mxnet.optimizer.contrib

Contrib optimizers.

Classes

GroupAdaGrad([learning_rate, epsilon, ...])

Adagrad optimizer with row-wise learning rates.
class mxnet.optimizer.contrib.GroupAdaGrad(learning_rate=0.01, epsilon=1e-06, use_fused_step=True, **kwargs)[source]

Bases: Optimizer

Adagrad optimizer with row-wise learning rates.

This class implements the AdaGrad optimizer described in Adaptive Subgradient Methods for Online Learning and Stochastic Optimization, and available at http://www.jmlr.org/papers/volume12/duchi11a/duchi11a.pdf but uses only a single learning rate for every row of the parameter array.

This optimizer updates each weight by:

grad = clip(grad * rescale_grad, clip_gradient)
history += mean(square(grad), axis=1, keepdims=True)
weight -= lr * grad / (sqrt(history) + epsilon)

Weights are updated lazily if the gradient is sparse.

For details of the update algorithm see group_adagrad_update.

This optimizer accepts the following parameters in addition to those accepted by Optimizer. Weight decay is not supported.

Parameters:

  • learning_rate (float, default 0.01) – The initial learning rate. If None, the optimization will use the learning rate from lr_scheduler. If not None, it will overwrite the learning rate in lr_scheduler. If None and lr_scheduler is also None, then it will be set to 0.01 by default.

  • epsilon (float, default 1e-6) – Small value to avoid division by 0.

  • use_fused_step (bool, default True) – Whether or not to use fused kernels for optimizer. When use_fused_step=False or grad is not sparse, step is called, otherwise, fused_step is called.

create_state(index, weight)[source]

Creates auxiliary state for a given weight.

Some optimizers require additional states, e.g. as momentum, in addition to gradients in order to update weights. This function creates state for a given weight which will be used in update. This function is called only once for each weight.

Parameters:

  • index (int) – An unique index to identify the weight.

  • weight (NDArray) – The weight.

Returns:

state – The state associated with the weight.

Return type:

any obj

fused_step(indices, weights, grads, states)[source]

Perform a fused optimization step using gradients and states. Fused kernel is used for update.

Parameters:

  • indices (list of int) – List of unique indices of the parameters into the individual learning rates and weight decays. Learning rates and weight decay may be set via set_lr_mult() and set_wd_mult(), respectively.

  • weights (list of NDArray) – List of parameters to be updated.

  • grads (list of NDArray) – List of gradients of the objective with respect to this parameter.

  • states (List of any obj) – List of state returned by create_state().

step(indices, weights, grads, states)[source]

Perform an optimization step using gradients and states.

Parameters:

  • indices (list of int) – List of unique indices of the parameters into the individual learning rates and weight decays. Learning rates and weight decay may be set via set_lr_mult() and set_wd_mult(), respectively.

  • weights (list of NDArray) – List of parameters to be updated.

  • grads (list of NDArray) – List of gradients of the objective with respect to this parameter.

  • states (List of any obj) – List of state returned by create_state().