mindspore.nn.Adam

class mindspore.nn.Adam(params, learning_rate=0.001, beta1=0.9, beta2=0.999, eps=1e-08, use_locking=False, use_nesterov=False, weight_decay=0.0, loss_scale=1.0, use_amsgrad=False, **kwargs)[source]

Implements the Adaptive Moment Estimation (Adam) algorithm.

The Adam optimizer can dynamically adjust the learning rate of each parameter using the first-order moment estimation and the second-order moment estimation of the gradient. The Adam algorithm is proposed in Adam: A Method for Stochastic Optimization.

The updating formulas are as follows:

\[\begin{split}\begin{array}{l} &\newline &\hline \\ &\textbf{Parameters}: \: 1^{\text {st }}\text {moment vector} \: m , \: 2^{\text {nd}} \: \text{moment vector} \: v , \\ &\:\text{gradients } g, \: \text{learning rate} \: \gamma, \text { exponential decay rates for the moment estimates} \: \beta_{1} \: \beta_{2} , \\ &\:\text {parameter vector} \: w_{0}, \:\text{timestep} \: t , \text{ weight decay } \lambda \\ &\textbf{Init}: m_{0} \leftarrow 0, \: v_{0} \leftarrow 0, \: t \leftarrow 0, \: \text{init parameter vector} \: w_{0} \\[-1.ex] &\newline &\hline \\ &\textbf{while} \: w_{t} \: \text{not converged} \: \textbf{do} \\ &\hspace{5mm}\boldsymbol{g}_{t} \leftarrow \nabla_{w} \boldsymbol{f}_{t}\left(\boldsymbol{w}_{t-1}\right) \\ &\hspace{5mm}\textbf {if } \lambda \neq 0 \\ &\hspace{10mm}\boldsymbol{g}_{t} \leftarrow \boldsymbol{g}_{t}+\lambda \boldsymbol{w}_{t-1} \\ &\hspace{5mm}\boldsymbol{m}_{t} \leftarrow \beta_{1} \boldsymbol{m}_{t-1}+\left(1-\beta_{1}\right) \boldsymbol{g}_{t} \\ &\hspace{5mm}\boldsymbol{v}_{t} \leftarrow \beta_{2} \boldsymbol{v}_{t-1}+\left(1-\beta_{2}\right) \boldsymbol{g}_{t}^{2} \\ &\hspace{5mm}\hat{\boldsymbol{m}}_{t} \leftarrow \boldsymbol{m}_{t} /\left(1-\beta_{1}^{t}\right) \\ &\hspace{5mm}\hat{\boldsymbol{v}}_{t} \leftarrow \boldsymbol{v}_{t} /\left(1-\beta_{2}^{t}\right) \\ &\hspace{5mm}\boldsymbol{w}_{t} \leftarrow \boldsymbol{w}_{t-1}-\gamma \hat{\boldsymbol{m}}_{t} /(\sqrt{\hat{\boldsymbol{v}}_{t}}+\epsilon) \\ &\textbf{end while} \\[-1.ex] &\newline &\hline \\[-1.ex] &\textbf{return} \: \boldsymbol{w}_{t} \\[-1.ex] &\newline &\hline \\[-1.ex] \end{array}\end{split}\]

\(m\) represents the 1st moment vector, \(v\) represents the 2nd moment vector, \(g\) represents gradients, \(\beta_1, \beta_2\) represent beta1 and beta2, \(t\) represents the current step while \(beta_1^t\) and \(beta_2^t\) represent beta1_power and beta2_power, \(\gamma\) represents learning_rate, \(w\) represents params, \(\epsilon\) represents eps.

Note

On Ascend, when use_amsgrad is set to True, it might have slightly larger accuracy error.

The sparse strategy is applied while the SparseGatherV2 operator is used for forward network. If the sparse strategy wants to be executed on the host, set the target to the CPU. The sparse feature is under continuous development.

If parameters are not grouped, the weight_decay in optimizer will be applied on the network parameters without ‘beta’ or ‘gamma’ in their names. Users can group parameters to change the strategy of decaying weight. When parameters are grouped, each group can set weight_decay. If not, the weight_decay in optimizer will be applied.

When using Adam with use_lazy=True:

Please note, the optimizer only updates the current index position of the network parameters when the gradient is sparse. The sparse behavior is not equivalent to the original Adam algorithm. If you want to execute a sparse policy, target needs to be set to CPU.

When using Adam with use_offload=True:

This optimizer only supports GRAPH_MODE.

Parameters

  • params (Union[list[Parameter], list[dict]]) –

    Must be list of Parameter or list of dict. When the params is a list of dict, the string “params”, “lr”, “weight_decay”, “grad_centralization” and “order_params” are the keys can be parsed.

    • params: Required. Parameters in current group. The value must be a list of Parameter.

    • lr: Optional. If “lr” in the keys, the value of corresponding learning rate will be used. If not, the learning_rate in optimizer will be used. Fixed and dynamic learning rate are supported.

    • weight_decay: Optional. If “weight_decay” in the keys, the value of corresponding weight decay will be used. If not, the weight_decay in the optimizer will be used. It should be noted that weight decay can be a constant value or a Cell. It is a Cell only when dynamic weight decay is applied. Dynamic weight decay is similar to dynamic learning rate, users need to customize a weight decay schedule only with global step as input, and during training, the optimizer calls the instance of WeightDecaySchedule to get the weight decay value of current step.

    • grad_centralization: Optional. Must be Boolean. If “grad_centralization” is in the keys, the set value will be used. If not, the grad_centralization is False by default. This configuration only works on the convolution layer.

    • order_params: Optional. When parameters is grouped, this usually is used to maintain the order of parameters that appeared in the network to improve performance. The value should be parameters whose order will be followed in optimizer. If order_params in the keys, other keys will be ignored and the element of ‘order_params’ must be in one group of params.

  • learning_rate (Union[float, int, Tensor, Iterable, LearningRateSchedule]) –

    Default: 1e-3 .

    • float: The fixed learning rate value. Must be equal to or greater than 0.

    • int: The fixed learning rate value. Must be equal to or greater than 0. It will be converted to float.

    • Tensor: Its value should be a scalar or a 1-D vector. For scalar, fixed learning rate will be applied. For vector, learning rate is dynamic, then the i-th step will take the i-th value as the learning rate.

    • Iterable: Learning rate is dynamic. The i-th step will take the i-th value as the learning rate.

    • LearningRateSchedule: Learning rate is dynamic. During training, the optimizer calls the instance of LearningRateSchedule with step as the input to get the learning rate of current step.

  • beta1 (float) – The exponential decay rate for the 1st moment estimations. Should be in range (0.0, 1.0). Default: 0.9 .

  • beta2 (float) – The exponential decay rate for the 2nd moment estimations. Should be in range (0.0, 1.0). Default: 0.999 .

  • eps (float) – Term added to the denominator to improve numerical stability. Should be greater than 0. Default: 1e-8 .

  • use_locking (bool) – Whether to enable a lock to protect the updating process of variable tensors. If true , updates of the w, m, and v tensors will be protected by a lock. If false , the result is unpredictable. Default: False .

  • use_nesterov (bool) – Whether to use Nesterov Accelerated Gradient (NAG) algorithm to update the gradients. If true , update the gradients using NAG. If false , update the gradients without using NAG. Default: False .

  • use_amsgrad (bool) – Whether to use Amsgrad algorithm to update the gradients. If true , update the gradients using Amsgrad. If false , update the gradients without using Amsgrad. Default: False .

  • weight_decay (Union[float, int, Cell]) –

    Weight decay (L2 penalty). Default: 0.0 .

    • float: The fixed weight decay value. Must be equal to or greater than 0.

    • int: The fixed weight decay value. Must be equal to or greater than 0. It will be converted to float.

    • Cell: Weight decay is dynamic. During training, the optimizer calls the instance of the Cell with step as the input to get the weight decay value of current step.

  • loss_scale (float) – A floating point value for the loss scale. Should be greater than 0. In general, use the default value. Only when FixedLossScaleManager is used for training and the drop_overflow_update in FixedLossScaleManager is set to False, then this value needs to be the same as the loss_scale in FixedLossScaleManager. Refer to class mindspore.amp.FixedLossScaleManager for more details. Default: 1.0.

  • kwargs –

    • use_lazy (bool): Whether to use Lazy Adam algorithm. Default: False . If true , apply lazy adam algorithm. If false , apply normal adam algorithm.

    • use_offload (bool): Whether to offload adam optimizer to host CPU and keep parameters being updated on the device in order to minimize the memory cost. Default: False . If true , apply offload adam. If false , apply normal adam.

Inputs:

  • gradients (tuple[Tensor]) - The gradients of params, the shape is the same as params.

Outputs:

Tensor[bool], the value is True.

Raises

  • KeyError – If kwargs got keys other than ‘use_lazy’ or ‘use_offload’.

  • TypeError – If learning_rate is not one of int, float, Tensor, Iterable, LearningRateSchedule.

  • TypeError – If element of parameters is neither Parameter nor dict.

  • TypeError – If beta1, beta2, eps or loss_scale is not a float.

  • TypeError – If weight_decay is neither float nor int.

  • TypeError – If use_locking, use_nesterov, use_amsgrad, use_lazy or use_offload is not a bool.

  • ValueError – If loss_scale or eps is less than or equal to 0.

  • ValueError – If beta1, beta2 is not in range (0.0, 1.0).

  • ValueError – If weight_decay is less than 0.

  • ValueError – If use_lazy and use_offload are both true .

  • ValueError – If use_amsgrad is true, use_lazy or use_offload is true .

  • ValueError – If use_amsgrad is True while using distributed training.

Supported Platforms:

Ascend GPU CPU

Examples

>>> import mindspore as ms
>>> from mindspore import nn
>>>
>>> # Define the network structure of LeNet5. Refer to
>>> # https://gitee.com/mindspore/docs/blob/master/docs/mindspore/code/lenet.py
>>> net = LeNet5()
>>> #1) All parameters use the same learning rate and weight decay
>>> optim = nn.Adam(params=net.trainable_params())
>>>
>>> #2) Use parameter groups and set different values
>>> conv_params = list(filter(lambda x: 'conv' in x.name, net.trainable_params()))
>>> no_conv_params = list(filter(lambda x: 'conv' not in x.name, net.trainable_params()))
>>> group_params = [{'params': conv_params, 'weight_decay': 0.01, 'grad_centralization':True},
...                 {'params': no_conv_params, 'lr': 0.01},
...                 {'order_params': net.trainable_params()}]
>>> optim = nn.Adam(group_params, learning_rate=0.1, weight_decay=0.0, use_lazy=False, use_offload=False)
>>> # The conv_params's parameters will use default learning rate of 0.1 and weight decay of 0.01 and grad
>>> # centralization of True.
>>> # The no_conv_params's parameters will use learning rate of 0.01 and default weight decay of 0.0 and grad
>>> # centralization of False.
>>> # The final parameters order in which the optimizer will be followed is the value of 'order_params'.
>>>
>>> loss = nn.SoftmaxCrossEntropyWithLogits()
>>> model = ms.Model(net, loss_fn=loss, optimizer=optim)