindspore.float32) >>> block_size = 2 - >>> op = SpaceToDepth(block_size) + >>> op = P.SpaceToDepth(block_size) >>> output = op(x) >>> output.asnumpy().shape == (1,12,1,1) """ @@ -2148,7 +2148,7 @@ class DepthToSpace(PrimitiveWithInfer): Examples: >>> x = Tensor(np.random.rand(1,12,1,1), mindspore.float32) >>> block_size = 2 - >>> op = DepthToSpace(block_size) + >>> op = P.DepthToSpace(block_size) >>> output = op(x) >>> output.asnumpy().shape == (1,3,2,2) """ @@ -2212,8 +2212,8 @@ class SpaceToBatch(PrimitiveWithInfer): >>> block_size = 2 >>> paddings = [[0, 0], [0, 0]] >>> space_to_batch = P.SpaceToBatch(block_size, paddings) - >>> x = Tensor(np.array([[[[1, 2], [3, 4]]]]), mindspore.float32) - >>> space_to_batch(x) + >>> input_x = Tensor(np.array([[[[1, 2], [3, 4]]]]), mindspore.float32) + >>> space_to_batch(input_x) [[[[1.]]], [[[2.]]], [[[3.]]], [[[4.]]]] """ @@ -2280,8 +2280,8 @@ class BatchToSpace(PrimitiveWithInfer): >>> block_size = 2 >>> crops = [[0, 0], [0, 0]] >>> op = P.BatchToSpace(block_size, crops) - >>> x = Tensor(np.array([[[[1]]], [[[2]]], [[[3]]], [[[4]]]]), mindspore.float32) - >>> output = op(x) + >>> input_x = Tensor(np.array([[[[1]]], [[[2]]], [[[3]]], [[[4]]]]), mindspore.float32) + >>> output = op(input_x) [[[[1., 2.], [3., 4.]]]] """ diff --git a/mindspore/ops/operations/math_ops.py b/mindspore/ops/operations/math_ops.py index d003f6ee8b..1294a65d02 100644 --- a/mindspore/ops/operations/math_ops.py +++ b/mindspore/ops/operations/math_ops.py @@ -112,9 +112,9 @@ class TensorAdd(_MathBinaryOp): Examples: >>> add = P.TensorAdd() - >>> x = Tensor(np.array([1,2,3]).astype(np.float32)) - >>> y = Tensor(np.array([4,5,6]).astype(np.float32)) - >>> add(x, y) + >>> input_x = Tensor(np.array([1,2,3]).astype(np.float32)) + >>> input_y = Tensor(np.array([4,5,6]).astype(np.float32)) + >>> add(input_x, input_y) [5,7,9] """ @@ -124,23 +124,24 @@ class AssignAdd(PrimitiveWithInfer): Updates a `Parameter` by adding a value to it. Inputs: - - **input_x** (Parameter) - The `Parameter`. - - **input_y** (Union[scalar, Tensor]) - Has the same shape as `input_x`. + - **variable** (Parameter) - The `Parameter`. + - **value** (Union[numbers.Number, Tensor]) - The value to be added to the `variable`. + It should have the same shape as `variable` if it is a Tensor. Examples: >>> class Net(Cell): >>> def __init__(self): >>> super(Net, self).__init__() >>> self.AssignAdd = P.AssignAdd() - >>> self.inputdata = Parameter(initializer(1, [1], mindspore.int64), name="global_step") + >>> self.variable = Parameter(initializer(1, [1], mindspore.int64), name="global_step") >>> >>> def construct(self, x): - >>> self.AssignAdd(self.inputdata, x) - >>> return self.inputdata + >>> self.AssignAdd(self.variable, x) + >>> return self.variable >>> >>> net = Net() - >>> x = Tensor(np.ones([1]).astype(np.int64)*100) - >>> net(x) + >>> value = Tensor(np.ones([1]).astype(np.int64)*100) + >>> net(value) """ __mindspore_signature__ = ( ('variable', sig_rw.RW_WRITE, sig_kind.KIND_POSITIONAL_KEYWORD), @@ -166,22 +167,24 @@ class AssignSub(PrimitiveWithInfer): Updates a `Parameter` by subtracting a value from it. Inputs: - - **input_x** (Parameter) - The `Parameter`. - - **input_y** (Union[scalar, Tensor]) - Has the same shape as `input_x`. + - **variable** (Parameter) - The `Parameter`. + - **value** (Union[numbers.Number, Tensor]) - The value to be subtracted from the `variable`. + It should have the same shape as `variable` if it is a Tensor. Examples: >>> class Net(Cell): >>> def __init__(self): + >>> super(Net, self).__init__() >>> self.AssignSub = P.AssignSub() - >>> self.inputdata = Parameter(initializer(1, [1], mindspore.int64), name="global_step") + >>> self.variable = Parameter(initializer(1, [1], mindspore.int64), name="global_step") >>> >>> def construct(self, x): - >>> self.AssignSub(self.inputdata, x) - >>> return self.inputdata + >>> self.AssignSub(self.variable, x) + >>> return self.variable >>> >>> net = Net() - >>> x = Tensor(np.ones([1]).astype(np.int64)*100) - >>> net(x) + >>> value = Tensor(np.ones([1]).astype(np.int64)*100) + >>> net(value) """ __mindspore_signature__ = ( @@ -263,9 +266,9 @@ class ReduceMean(_Reduce): the shape of output is :math:`(x_1, x_4, ..., x_R)`. Examples: - >>> data = Tensor(np.random.randn(3, 4, 5, 6).astype(np.float32)) + >>> input_x = Tensor(np.random.randn(3, 4, 5, 6).astype(np.float32)) >>> op = P.ReduceMean(keep_dims=True) - >>> output = op(data, 1) + >>> output = op(input_x, 1) """ @@ -295,9 +298,9 @@ class ReduceSum(_Reduce): the shape of output is :math:`(x_1, x_4, ..., x_R)`. Examples: - >>> data = Tensor(np.random.randn(3, 4, 5, 6).astype(np.float32)) + >>> input_x = Tensor(np.random.randn(3, 4, 5, 6).astype(np.float32)) >>> op = P.ReduceSum(keep_dims=True) - >>> output = op(data, 1) + >>> output = op(input_x, 1) """ @@ -328,9 +331,9 @@ class ReduceAll(_Reduce): the shape of output is :math:`(x_1, x_4, ..., x_R)`. Examples: - >>> data = Tensor(np.array([[True, False], [True, True]])) + >>> input_x = Tensor(np.array([[True, False], [True, True]])) >>> op = P.ReduceAll(keep_dims=True) - >>> output = op(data, 1) + >>> output = op(input_x, 1) """ def __infer__(self, input_x, axis): @@ -364,9 +367,9 @@ class ReduceMax(_Reduce): the shape of output is :math:`(x_1, x_4, ..., x_R)`. Examples: - >>> data = Tensor(np.random.randn(3, 4, 5, 6).astype(np.float32)) + >>> input_x = Tensor(np.random.randn(3, 4, 5, 6).astype(np.float32)) >>> op = P.ReduceMax(keep_dims=True) - >>> output = op(data, 1) + >>> output = op(input_x, 1) """ @@ -397,9 +400,9 @@ class ReduceMin(_Reduce): the shape of output is :math:`(x_1, x_4, ..., x_R)`. Examples: - >>> data = Tensor(np.random.randn(3, 4, 5, 6).astype(np.float32)) + >>> input_x = Tensor(np.random.randn(3, 4, 5, 6).astype(np.float32)) >>> op = P.ReduceMin(keep_dims=True) - >>> output = op(data, 1) + >>> output = op(input_x, 1) """ @@ -429,9 +432,9 @@ class ReduceProd(_Reduce): the shape of output is :math:`(x_1, x_4, ..., x_R)`. Examples: - >>> data = Tensor(np.random.randn(3, 4, 5, 6).astype(np.float32)) + >>> input_x = Tensor(np.random.randn(3, 4, 5, 6).astype(np.float32)) >>> op = P.ReduceProd(keep_dims=True) - >>> output = op(data, 1) + >>> output = op(input_x, 1) """ @@ -451,15 +454,15 @@ class CumProd(PrimitiveWithInfer): Tensor, has the same shape and dtype as the 'input_x'. Examples: - >>> data = Tensor(np.array([a, b, c]).astype(np.float32)) + >>> input_x = Tensor(np.array([a, b, c]).astype(np.float32)) >>> op0 = P.CumProd() - >>> output = op0(data, 0) # output=[a, a * b, a * b * c] + >>> output = op0(input_x, 0) # output=[a, a * b, a * b * c] >>> op1 = P.CumProd(exclusive=True) - >>> output = op1(data, 0) # output=[1, a, a * b] + >>> output = op1(input_x, 0) # output=[1, a, a * b] >>> op2 = P.CumProd(reverse=True) - >>> output = op2(data, 0) # output=[a * b * c, b * c, c] + >>> output = op2(input_x, 0) # output=[a * b * c, b * c, c] >>> op3 = P.CumProd(exclusive=True, reverse=True) - >>> output = op3(data, 0) # output=[b * c, c, 1] + >>> output = op3(input_x, 0) # output=[b * c, c, 1] """ @prim_attr_register def __init__(self, exclusive=False, reverse=False): @@ -1190,7 +1193,7 @@ class FloorMod(_MathBinaryOp): Examples: >>> input_x = Tensor(np.array([2, 4, -1]), mindspore.int32) >>> input_y = Tensor(np.array([3, 3, 3]), mindspore.int32) - >>> floor_mod = FloorMod() + >>> floor_mod = P.FloorMod() >>> floor_mod(input_x, input_y) [2, 1, 2] """ @@ -1207,9 +1210,9 @@ class Acosh(PrimitiveWithInfer): Tensor, has the same shape as `input_x`. Examples: - >>> acosh = Acosh() - >>> X = Tensor(np.array([1.0, 1.5, 3.0, 100.0]), mindspore.float32) - >>> output = acosh(X) + >>> acosh = P.Acosh() + >>> input_x = Tensor(np.array([1.0, 1.5, 3.0, 100.0]), mindspore.float32) + >>> output = acosh(input_x) """ @prim_attr_register @@ -1286,7 +1289,7 @@ class EqualCount(PrimitiveWithInfer): - **input_y** (Tensor) - The second input tensor. Outputs: - Tensor, has the same shape as the `input_x`. + Tensor, with the type as `mindspore.int32` and size as (1,). Examples: >>> input_x = Tensor(np.array([1, 2, 3]), mindspore.int32) @@ -1324,7 +1327,7 @@ class NotEqual(_LogicBinaryOp): Inputs: - **input_x** (Union[Tensor, Number, bool]) - The first input is a tensor whose data type is number or bool, or a number or a bool object. - - **input_y** (Union[Tensor, Number, bool]) - The second input tensor whose data type is same as 'input_x' or + - **input_y** (Union[Tensor, Number, bool]) - The second input tensor whose data type is same as `input_x` or a number or a bool object. Outputs: @@ -1359,11 +1362,11 @@ class Greater(_LogicBinaryOp): Inputs: - **input_x** (Union[Tensor, Number]) - The first input is a tensor whose data type is number or a number. - - **input_y** (Union[Tensor, Number]) - The second input is a tensor whose data type is same as 'input_x' or + - **input_y** (Union[Tensor, Number]) - The second input is a tensor whose data type is same as `input_x` or a number. Outputs: - Tensor, the shape is same as the shape after broadcasting, and the data type is same as 'input_x'. + Tensor, the shape is same as the shape after broadcasting, and the data type is bool. Examples: >>> input_x = Tensor(np.array([1, 2, 3]), mindspore.int32) @@ -1386,11 +1389,11 @@ class GreaterEqual(_LogicBinaryOp): Inputs: - **input_x** (Union[Tensor, Number]) - The first input is a tensor whose data type is number or a number. - - **input_y** (Union[Tensor, Number]) - The second input is a tensor whose data type is same as 'input_x' or + - **input_y** (Union[Tensor, Number]) - The second input is a tensor whose data type is same as `input_x` or a number. Outputs: - Tensor, the shape is same as the shape after broadcasting, and the data type is bool'. + Tensor, the shape is same as the shape after broadcasting, and the data type is bool. Examples: >>> input_x = Tensor(np.array([1, 2, 3]), mindspore.int32) @@ -1413,7 +1416,7 @@ class Less(_LogicBinaryOp): Inputs: - **input_x** (Union[Tensor, Number]) - The first input is a tensor whose data type is number or a number. - - **input_y** (Union[Tensor, Number]) - The second input is a tensor whose data type is same as 'input_x' or + - **input_y** (Union[Tensor, Number]) - The second input is a tensor whose data type is same as `input_x` or a number. Outputs: @@ -1440,7 +1443,7 @@ class LessEqual(_LogicBinaryOp): Inputs: - **input_x** (Union[Tensor, Number]) - The first input is a tensor whose data type is number or a number. - - **input_y** (Union[Tensor, Number]) - The second input is a tensor whose data type is same as 'input_x' or + - **input_y** (Union[Tensor, Number]) - The second input is a tensor whose data type is same as `input_x` or a number. Outputs: @@ -1752,8 +1755,8 @@ class Cos(PrimitiveWithInfer): Examples: >>> cos = P.Cos() - >>> X = Tensor(np.array([0.24, 0.83, 0.31, 0.09]), mindspore.float32) - >>> output = cos(X) + >>> input_x = Tensor(np.array([0.24, 0.83, 0.31, 0.09]), mindspore.float32) + >>> output = cos(input_x) """ @prim_attr_register @@ -1780,8 +1783,8 @@ class ACos(PrimitiveWithInfer): Examples: >>> acos = P.ACos() - >>> X = Tensor(np.array([0.74, 0.04, 0.30, 0.56]), mindspore.float32) - >>> output = acos(X) + >>> input_x = Tensor(np.array([0.74, 0.04, 0.30, 0.56]), mindspore.float32) + >>> output = acos(input_x) """ @prim_attr_register @@ -1993,7 +1996,7 @@ class Atan2(_MathBinaryOp): - **input_y** (Tensor) - The input tensor. Outputs: - Tensor, the shape is same as the shape after broadcasting, and the data type is same as 'input_x'. + Tensor, the shape is same as the shape after broadcasting, and the data type is same as `input_x`. Examples: >>> input_x = Tensor(np.array([[0, 1]]), mindspore.float32) diff --git a/mindspore/ops/operations/nn_ops.py b/mindspore/ops/operations/nn_ops.py index 9ee98d174e..83f76455e0 100644 --- a/mindspore/ops/operations/nn_ops.py +++ b/mindspore/ops/operations/nn_ops.py @@ -41,7 +41,7 @@ class Flatten(PrimitiveWithInfer): Examples: >>> input_tensor = Tensor(np.ones(shape=[1, 2, 3, 4]), mindspore.float32) - >>> flatten = Flatten() + >>> flatten = P.Flatten() >>> output = flatten(input_tensor) >>> assert output.shape() == (1, 24) """ @@ -155,7 +155,7 @@ class ReLU(PrimitiveWithInfer): Examples: >>> input_x = Tensor(np.array([[-1.0, 4.0, -8.0], [2.0, -5.0, 9.0]], np.float32)) - >>> relu = ReLU() + >>> relu = P.ReLU() >>> result = relu(input_x) [[0, 4.0, 0.0], [2.0, 0.0, 9.0]] """ @@ -188,7 +188,7 @@ class ReLU6(PrimitiveWithInfer): Examples: >>> input_x = Tensor(np.array([[-1.0, 4.0, -8.0], [2.0, -5.0, 9.0]], np.float32)) - >>> relu6 = ReLU6() + >>> relu6 = P.ReLU6() >>> result = relu6(input_x) """ @@ -222,10 +222,10 @@ class Elu(PrimitiveWithInfer): Examples: >>> input_x = Tensor(np.array([[-1.0, 4.0, -8.0], [2.0, -5.0, 9.0]], np.float32)) - >>> elu = Elu() + >>> elu = P.Elu() >>> result = elu(input_x) Tensor([[-0.632 4.0 -0.999] - [2.0 -0.993 9.0 ]], shape=(2, 3), dtype=ms.float32) + [2.0 -0.993 9.0 ]], shape=(2, 3), dtype=mindspore.float32) """ @prim_attr_register @@ -1082,7 +1082,7 @@ class TopK(PrimitiveWithInfer): Examples: >>> topk = P.TopK(sorted=True) - >>> input_x = Tensor([1, 2, 3, 4, 5], mindspore.float16)) + >>> input_x = Tensor([1, 2, 3, 4, 5], mindspore.float16) >>> k = 3 >>> values, indices = topk(input_x, k) >>> assert values == Tensor(np.array([5, 4, 3])) @@ -1223,8 +1223,8 @@ class ApplyMomentum(PrimitiveWithInfer): Examples: >>> net = ResNet50() - >>> loss = SoftmaxCrossEntropyWithLogits() - >>> opt = ApplyMomentum(Tensor(np.array([0.001])), Tensor(np.array([0.9])), + >>> loss = nn.SoftmaxCrossEntropyWithLogits() + >>> opt = P.ApplyMomentum(Tensor(np.array([0.001])), Tensor(np.array([0.9])), filter(lambda x: x.requires_grad, net.get_parameters())) >>> model = Model(net, loss, opt) """ @@ -1351,6 +1351,7 @@ class SGD(PrimitiveWithInfer): class ApplyRMSProp(PrimitiveWithInfer): """ Optimizer that implements the Root Mean Square prop(RMSProp) algorithm. + Please refer to the usage in source code of `nn.RMSProp`. Note: Update `var` according to the RMSProp algorithm. @@ -1386,12 +1387,6 @@ class ApplyRMSProp(PrimitiveWithInfer): Outputs: Tensor, parameters to be update. - - Examples: - >>> net = Net() - >>> loss = nn.SoftmaxCrossEntropyWithLogits() - >>> opt = RMSProp(params=net.trainable_params(), learning_rate=learning_rate) - >>> model = Model(net, loss, opt) """ @prim_attr_register @@ -1424,6 +1419,7 @@ class ApplyRMSProp(PrimitiveWithInfer): class ApplyCenteredRMSProp(PrimitiveWithInfer): """ Optimizer that implements the centered RMSProp algorithm. + Please refer to the usage in source code of `nn.RMSProp`. Note: Update `var` according to the centered RMSProp algorithm. @@ -1464,12 +1460,6 @@ class ApplyCenteredRMSProp(PrimitiveWithInfer): Outputs: Tensor, parameters to be update. - - Examples: - >>> net = Net() - >>> loss = nn.SoftmaxCrossEntropyWithLogits() - >>> opt = RMSProp(params=net.trainable_params(), learning_rate=learning_rate, centered=True) - >>> model = Model(net, loss, opt) """ @prim_attr_register @@ -1596,7 +1586,7 @@ class DropoutGenMask(Primitive): Tensor, the value of generated mask for input shape. Examples: - >>> dropout_gen_mask = DropoutGenMask() + >>> dropout_gen_mask = P.DropoutGenMask() >>> shape = (20, 16, 50) >>> keep_prob = Tensor(0.5, mindspore.float32) >>> mask = dropout_gen_mask(shape, keep_prob) @@ -1631,8 +1621,8 @@ class DropoutDoMask(PrimitiveWithInfer): >>> x = Tensor(np.ones([20, 16, 50]), mindspore.float32) >>> shape = (20, 16, 50) >>> keep_prob = Tensor(0.5, mindspore.float32) - >>> dropout_gen_mask = DropoutGenMask() - >>> dropout_do_mask = DropoutDoMask() + >>> dropout_gen_mask = P.DropoutGenMask() + >>> dropout_do_mask = P.DropoutDoMask() >>> mask = dropout_gen_mask(shape, keep_prob) >>> output = dropout_do_mask(x, mask, keep_prob) >>> assert output.shape() == (20, 16, 50) @@ -1737,7 +1727,7 @@ class OneHot(PrimitiveWithInfer): Examples: >>> indices = Tensor(np.array([0, 1, 2]), mindspore.int32) >>> depth, on_value, off_value = 3, Tensor(1.0, mindspore.float32), Tensor(0.0, mindspore.float32) - >>> onehot = OneHot() + >>> onehot = P.OneHot() >>> result = onehot(indices, depth, on_value, off_value) [[1, 0, 0], [0, 1, 0], [0, 0, 1]] """ @@ -1793,7 +1783,7 @@ class Gelu(PrimitiveWithInfer): Examples: >>> tensor = Tensor(np.array([1.0, 2.0, 3.0]), mindspore.float32) - >>> gelu = Gelu() + >>> gelu = P.Gelu() >>> result = gelu(tensor) """ @@ -1834,7 +1824,7 @@ class GetNext(PrimitiveWithInfer): and the type is described is `types`. Examples: - >>> get_next = GetNext([mindspore.float32, mindspore.int32], [[32, 1, 28, 28], [10]], 'shared_name') + >>> get_next = P.GetNext([mindspore.float32, mindspore.int32], [[32, 1, 28, 28], [10]], 'shared_name') >>> feature, label = get_next() """ @@ -2015,7 +2005,7 @@ class Pad(PrimitiveWithInfer): Examples: >>> input_tensor = Tensor(np.array([[-0.1, 0.3, 3.6], [0.4, 0.5, -3.2]]), mindspore.float32) - >>> pad_op = Pad(((1, 2), (2, 1))) + >>> pad_op = P.Pad(((1, 2), (2, 1))) >>> output_tensor = pad_op(input_tensor) >>> assert output_tensor == Tensor(np.array([[ 0. , 0. , 0. , 0. , 0. , 0. ], >>> [ 0. , 0. , -0.1, 0.3, 3.6, 0. ], diff --git a/mindspore/train/serialization.py b/mindspore/train/serialization.py index b334c3e9d8..90d8816094 100644 --- a/mindspore/train/serialization.py +++ b/mindspore/train/serialization.py @@ -406,7 +406,7 @@ def export(net, *inputs, file_name, file_format='GEIR'): file_format (str): MindSpore currently supports 'GEIR', 'ONNX' and 'LITE' format for exported model. - GEIR: Graph Engine Intermidiate Representation. An intermidiate representation format of - Ascend model. + Ascend model. - ONNX: Open Neural Network eXchange. An open format built to represent machine learning models. - LITE: Huawei model format for mobile. """