mindspore.nn.dynamic_lr

Dynamic Learning Rate

mindspore.nn.dynamic_lr.piecewise_constant_lr(milestone, learning_rates)[source]

Get piecewise constant learning rate.

Calculate learning rate by given milestone and learning_rates. Let the value of milestone be \((M_1, M_2, ..., M_N)\) and the value of learning_rates be \((x_1, x_2, ..., x_N)\). N is the length of milestone. Let the output learning rate be y.

\[y[i] = x_t,\ for\ i \in [M_{t-1}, M_t)\]
Parameters
  • milestone (Union[list[int], tuple[int]]) – A list of milestone. This list is a monotone increasing list. Every element is a milestone step, and must be greater than 0.

  • learning_rates (Union[list[float], tuple[float]]) – A list of learning rates.

Returns

list[float]. The size of list is \(M_N\).

Examples

>>> milestone = [2, 5, 10]
>>> learning_rates = [0.1, 0.05, 0.01]
>>> piecewise_constant_lr(milestone, learning_rates)
[0.1, 0.1, 0.05, 0.05, 0.05, 0.01, 0.01, 0.01, 0.01, 0.01]
mindspore.nn.dynamic_lr.exponential_decay_lr(learning_rate, decay_rate, total_step, step_per_epoch, decay_epoch, is_stair=False)[source]

Calculate learning rate base on exponential decay function.

For the i-th step, the formula of computing decayed_learning_rate[i] is:

\[decayed\_learning\_rate[i] = learning\_rate * decay\_rate^{\frac{current\_epoch}{decay\_epoch}}\]

Where \(current\_epoch=floor(\frac{i}{step\_per\_epoch})\).

Parameters
  • learning_rate (float) – The initial value of learning rate.

  • decay_rate (float) – The decay rate.

  • total_step (int) – The total number of steps.

  • step_per_epoch (int) – The number of steps in per epoch.

  • decay_epoch (int) – A value used to calculate decayed learning rate.

  • is_stair (bool) – If true, learning rate is decayed once every decay_epoch times. Default: False.

Returns

list[float]. The size of list is total_step.

Examples

>>> learning_rate = 0.1
>>> decay_rate = 0.9
>>> total_step = 6
>>> step_per_epoch = 2
>>> decay_epoch = 1
>>> exponential_decay_lr(learning_rate, decay_rate, total_step, step_per_epoch, decay_epoch)
[0.1, 0.1, 0.09000000000000001, 0.09000000000000001, 0.08100000000000002, 0.08100000000000002]
mindspore.nn.dynamic_lr.natural_exp_decay_lr(learning_rate, decay_rate, total_step, step_per_epoch, decay_epoch, is_stair=False)[source]

Calculate learning rate base on natural exponential decay function.

For the i-th step, the formula of computing decayed_learning_rate[i] is:

\[decayed\_learning\_rate[i] = learning\_rate * e^{-decay\_rate * current\_epoch}\]

Where \(current\_epoch=floor(\frac{i}{step\_per\_epoch})\).

Parameters
  • learning_rate (float) – The initial value of learning rate.

  • decay_rate (float) – The decay rate.

  • total_step (int) – The total number of steps.

  • step_per_epoch (int) – The number of steps in per epoch.

  • decay_epoch (int) – A value used to calculate decayed learning rate.

  • is_stair (bool) – If true, learning rate is decayed once every decay_epoch times. Default: False.

Returns

list[float]. The size of list is total_step.

Examples

>>> learning_rate = 0.1
>>> decay_rate = 0.9
>>> total_step = 6
>>> step_per_epoch = 2
>>> decay_epoch = 2
>>> natural_exp_decay_lr(learning_rate, decay_rate, total_step, step_per_epoch, decay_epoch, True)
[0.1, 0.1, 0.1, 0.1, 0.016529888822158657, 0.016529888822158657]
mindspore.nn.dynamic_lr.inverse_decay_lr(learning_rate, decay_rate, total_step, step_per_epoch, decay_epoch, is_stair=False)[source]

Calculate learning rate base on inverse-time decay function.

For the i-th step, the formula of computing decayed_learning_rate[i] is:

\[decayed\_learning\_rate[i] = learning\_rate / (1 + decay\_rate * current\_epoch / decay\_epoch)\]

Where \(current\_epoch=floor(\frac{i}{step\_per\_epoch})\).

Parameters
  • learning_rate (float) – The initial value of learning rate.

  • decay_rate (float) – The decay rate.

  • total_step (int) – The total number of steps.

  • step_per_epoch (int) – The number of steps in per epoch.

  • decay_epoch (int) – A value used to calculate decayed learning rate.

  • is_stair (bool) – If true, learning rate is decayed once every decay_epoch times. Default: False.

Returns

list[float]. The size of list is total_step.

Examples

>>> learning_rate = 0.1
>>> decay_rate = 0.5
>>> total_step = 6
>>> step_per_epoch = 1
>>> decay_epoch = 1
>>> inverse_decay_lr(learning_rate, decay_rate, total_step, step_per_epoch, decay_epoch, True)
[0.1, 0.06666666666666667, 0.05, 0.04, 0.03333333333333333, 0.028571428571428574]
mindspore.nn.dynamic_lr.cosine_decay_lr(min_lr, max_lr, total_step, step_per_epoch, decay_epoch)[source]

Calculate learning rate base on cosine decay function.

For the i-th step, the formula of computing decayed_learning_rate[i] is:

\[decayed\_learning\_rate[i] = min\_learning\_rate + 0.5 * (max\_learning\_rate - min\_learning\_rate) * (1 + cos(\frac{current\_epoch}{decay\_epoch}\pi))\]

Where \(current\_epoch=floor(\frac{i}{step\_per\_epoch})\).

Parameters
  • min_lr (float) – The minimum value of learning rate.

  • max_lr (float) – The maximum value of learning rate.

  • total_step (int) – The total number of steps.

  • step_per_epoch (int) – The number of steps in per epoch.

  • decay_epoch (int) – A value used to calculate decayed learning rate.

Returns

list[float]. The size of list is total_step.

Examples

>>> min_lr = 0.01
>>> max_lr = 0.1
>>> total_step = 6
>>> step_per_epoch = 2
>>> decay_epoch = 2
>>> cosine_decay_lr(min_lr, max_lr, total_step, step_per_epoch, decay_epoch)
[0.1, 0.1, 0.05500000000000001, 0.05500000000000001, 0.01, 0.01]
mindspore.nn.dynamic_lr.polynomial_decay_lr(learning_rate, end_learning_rate, total_step, step_per_epoch, decay_epoch, power, update_decay_epoch=False)[source]

Calculate learning rate base on polynomial decay function.

For the i-th step, the formula of computing decayed_learning_rate[i] is:

\[decayed\_learning\_rate[i] = (learning\_rate - end\_learning\_rate) * (1 - tmp\_epoch / tmp\_decay\_epoch)^{power} + end\_learning\_rate\]

Where:

\[tmp\_epoch = min(current\_epoch, decay\_epoch)\]
\[current\_epoch=floor(\frac{i}{step\_per\_epoch})\]
\[tmp\_decay\_epoch = decay\_epoch\]

If update_decay_epoch is true, update the value of tmp_decay_epoch every epoch. The formula is:

\[tmp\_decay\_epoch = decay\_epoch * ceil(current\_epoch / decay\_epoch)\]
Parameters
  • learning_rate (float) – The initial value of learning rate.

  • end_learning_rate (float) – The end value of learning rate.

  • total_step (int) – The total number of steps.

  • step_per_epoch (int) – The number of steps in per epoch.

  • decay_epoch (int) – A value used to calculate decayed learning rate.

  • power (float) – A value used to calculate decayed learning rate. This parameter must be greater than 0.

  • update_decay_epoch (bool) – If true, update decay_epoch. Default: False.

Returns

list[float]. The size of list is total_step.

Examples

>>> learning_rate = 0.1
>>> end_learning_rate = 0.01
>>> total_step = 6
>>> step_per_epoch = 2
>>> decay_epoch = 2
>>> power = 0.5
>>> polynomial_decay_lr(learning_rate, end_learning_rate, total_step, step_per_epoch, decay_epoch, power)
[0.1, 0.1, 0.07363961030678928, 0.07363961030678928, 0.01, 0.01]
mindspore.nn.dynamic_lr.warmup_lr(learning_rate, total_step, step_per_epoch, warmup_epoch)[source]

Get learning rate warming up.

For the i-th step, the formula of computing warmup_learning_rate[i] is:

\[warmup\_learning\_rate[i] = learning\_rate * tmp\_epoch / tmp\_warmup\_epoch\]

Where \(tmp\_epoch=min(current\_epoch, warmup\_epoch),\ current\_epoch=floor(\frac{i}{step\_per\_epoch})\)

Parameters
  • learning_rate (float) – The initial value of learning rate.

  • warmup_steps (int) – The warm up steps of learning rate.

Inputs:

Tensor. The current step number.

Returns

Tensor. The learning rate value for the current step.

Examples

>>> learning_rate = 0.1
>>> total_step = 6
>>> step_per_epoch = 2
>>> warmup_epoch = 2
>>> warmup_lr(learning_rate, total_step, step_per_epoch, warmup_epoch)
[0.0, 0.0, 0.05, 0.05, 0.1, 0.1]