mindspore.ops.ApplyPowerSign

class mindspore.ops.ApplyPowerSign[源代码]

Updates relevant entries according to the AddSign algorithm.

\[\begin{split}\begin{array}{ll} \\ m_{t+1} = \beta * m_{t} + (1 - \beta) * g \\ \text{update} = \exp(\text{logbase} * \text{sign_decay} * sign(g) * sign(m)) * g \\ var = var - lr_{t+1} * \text{update} \end{array}\end{split}\]

\(t\) represents updating step while \(m\) represents the 1st moment vector, \(m_{t}\) is the last moment of \(m_{t+1}\), \(lr\) represents scaling factor lr, \(g\) represents grad, \(\beta\) represents beta.

All of inputs comply with the implicit type conversion rules to make the data types consistent. If lr, logbase, sign_decay or beta is a number, the number is automatically converted to Tensor, and the data type is consistent with the Tensor data type involved in the operation. If inputs are tensors and have different data types, the lower priority data type will be converted to the relatively highest priority data type.

Inputs:

  • var (Parameter) - Variable tensor to be updated. With float32 or float16 data type. If data type of var is float16, all inputs must have the same data type as var. The shape is \((N, *)\) where \(*\) means, any number of additional dimensions.

  • m (Parameter) - Variable tensor to be updated, has the same shape and data type as var.

  • lr (Union[Number, Tensor]) - The learning rate value, should be a scalar or Tensor with float32 or float16 data type.

  • logbase (Union[Number, Tensor]) - Should be a scalar or Tensor with float32 or float16 data type.

  • sign_decay (Union[Number, Tensor]) - Should be a scalar or Tensor with float32 or float16 data type.

  • beta (Union[Number, Tensor]) - The exponential decay rate, should be a scalar or Tensor with float32 or float16 data type.

  • grad (Tensor) - A tensor of the same shape and data type as var, for the gradient.

Outputs:

Tuple of 2 Tensors, the updated parameters.

  • var (Tensor) - The same shape and data type as var.

  • m (Tensor) - The same shape and data type as m.

Raises

  • TypeError – If dtype of var, lr, logbase, sign_decay, beta or grad is neither float16 nor float32.

  • TypeError – If lr, logbase, sign_decay or beta is neither a Number nor a Tensor.

  • TypeError – If grad is not a Tensor.

  • RuntimeError – If the data type of lr, logbase, sign_decay and grad conversion of Parameter is not supported.

Supported Platforms:

Ascend

Examples

>>> class Net(nn.Cell):
...     def __init__(self):
...         super(Net, self).__init__()
...         self.apply_power_sign = ops.ApplyPowerSign()
...         self.var = Parameter(Tensor(np.array([[0.6, 0.4],
...                                               [0.1, 0.5]]).astype(np.float32)), name="var")
...         self.m = Parameter(Tensor(np.array([[0.6, 0.5],
...                                             [0.2, 0.6]]).astype(np.float32)), name="m")
...         self.lr = 0.001
...         self.logbase = np.e
...         self.sign_decay = 0.99
...         self.beta = 0.9
...     def construct(self, grad):
...         out = self.apply_power_sign(self.var, self.m, self.lr, self.logbase,
...                                        self.sign_decay, self.beta, grad)
...         return out
...
>>> net = Net()
>>> grad = Tensor(np.array([[0.3, 0.7], [0.1, 0.8]]).astype(np.float32))
>>> output = net(grad)
>>> print(output)
(Tensor(shape=[2, 2], dtype=Float32, value=
[[ 5.95575690e-01,  3.89676481e-01],
 [ 9.85252112e-02,  4.88201708e-01]]), Tensor(shape=[2, 2], dtype=Float32, value=
[[ 5.70000052e-01,  5.19999981e-01],
 [ 1.89999998e-01,  6.20000064e-01]]))