|
1234567891011121314151617181920212223242526272829303132333435363738394041424344454647484950515253545556575859 |
- mindspore.FixedLossScaleManager
- ===============================
-
- .. py:class:: mindspore.FixedLossScaleManager(loss_scale=128.0, drop_overflow_update=True)
-
- ݶȷŴϵĹ̳ :class:`mindspore.LossScaleManager`
-
- ****
-
- - **loss_scale** (float) - ݶȷŴϵע `drop_overflow_update` ΪFalseŻʱҪŻ `loss_scale` ΪֵͬĬֵ128.0
- - **drop_overflow_update** (bool) - ʱǷִŻֵΪTrueʱִŻĬֵTrue
-
- ****
-
- >>> from mindspore import Model, nn, FixedLossScaleManager
- >>>
- >>> net = Net()
- >>> # 1) ִв
- >>> loss_scale_manager = FixedLossScaleManager()
- >>> optim = nn.Momentum(params=net.trainable_params(), learning_rate=0.1, momentum=0.9)
- >>> model = Model(net, loss_scale_manager=loss_scale_manager, optimizer=optim)
- >>>
- >>> # 2) ʹҲִв
- >>> loss_scale = 1024.0
- >>> loss_scale_manager = FixedLossScaleManager(loss_scale, False)
- >>> optim = nn.Momentum(params=net.trainable_params(), learning_rate=0.1, momentum=0.9, loss_scale=loss_scale)
- >>> model = Model(net, loss_scale_manager=loss_scale_manager, optimizer=optim)
-
- .. py:method:: get_drop_overflow_update()
-
- `drop_overflow_update` ֵʾǷڷʱֲ¡
-
- **أ**
-
- bool, `drop_overflow_update` ֵ
-
- .. py:method:: get_loss_scale()
-
- ȡloss scaleֵ
-
- **أ**
-
- bool`loss_scale` ֵ
-
- .. py:method:: get_update_cell()
-
- ڸ `loss_scale` ֵ `Cell` ʵ :class:`mindspore.TrainOneStepWithLossScaleCell` øʵʹù̶ݶȷŴϵ˸ʵִκβ
-
- **أ**
-
- None `Cell` `drop_overflow_update` ΪTrueʱ :class:`mindspore.FixedLossScaleUpdateCell` ʵ `drop_overflow_update` ΪFalseʱNone
-
- .. py:method:: update_loss_scale(overflow)
-
- loss scaleֵ :class:`mindspore.FixedLossScaleManager` У÷ִκβ
-
- ****
-
- - **overflow** (bool) - ʾǷ
|