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

fix bugs

tags/v1.1.0
lihongkang 5 years ago
parent
6999cf9c55
commit
6deefb0a32
7 changed files with 49 additions and 18 deletions
Split View
Diff Options
Show Stats Download Patch File Download Diff File
  1. +1
    -1
      mindspore/nn/layer/image.py
  2. +1
    -0
      mindspore/nn/loss/loss.py
  3. +3
    -1
      mindspore/ops/composite/random_ops.py
  4. +16
    -3
      mindspore/ops/operations/array_ops.py
  5. +4
    -0
      mindspore/ops/operations/math_ops.py
  6. +20
    -13
      mindspore/ops/operations/nn_ops.py
  7. +4
    -0
      mindspore/ops/operations/random_ops.py

+ 1
- 1
mindspore/nn/layer/image.py View File

@@ -378,7 +378,7 @@ class PSNR(Cell):
>>> img1 = Tensor(np.random.random((1,3,16,16)))
>>> img2 = Tensor(np.random.random((1,3,16,16)))
>>> psnr = net(img1, img2)
[7.8297315]
"""
def __init__(self, max_val=1.0):
super(PSNR, self).__init__()


+ 1
- 0
mindspore/nn/loss/loss.py View File

@@ -137,6 +137,7 @@ class MSELoss(_Loss):
>>> input_data = Tensor(np.array([1, 2, 3]), mindspore.float32)
>>> target_data = Tensor(np.array([1, 2, 2]), mindspore.float32)
>>> loss(input_data, target_data)
0.33333334
"""
def construct(self, base, target):
x = F.square(base - target)


+ 3
- 1
mindspore/ops/composite/random_ops.py View File

@@ -79,10 +79,12 @@ def normal(shape, mean, stddev, seed=None):
The dtype is float32.

Examples:
>>> shape = (4, 16)
>>> shape = (2, 4)
>>> mean = Tensor(1.0, mstype.float32)
>>> stddev = Tensor(1.0, mstype.float32)
>>> output = C.normal(shape, mean, stddev, seed=5)
[[1.0996436 0.44371283 0.11127508 -0.48055804]
[0.31989878 -1.0644426 1.5076542 1.2290289 ]]
"""
mean_dtype = F.dtype(mean)
stddev_dtype = F.dtype(stddev)


+ 16
- 3
mindspore/ops/operations/array_ops.py View File

@@ -446,6 +446,7 @@ class Shape(Primitive):
>>> input_tensor = Tensor(np.ones(shape=[3, 2, 1]), mindspore.float32)
>>> shape = P.Shape()
>>> output = shape(input_tensor)
(3, 2, 1)
"""

@prim_attr_register
@@ -501,6 +502,9 @@ class Squeeze(PrimitiveWithInfer):
>>> input_tensor = Tensor(np.ones(shape=[3, 2, 1]), mindspore.float32)
>>> squeeze = P.Squeeze(2)
>>> output = squeeze(input_tensor)
[[1. 1.]
[1. 1.]
[1. 1.]]
"""

@prim_attr_register
@@ -748,6 +752,10 @@ class Split(PrimitiveWithInfer):
>>> split = P.Split(1, 2)
>>> x = Tensor(np.array([[1, 1, 1, 1], [2, 2, 2, 2]]))
>>> output = split(x)
([[1, 1],
[2, 2]],
[[1, 1],
[2, 2]])
"""

@prim_attr_register
@@ -828,7 +836,7 @@ class TruncatedNormal(PrimitiveWithInfer):
- **shape** (tuple[int]) - The shape of the output tensor, is a tuple of positive integer.

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

Examples:
>>> shape = (1, 2, 3)
@@ -2451,6 +2459,9 @@ class ScatterNd(PrimitiveWithInfer):
>>> update = Tensor(np.array([3.2, 1.1]), mindspore.float32)
>>> shape = (3, 3)
>>> output = op(indices, update, shape)
[[0. 3.2 0.]
[0. 1.1 0.]
[0. 0. 0. ]]
"""

@prim_attr_register
@@ -2676,6 +2687,8 @@ class ScatterNdUpdate(_ScatterNdOp):
>>> update = Tensor(np.array([1.0, 2.2]), mindspore.float32)
>>> op = P.ScatterNdUpdate()
>>> output = op(input_x, indices, update)
[[1. 0.3 3.6]
[0.4 2.2 -3.2]]
"""


@@ -3136,7 +3149,7 @@ class SpaceToBatch(PrimitiveWithInfer):

Args:
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.
paddings (Union[tuple, 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]
@@ -3284,7 +3297,7 @@ class SpaceToBatchND(PrimitiveWithInfer):
Args:
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. M must be 2.
paddings (list): The padding values for H and W dimension, containing 2 subtraction list.
paddings (Union[tuple, list]): The padding values for H and W dimension, containing 2 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.


+ 4
- 0
mindspore/ops/operations/math_ops.py View File

@@ -413,6 +413,8 @@ class ReduceAll(_Reduce):
>>> input_x = Tensor(np.array([[True, False], [True, True]]))
>>> op = P.ReduceAll(keep_dims=True)
>>> output = op(input_x, 1)
[[False]
[True ]]
"""

def __infer__(self, input_x, axis):
@@ -1942,6 +1944,7 @@ class Mod(_MathBinaryOp):
>>> input_y = Tensor(np.array([3.0, 2.0, 3.0]), mindspore.float32)
>>> mod = P.Mod()
>>> mod(input_x, input_y)
[-1. 1. 0.]
"""

def infer_value(self, x, y):
@@ -3269,6 +3272,7 @@ class SquareSumAll(PrimitiveWithInfer):
>>> input_x2 = Tensor(np.random.randint([3, 2, 5, 7]), mindspore.float32)
>>> square_sum_all = P.SquareSumAll()
>>> square_sum_all(input_x1, input_x2)
(27, 26)
"""

@prim_attr_register


+ 20
- 13
mindspore/ops/operations/nn_ops.py View File

@@ -2955,6 +2955,10 @@ class MirrorPad(PrimitiveWithInfer):
>>> paddings = Tensor([[1,1],[2,2]])
>>> pad = Net()
>>> ms_output = pad(Tensor(x), paddings)
[[0.5525309 0.49183875 0.99110144 0.49183875 0.5525309 0.49183875 0.99110144]
[0.31417271 0.96308136 0.934709 0.96308136 0.31417271 0.96308136 0.934709 ]
[0.5525309 0.49183875 0.99110144 0.49183875 0.5525309 0.49183875 0.99110144]
[0.31417271 0.96308136 0.934709 0.96308136 0.31417271 0.96308136 0.934709 ]]
"""

@prim_attr_register
@@ -4196,15 +4200,16 @@ class SparseApplyAdagrad(PrimitiveWithInfer):
>>> def __init__(self):
>>> super(Net, self).__init__()
>>> self.sparse_apply_adagrad = P.SparseApplyAdagrad(lr=1e-8)
>>> self.var = Parameter(Tensor(np.ones([3, 3, 3]).astype(np.float32)), name="var")
>>> self.accum = Parameter(Tensor(np.ones([3, 3, 3]).astype(np.float32)), name="accum")
>>> self.var = Parameter(Tensor(np.ones([1, 1, 1]).astype(np.float32)), name="var")
>>> self.accum = Parameter(Tensor(np.ones([1, 1, 1]).astype(np.float32)), name="accum")
>>> def construct(self, grad, indices):
>>> out = self.sparse_apply_adagrad(self.var, self.accum, grad, indices)
>>> return out
>>> net = Net()
>>> grad = Tensor(np.random.rand(3, 3, 3).astype(np.float32))
>>> indices = Tensor([0, 1, 2], mstype.int32)
>>> grad = Tensor(np.random.rand(1, 1, 1).astype(np.float32))
>>> indices = Tensor([0], mstype.int32)
>>> result = net(grad, indices)
([[[1.0]]], [[[1.0]]])
"""

__mindspore_signature__ = (
@@ -4283,16 +4288,17 @@ class SparseApplyAdagradV2(PrimitiveWithInfer):
>>> def __init__(self):
>>> super(Net, self).__init__()
>>> self.sparse_apply_adagrad_v2 = P.SparseApplyAdagradV2(lr=1e-8, epsilon=1e-6)
>>> self.var = Parameter(Tensor(np.ones([3, 3, 3]).astype(np.float32)), name="var")
>>> self.accum = Parameter(Tensor(np.ones([3, 3, 3]).astype(np.float32)), name="accum")
>>> self.var = Parameter(Tensor(np.ones([1, 1, 1]).astype(np.float32)), name="var")
>>> self.accum = Parameter(Tensor(np.ones([1, 1, 1]).astype(np.float32)), name="accum")
>>>
>>> def construct(self, grad, indices):
>>> out = self.sparse_apply_adagrad_v2(self.var, self.accum, grad, indices)
>>> return out
>>> net = Net()
>>> grad = Tensor(np.random.rand(3, 3, 3).astype(np.float32))
>>> indices = Tensor([0, 1, 2], mstype.int32)
>>> grad = Tensor(np.random.rand(1, 1, 1).astype(np.float32))
>>> indices = Tensor([0], mstype.int32)
>>> result = net(grad, indices)
([[[1.0]]], [[[1.67194188]]])
"""

__mindspore_signature__ = (
@@ -5089,18 +5095,19 @@ class SparseApplyFtrl(PrimitiveWithCheck):
>>> def __init__(self):
>>> super(SparseApplyFtrlNet, self).__init__()
>>> self.sparse_apply_ftrl = P.SparseApplyFtrl(lr=0.01, l1=0.0, l2=0.0, lr_power=-0.5)
>>> self.var = Parameter(Tensor(np.random.rand(3, 3).astype(np.float32)), name="var")
>>> self.accum = Parameter(Tensor(np.random.rand(3, 3).astype(np.float32)), name="accum")
>>> self.linear = Parameter(Tensor(np.random.rand(3, 3).astype(np.float32)), name="linear")
>>> self.var = Parameter(Tensor(np.random.rand(1, 1).astype(np.float32)), name="var")
>>> self.accum = Parameter(Tensor(np.random.rand(1, 1).astype(np.float32)), name="accum")
>>> self.linear = Parameter(Tensor(np.random.rand(1, 1).astype(np.float32)), name="linear")
>>>
>>> def construct(self, grad, indices):
>>> out = self.sparse_apply_ftrl(self.var, self.accum, self.linear, grad, indices)
>>> return out
>>>
>>> net = SparseApplyFtrlNet()
>>> grad = Tensor(np.random.rand(3, 3).astype(np.float32))
>>> indices = Tensor(np.ones([3]), mindspore.int32)
>>> grad = Tensor(np.random.rand(1, 1).astype(np.float32))
>>> indices = Tensor(np.ones([1]), mindspore.int32)
>>> output = net(grad, indices)
([[1.02914639e-01]], [[7.60280550e-01]], [[7.64630079e-01]])
"""

__mindspore_signature__ = (


+ 4
- 0
mindspore/ops/operations/random_ops.py View File

@@ -341,6 +341,10 @@ class RandomChoiceWithMask(PrimitiveWithInfer):
>>> rnd_choice_mask = P.RandomChoiceWithMask()
>>> input_x = Tensor(np.ones(shape=[240000, 4]).astype(np.bool))
>>> output_y, output_mask = rnd_choice_mask(input_x)
>>> output_y.shape
(256, 2)
>>> output_mask.shape
(256,)
"""

@prim_attr_register


Write Preview
Loading…
Cancel
Save