m53297601
/
mindspore
1
0
Fork 0
Code Issues Pull Requests Releases Wiki Activity

!9871 fix bugs of op Sigmoid, LSTM, ACos, Equal, MatMul, NMSWithMask, SSIM, DiGamma and so on

Browse Source 
From: @lihongkang1
Reviewed-by: @linqingke,@liangchenghui
Signed-off-by: @liangchenghui
pull/9871/MERGE
mindspore-ci-bot 4 years ago committed by Gitee
parent 19faf00d28 9b2265a66f
commit 17683091d3
7 changed files with 23 additions and 21 deletions
Show Stats Download Patch File Download Diff File

2
mindspore/nn/layer/activation.py
Unescape View File

@ -408,7 +408,7 @@ class Sigmoid(Cell):
Applies sigmoid-type activation element-wise.
Sigmoid function is defined as:
:math:`\text{sigmoid}(x_i) = \frac{1}{1 + \exp(-x_i)}`, where :math:`x_i` is the element of the input.
:math:`\text{sigmoid}(x_i) = \frac{1}{1 + \exp(-x_i)}`, where :math:`x_i` is the element of the input.
Inputs:
- **input_data** (Tensor) - The input of Tanh.

2
mindspore/nn/layer/basic.py
Unescape View File

@ -98,7 +98,7 @@ class Dropout(Cell):
Note:
Each channel will be zeroed out independently on every construct call.
The outputs are scaled by a factor of :math:`\frac{1}{keep\_prob}` during training so
The outputs are scaled by a factor of :math:`\frac{1}{keep\_prob}` during training so
that the output layer remains at a similar scale. During inference, this
layer returns the same tensor as the input.

14
mindspore/nn/layer/embedding.py
Unescape View File

@ -176,11 +176,8 @@ class EmbeddingLookup(Cell):
Examples:
>>> input_indices = Tensor(np.array([[1, 0], [3, 2]]), mindspore.int32)
>>> result = nn.EmbeddingLookup(4,2)(input_indices)
>>> print(result)
[[[ 0.00856617 0.01039034]
[ 0.00196276 -0.00094072]]
[[ 0.01279703 0.00078912]
[ 0.00084863 -0.00742412]]]
>>> print(result.shape)
(2, 2, 2)
"""
BATCH_SLICE = "batch_slice"
FIELD_SLICE = "field_slice"
@ -350,11 +347,8 @@ class MultiFieldEmbeddingLookup(EmbeddingLookup):
>>> field_ids = Tensor([[0, 1, 1, 0, 0], [0, 0, 1, 0, 0]], mindspore.int32)
>>> net = nn.MultiFieldEmbeddingLookup(10, 2, field_size=2, operator='SUM')
>>> out = net(input_indices, input_values, field_ids)
>>> print(out)
[[[-0.00478983 -0.00772568]
[-0.00968955 -0.00064902]]
[[-0.01251151 -0.01251151]
[-0.00196387 -0.00196387]
>>> print(out.shape)
(2, 2, 2)
"""
OPERATOR_SUM = 'SUM'
OPERATOR_MEAN = 'MEAN'

3
mindspore/nn/layer/image.py
Unescape View File

@ -219,11 +219,12 @@ class SSIM(Cell):
Examples:
>>> net = nn.SSIM()
>>> np.random.seed(0)
>>> img1 = Tensor(np.random.random((1, 3, 16, 16)), mindspore.float32)
>>> img2 = Tensor(np.random.random((1, 3, 16, 16)), mindspore.float32)
>>> output = net(img1, img2)
>>> print(output)
[0.12174469]
[-0.15189075]
"""
def __init__(self, max_val=1.0, filter_size=11, filter_sigma=1.5, k1=0.01, k2=0.03):
super(SSIM, self).__init__()

3
mindspore/nn/layer/math.py
Unescape View File

@ -315,6 +315,7 @@ class DiGamma(Cell):
>>> input_x = Tensor(np.array([2, 3, 4]).astype(np.float32))
>>> op = nn.DiGamma()
>>> output = op(input_x)
>>> print(output)
[0.42278463 0.92278427 1.2561178]
"""
@ -661,7 +662,7 @@ class LBeta(Cell):
>>> input_y = Tensor(np.array([2.0, 3.0, 14.0, 15.0]).astype(np.float32))
>>> lbeta = nn.LBeta()
>>> output = lbeta(input_y, input_x)
>>> print (output)
>>> print(output)
[-1.7917596 -4.094345 -12.000229 -14.754799]
"""

16
mindspore/ops/operations/math_ops.py
Unescape View File

@ -716,6 +716,8 @@ class MatMul(PrimitiveWithInfer):
>>> input_x2 = Tensor(np.ones(shape=[3, 4]), mindspore.float32)
>>> matmul = ops.MatMul()
>>> output = matmul(input_x1, input_x2)
>>> print(output)
[[3. 3. 3. 3.]]
"""
@prim_attr_register
@ -2368,6 +2370,8 @@ class Acosh(PrimitiveWithInfer):
>>> acosh = ops.Acosh()
>>> input_x = Tensor(np.array([1.0, 1.5, 3.0, 100.0]), mindspore.float32)
>>> output = acosh(input_x)
>>> print(output)
[0. 0.9624236 1.7627472 5.298292]
"""
@prim_attr_register
@ -2522,7 +2526,7 @@ class Equal(_LogicBinaryOp):
>>> equal = ops.Equal()
>>> output = equal(input_x, 2.0)
>>> print(output)
Tensor(shape=[3], dtype=Bool, value= [False, True, False])
[False True False]
>>>
>>> input_x = Tensor(np.array([1, 2, 3]), mindspore.int32)
>>> input_y = Tensor(np.array([1, 2, 4]), mindspore.int32)
@ -3246,6 +3250,8 @@ class ACos(PrimitiveWithInfer):
>>> acos = ops.ACos()
>>> input_x = Tensor(np.array([0.74, 0.04, 0.30, 0.56]), mindspore.float32)
>>> output = acos(input_x)
>>> print(output)
[0.7377037 1.5307858 1.2661037 0.97641146]
"""
@prim_attr_register
@ -3360,12 +3366,15 @@ class NMSWithMask(PrimitiveWithInfer):
``Ascend`` ``GPU``
Examples:
>>> bbox = np.random.rand(128, 5)
>>> bbox = np.array([[0.4, 0.2, 0.4, 0.3, 0.1], [0.4, 0.3, 0.6, 0.8, 0.7]])
>>> bbox[:, 2] += bbox[:, 0]
>>> bbox[:, 3] += bbox[:, 1]
>>> inputs = Tensor(bbox, mindspore.float32)
>>> nms = ops.NMSWithMask(0.5)
>>> output_boxes, indices, mask = nms(inputs)
>>> print(output_boxes)
[[0.39990234 0.19995117 0.7998047 0.5 0.09997559]
[0.39990234 0.30004883 1. 1.0996094 0.7001953 ]]
"""
@prim_attr_register
@ -3520,6 +3529,7 @@ class Tan(PrimitiveWithInfer):
>>> tan = ops.Tan()
>>> input_x = Tensor(np.array([-1.0, 0.0, 1.0]), mindspore.float32)
>>> output = tan(input_x)
>>> print(output)
[-1.5574081 0. 1.5574081]
"""
@ -3733,7 +3743,7 @@ class BitwiseOr(_BitwiseBinaryOp):
>>> input_x1 = Tensor(np.array([0, 0, 1, -1, 1, 1, 1]), mindspore.int16)
>>> input_x2 = Tensor(np.array([0, 1, 1, -1, -1, 2, 3]), mindspore.int16)
>>> bitwise_or = ops.BitwiseOr()
>>> boutput = itwise_or(input_x1, input_x2)
>>> output = bitwise_or(input_x1, input_x2)
>>> print(output)
[ 0 1 1 -1 -1 3 3]
"""

4
mindspore/ops/operations/nn_ops.py
Unescape View File

@ -3159,16 +3159,12 @@ class LSTM(PrimitiveWithInfer):
>>> print(output)
[[[0.9640267 0.9640267 ]
[0.9640267 0.9640267 ]]
[[0.9950539 0.9950539 ]
[0.9950539 0.9950539 ]]
[[0.99932843 0.99932843]
[0.99932843 0.99932843]]
[[0.9999084 0.9999084 ]
[0.9999084 0.9999084 ]]
[[0.9999869 0.9999869 ]
[0.9999869 0.9999869 ]]]
"""

Write Preview
Loading…
Cancel
Save