mindspore.ops.ApplyFtrl

class mindspore.ops.ApplyFtrl(use_locking=False)[源代码]

Updates relevant entries according to the FTRL scheme.

For more details, please refer to mindspore.nn.FTRL.

Parameters

use_locking (bool) – Use locks for updating operation if true . Default: False.

Inputs:

  • var (Parameter) - The variable to be updated. The data type must be float16 or float32. The shape is \((N, *)\) where \(*\) means, any number of additional dimensions.

  • accum (Parameter) - The accumulation to be updated, must be same shape and data type as var.

  • linear (Parameter) - The linear coefficient to be updated, must be same shape and data type as var.

  • grad (Tensor) - Gradient. The data type must be float16 or float32.

  • lr (Union[Number, Tensor]) - The learning rate value, must be positive. Default: 0.001. It must be a float number or a scalar tensor with float16 or float32 data type.

  • l1 (Union[Number, Tensor]) - l1 regularization strength, must be greater than or equal to zero. Default: 0.0. It must be a float number or a scalar tensor with float16 or float32 data type.

  • l2 (Union[Number, Tensor]) - l2 regularization strength, must be greater than or equal to zero. Default: 0.0. It must be a float number or a scalar tensor with float16 or float32 data type.

  • lr_power (Union[Number, Tensor]) - Learning rate power controls how the learning rate decreases during training, must be less than or equal to zero. Use fixed learning rate if lr_power is zero. Default: -0.5. It must be a float number or a scalar tensor with float16 or float32 data type.

Outputs:

  • var (Tensor) - Represents the updated var. As the input parameters has been updated in-place, this value is always zero when the platform is GPU.

Raises

  • TypeError – If use_locking is not a bool.

  • TypeError – If dtype of var, grad, lr, l1, l2 or lr_power is neither float16 nor float32.

  • TypeError – If lr, l1, l2 or lr_power is neither a Number nor a Tensor.

  • TypeError – If grad is not a Tensor.

Supported Platforms:

Ascend GPU

Examples

>>> class ApplyFtrlNet(nn.Cell):
...     def __init__(self):
...         super(ApplyFtrlNet, self).__init__()
...         self.apply_ftrl = ops.ApplyFtrl()
...         self.lr = 0.001
...         self.l1 = 0.0
...         self.l2 = 0.0
...         self.lr_power = -0.5
...         self.var = Parameter(Tensor(np.array([[0.6, 0.4],
...                                               [0.1, 0.5]]).astype(np.float32)), name="var")
...         self.accum = Parameter(Tensor(np.array([[0.6, 0.5],
...                                                 [0.2, 0.6]]).astype(np.float32)), name="accum")
...         self.linear = Parameter(Tensor(np.array([[0.9, 0.1],
...                                                  [0.7, 0.8]]).astype(np.float32)), name="linear")
...
...     def construct(self, grad):
...         out = self.apply_ftrl(self.var, self.accum, self.linear, grad, self.lr, self.l1, self.l2,
...                               self.lr_power)
...         return out
...
>>> net = ApplyFtrlNet()
>>> input_x = Tensor(np.array([[0.3, 0.7], [0.1, 0.8]]).astype(np.float32))
>>> output = net(input_x)
>>> print(net.var.asnumpy())
[[ 0.0390525  0.11492836]
 [ 0.00066425 0.15075898]]