p47068915
/
mindspore
forked from mindspore-Ecosystem/mindspore
1
0
Fork 0
Code Issues Pull Requests Projects Releases Wiki Activity

!13523 Update doc for forwardvalueandgrad 

From: @joylvliang
Reviewed-by: @chujinjin,@kisnwang
Signed-off-by: @chujinjin
This commit is contained in:
mindspore-ci-bot 2021-03-18 20:20:46 +08:00 committed by Gitee
parent 64a93c2089 c66f21b1e8
commit 103922cde5
1 changed files with 6 additions and 6 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

12
mindspore/nn/wrap/cell_wrapper.py
View File

@ -202,7 +202,7 @@ class ForwardValueAndGrad(Cell):
sens_param (bool): Whether to append sensitivity (gradient with respect to output) as input. sens_param (bool): Whether to append sensitivity (gradient with respect to output) as input.
If sens_param is False, a 'ones_like(outputs)' sensitivity will be attached automatically. If sens_param is False, a 'ones_like(outputs)' sensitivity will be attached automatically.
Default: False. Default: False.
If the sensor_param is True, a sensitivity (gradient with respect to output) needs to be transferred through If the sens_param is True, a sensitivity (gradient with respect to output) needs to be transferred through
the input parameter. the input parameter.
Inputs: Inputs:
@ -227,11 +227,11 @@ class ForwardValueAndGrad(Cell):
... ...
... def construct(self, x): ... def construct(self, x):
... out = self.matmul(x, self.weight) ... out = self.matmul(x, self.weight)
... return x ... return out
... ...
>>> net = Net() >>> net = Net()
>>> criterion = nn.SoftmaxCrossEntropyWithLogits() >>> criterion = nn.SoftmaxCrossEntropyWithLogits()
>>> net_with_criterion = WithLossCell(net, criterion) >>> net_with_criterion = nn.WithLossCell(net, criterion)
>>> weight = ParameterTuple(net.trainable_params()) >>> weight = ParameterTuple(net.trainable_params())
>>> train_network = nn.ForwardValueAndGrad(net_with_criterion, weights=weight, get_all=True, get_by_list=True) >>> train_network = nn.ForwardValueAndGrad(net_with_criterion, weights=weight, get_all=True, get_by_list=True)
>>> inputs = Tensor(np.ones([1, 2]).astype(np.float32)) >>> inputs = Tensor(np.ones([1, 2]).astype(np.float32))
@ -239,10 +239,10 @@ class ForwardValueAndGrad(Cell):
>>> result = train_network(inputs, labels) >>> result = train_network(inputs, labels)
>>> print(result) >>> print(result)
(Tensor(shape=[1], dtype=Float32, value=[0]), ((Tensor(shape=[1, 2], dtype=Float32, value= (Tensor(shape=[1], dtype=Float32, value=[0]), ((Tensor(shape=[1, 2], dtype=Float32, value=
[[0.5, 0.5]]), Tensor(shape=[1, 2], dtype=Float32, value= [[1, 1]]), Tensor(shape=[1, 2], dtype=Float32, value=
[[0, 0]])), (Tensor(shape=[2, 2], dtype=Float32, value= [[0, 0]])), (Tensor(shape=[2, 2], dtype=Float32, value=
[[0, 0], [[0.5, 0.5],
[0, 0]]),))) [0.5, 0.5]]),)))
""" """
def __init__(self, network, weights=None, get_all=False, get_by_list=False, sens_param=False): def __init__(self, network, weights=None, get_all=False, get_by_list=False, sens_param=False):