mindspore-Ecosystem
/
mindspore
21
0
Fork 27
Code Issues Pull Requests 6 Projects Releases Wiki Activity

!18953 fix docs error of TransformToBNN 

Merge pull request !18953 from byweng/master
This commit is contained in:
i-robot 2021-06-29 07:59:59 +00:00 committed by Gitee
parent 8102d68b8b ea3471d5a6
commit 875003dfad
1 changed files with 14 additions and 12 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

26
mindspore/nn/probability/transforms/transform_bnn.py
View File

@ -38,6 +38,8 @@ class TransformToBNN:
``Ascend`` ``GPU``
Examples:
>>> from mindspore.nn.probability import bnn_layers
>>>
>>> class Net(nn.Cell):
... def __init__(self):
... super(Net, self).__init__()
@ -57,9 +59,9 @@ class TransformToBNN:
>>>
>>> net = Net()
>>> criterion = nn.SoftmaxCrossEntropyWithLogits(sparse=True)
>>> optim = Momentum(params=net.trainable_params(), learning_rate=0.1, momentum=0.9)
>>> net_with_loss = WithLossCell(net, criterion)
>>> train_network = TrainOneStepCell(net_with_loss, optim)
>>> optim = nn.AdamWeightDecay(params=net.trainable_params(), learning_rate=0.0001)
>>> net_with_loss = nn.WithLossCell(net, criterion)
>>> train_network = nn.TrainOneStepCell(net_with_loss, optim)
>>> bnn_transformer = TransformToBNN(train_network, 60000, 0.0001)
"""
@ -111,14 +113,14 @@ class TransformToBNN:
Cell, a trainable BNN model wrapped by TrainOneStepCell.
Supported Platforms:
``Ascend`` ``GPU``
``Ascend`` ``GPU``
Examples:
>>> net = Net()
>>> criterion = nn.SoftmaxCrossEntropyWithLogits(sparse=True)
>>> optim = Momentum(params=net.trainable_params(), learning_rate=0.1, momentum=0.9)
>>> net_with_loss = WithLossCell(net, criterion)
>>> train_network = TrainOneStepCell(net_with_loss, optim)
>>> optim = nn.AdamWeightDecay(params=net.trainable_params(), learning_rate=0.0001)
>>> net_with_loss = nn.WithLossCell(net, criterion)
>>> train_network = nn.TrainOneStepCell(net_with_loss, optim)
>>> bnn_transformer = TransformToBNN(train_network, 60000, 0.1)
>>> train_bnn_network = bnn_transformer.transform_to_bnn_model()
"""
@ -156,16 +158,16 @@ class TransformToBNN:
corresponding bayesian layer.
Supported Platforms:
``Ascend`` ``GPU``
``Ascend`` ``GPU``
Examples:
>>> net = Net()
>>> criterion = nn.SoftmaxCrossEntropyWithLogits(sparse=True)
>>> optim = Momentum(params=net.trainable_params(), learning_rate=0.1, momentum=0.9)
>>> net_with_loss = WithLossCell(net, criterion)
>>> train_network = TrainOneStepCell(net_with_loss, optim)
>>> optim = nn.AdamWeightDecay(params=net.trainable_params(), learning_rate=0.0001)
>>> net_with_loss = nn.WithLossCell(net, criterion)
>>> train_network = nn.TrainOneStepCell(net_with_loss, optim)
>>> bnn_transformer = TransformToBNN(train_network, 60000, 0.1)
>>> train_bnn_network = bnn_transformer.transform_to_bnn_layer(Dense, DenseReparam)
>>> train_bnn_network = bnn_transformer.transform_to_bnn_layer(nn.Dense, bnn_layers.DenseReparam)
"""
if dnn_layer_type.__name__ not in ["Dense", "Conv2d"]:
raise ValueError(' \'dnn_layer\'' + str(dnn_layer_type) +