# Copyright 2020 Huawei Technologies Co., Ltd
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
# ============================================================================
"""Dynamic Learning Rate"""
from __future__ import absolute_import
import math
from mindspore import _checkparam as validator
[文档]def piecewise_constant_lr(milestone, learning_rates):
r"""
Get piecewise constant learning rate. The learning rate for each step will be stored in a list.
Calculate learning rate by the given `milestone` and `learning_rates`. Let the value of `milestone` be
:math:`(M_1, M_2, ..., M_t, ..., M_N)` and the value of `learning_rates` be :math:`(x_1, x_2, ..., x_t, ..., x_N)`.
N is the length of `milestone`.
Let the output learning rate be :math:`y[i]`, then for the :math:`i`-th step, the formula of
computing decayed_learning_rate[i] is:
.. math::
y[i] = x_t,\ for\ i \in [M_{t-1}, M_t)
Args:
milestone (Union[list[int], tuple[int]]): A list of milestone. When the specified step is reached, use the
corresponding `learning_rates`. This list is a monotone increasing list.
Every element in the list 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 :math:`M_N`.
Raises:
TypeError: If `milestone` or `learning_rates` is neither a tuple nor a list.
ValueError: If the length of `milestone` and `learning_rates` is not same.
ValueError: If the value in `milestone` is not monotonically decreasing.
Supported Platforms:
``Ascend`` ``GPU`` ``CPU``
Examples:
>>> import mindspore.nn as nn
>>>
>>> milestone = [2, 5, 10]
>>> learning_rates = [0.1, 0.05, 0.01]
>>> lr = nn.piecewise_constant_lr(milestone, learning_rates)
>>> # learning_rates = 0.1 if step <= 2
>>> # learning_rates = 0.05 if 2 < step <= 5
>>> # learning_rates = 0.01 if 5 < step <= 10
>>> net = nn.Dense(2, 3)
>>> optim = nn.SGD(net.trainable_params(), learning_rate=lr)
"""
validator.check_value_type('milestone', milestone, (tuple, list))
validator.check_value_type('learning_rates', learning_rates, (tuple, list))
if len(milestone) != len(learning_rates):
raise ValueError(f"For 'piecewise_constant_lr', "
f"the size of 'milestone' must be same with the size of 'learning_rates', "
f"but got 'milestone' size: {len(milestone)}, 'learning_rates' size: {len(learning_rates)}.")
lr = []
last_item = 0
for i, item in enumerate(milestone):
validator.check_positive_int(item, f'milestone[{i}]')
validator.check_is_float(learning_rates[i], f'learning_rates[{i}]')
if item < last_item:
raise ValueError(f"For 'piecewise_constant_lr', "
f"the value of milestone[{i}] must be greater than milestone[{i - 1}], "
f"but got milestone[{i}]: {milestone[i]}, "
f"milestone[{i - 1}]: {milestone[i - 1]}.")
lr += [learning_rates[i]] * (item - last_item)
last_item = item
return lr
def _check_inputs(learning_rate, decay_rate, total_step, step_per_epoch, decay_epoch, is_stair):
validator.check_positive_int(total_step, 'total_step')
validator.check_positive_int(step_per_epoch, 'step_per_epoch')
validator.check_positive_int(decay_epoch, 'decay_epoch')
validator.check_positive_float(learning_rate, 'learning_rate')
validator.check_is_float(learning_rate, 'learning_rate')
validator.check_positive_float(decay_rate, 'decay_rate')
validator.check_is_float(decay_rate, 'decay_rate')
validator.check_value_type('is_stair', is_stair, [bool])
[文档]def exponential_decay_lr(learning_rate, decay_rate, total_step, step_per_epoch, decay_epoch, is_stair=False):
r"""
Calculates learning rate base on exponential decay function. The learning rate for each step will
be stored in a list.
For the i-th step, the formula of computing decayed_learning_rate[i] is:
.. math::
decayed\_learning\_rate[i] = learning\_rate * decay\_rate^{\frac{current\_epoch}{decay\_epoch}}
Where :math:`current\_epoch=floor(\frac{i}{step\_per\_epoch})`.
Args:
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): Number of epochs to decay over.
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`.
Raises:
TypeError: If `total_step` or `step_per_epoch` or `decay_epoch` is not an int.
TypeError: If `is_stair` is not a bool.
TypeError: If `learning_rate` or `decay_rate` is not a float.
ValueError: If `learning_rate` or `decay_rate` is less than or equal to 0.
Supported Platforms:
``Ascend`` ``GPU`` ``CPU``
Examples:
>>> import mindspore.nn as nn
>>>
>>> learning_rate = 0.1
>>> decay_rate = 0.9
>>> total_step = 6
>>> step_per_epoch = 2
>>> decay_epoch = 1
>>> lr = nn.exponential_decay_lr(learning_rate, decay_rate, total_step, step_per_epoch, decay_epoch)
>>> net = nn.Dense(2, 3)
>>> optim = nn.SGD(net.trainable_params(), learning_rate=lr)
"""
_check_inputs(learning_rate, decay_rate, total_step, step_per_epoch, decay_epoch, is_stair)
lr = []
for i in range(total_step):
if is_stair:
lr.append(learning_rate * decay_rate ** math.floor(math.floor(i / step_per_epoch) / decay_epoch))
else:
lr.append(learning_rate * decay_rate ** (math.floor(i / step_per_epoch) / decay_epoch))
return lr
[文档]def natural_exp_decay_lr(learning_rate, decay_rate, total_step, step_per_epoch, decay_epoch, is_stair=False):
r"""
Calculates learning rate base on natural exponential decay function. The learning rate for each step will be
stored in a list.
For the i-th step, the formula of computing decayed_learning_rate[i] is:
.. math::
decayed\_learning\_rate[i] = learning\_rate * e^{-decay\_rate * current\_epoch}
Where :math:`current\_epoch=floor(\frac{i}{step\_per\_epoch})`.
Args:
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): Number of epochs to decay over.
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`.
Raises:
TypeError: If `total_step` or `step_per_epoch` or `decay_epoch` is not an int.
TypeError: If `is_stair` is not a bool.
TypeError: If `learning_rate` or `decay_rate` is not a float.
ValueError: If `learning_rate` or `decay_rate` is less than or equal to 0.
Supported Platforms:
``Ascend`` ``GPU`` ``CPU``
Examples:
>>> import mindspore.nn as nn
>>>
>>> learning_rate = 0.1
>>> decay_rate = 0.9
>>> total_step = 6
>>> step_per_epoch = 2
>>> decay_epoch = 2
>>> lr = nn.natural_exp_decay_lr(learning_rate, decay_rate, total_step, step_per_epoch, decay_epoch, True)
>>> net = nn.Dense(2, 3)
>>> optim = nn.SGD(net.trainable_params(), learning_rate=lr)
"""
_check_inputs(learning_rate, decay_rate, total_step, step_per_epoch, decay_epoch, is_stair)
function = lambda x, y: x
if is_stair:
function = lambda x, y: math.floor(x / y) * y
lr = []
for i in range(total_step):
lr.append(learning_rate * math.e ** (-decay_rate * function(math.floor(i / step_per_epoch), decay_epoch)))
return lr
[文档]def inverse_decay_lr(learning_rate, decay_rate, total_step, step_per_epoch, decay_epoch, is_stair=False):
r"""
Calculates learning rate base on inverse-time decay function. The learning rate for each step
will be stored in a list.
For the i-th step, the formula of computing decayed_learning_rate[i] is:
.. math::
decayed\_learning\_rate[i] = learning\_rate / (1 + decay\_rate * current\_epoch / decay\_epoch)
Where :math:`current\_epoch=floor(\frac{i}{step\_per\_epoch})`.
Args:
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): Number of epochs to decay over.
is_stair (bool): If true, learning rate is decayed once every `decay_epoch` times. If False, the learning rate
decays for every epoch. Default: ``False`` .
Returns:
list[float]. The size of list is `total_step`.
Raises:
TypeError: If `total_step` or `step_per_epoch` or `decay_epoch` is not an int.
TypeError: If `is_stair` is not a bool.
TypeError: If `learning_rate` or `decay_rate` is not a float.
ValueError: If `learning_rate` or `decay_rate` is less than or equal to 0.
Supported Platforms:
``Ascend`` ``GPU`` ``CPU``
Examples:
>>> import mindspore.nn as nn
>>>
>>> learning_rate = 0.1
>>> decay_rate = 0.5
>>> total_step = 6
>>> step_per_epoch = 1
>>> decay_epoch = 1
>>> lr = nn.inverse_decay_lr(learning_rate, decay_rate, total_step, step_per_epoch, decay_epoch, True)
>>> net = nn.Dense(2, 3)
>>> optim = nn.SGD(net.trainable_params(), learning_rate=lr)
"""
_check_inputs(learning_rate, decay_rate, total_step, step_per_epoch, decay_epoch, is_stair)
lr = []
for i in range(total_step):
if is_stair:
lr.append(learning_rate / (1 + decay_rate * math.floor(math.floor(i / step_per_epoch) / decay_epoch)))
else:
lr.append(learning_rate / (1 + decay_rate * math.floor(i / step_per_epoch) / decay_epoch))
return lr
[文档]def cosine_decay_lr(min_lr, max_lr, total_step, step_per_epoch, decay_epoch):
r"""
Calculates learning rate base on cosine decay function. The learning rate for each step will be stored in a list.
For the i-th step, the formula of computing decayed_learning_rate[i] is:
.. math::
decayed\_learning\_rate[i] = min\_lr + 0.5 * (max\_lr - min\_lr) *
(1 + cos(\frac{current\_epoch}{decay\_epoch}\pi))
Where :math:`current\_epoch=floor(\frac{i}{step\_per\_epoch})`.
Args:
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): Number of epochs to decay over.
Returns:
list[float]. The size of list is `total_step`.
Raises:
TypeError: If `min_lr` or `max_lr` is not a float.
TypeError: If `total_step` or `step_per_epoch` or `decay_epoch` is not an int.
ValueError: If `max_lr` is not greater than 0 or `min_lr` is less than 0.
ValueError: If `total_step` or `step_per_epoch` or `decay_epoch` is less than 0.
ValueError: If `min_lr` is greater than or equal to `max_lr`.
Supported Platforms:
``Ascend`` ``GPU`` ``CPU``
Examples:
>>> import mindspore.nn as nn
>>>
>>> min_lr = 0.01
>>> max_lr = 0.1
>>> total_step = 6
>>> step_per_epoch = 2
>>> decay_epoch = 2
>>> lr = nn.cosine_decay_lr(min_lr, max_lr, total_step, step_per_epoch, decay_epoch)
>>> net = nn.Dense(2, 3)
>>> optim = nn.SGD(net.trainable_params(), learning_rate=lr)
"""
if not isinstance(min_lr, float):
raise TypeError(f"For 'cosine_decay_lr', the argument 'min_lr' must be type of float, "
f"but got 'min_lr' type: {type(min_lr)}.")
validator.check_non_negative_float(min_lr, "min_lr", None)
validator.check_positive_float(max_lr, 'max_lr')
validator.check_is_float(max_lr, 'max_lr')
validator.check_positive_int(total_step, 'total_step')
validator.check_positive_int(step_per_epoch, 'step_per_epoch')
validator.check_positive_int(decay_epoch, 'decay_epoch')
if min_lr >= max_lr:
raise ValueError(f"For 'cosine_decay_lr', the 'max_lr' must be greater than the 'min_lr', "
f"but got 'max_lr' value: {max_lr}, 'min_lr' value: {min_lr}.")
delta = 0.5 * (max_lr - min_lr)
lr = []
for i in range(total_step):
tmp_epoch = min(math.floor(i / step_per_epoch), decay_epoch)
lr.append(min_lr + delta * (1 + math.cos(math.pi * tmp_epoch / decay_epoch)))
return lr
[文档]def polynomial_decay_lr(learning_rate, end_learning_rate, total_step, step_per_epoch, decay_epoch, power,
update_decay_epoch=False):
r"""
Calculates learning rate base on polynomial decay function. The learning rate for each step
will be stored in a list.
For the i-th step, the formula of computing decayed_learning_rate[i] is:
.. math::
decayed\_learning\_rate[i] = (learning\_rate - end\_learning\_rate) *
(1 - tmp\_epoch / tmp\_decay\_epoch)^{power} + end\_learning\_rate
Where:
.. math::
tmp\_epoch = \min(current\_epoch, decay\_epoch)
.. math::
current\_epoch=floor(\frac{i}{step\_per\_epoch})
.. math::
tmp\_decay\_epoch = decay\_epoch
If `update_decay_epoch` is true, update the value of :math:`tmp\_decay\_epoch` every epoch. The formula is:
.. math::
tmp\_decay\_epoch = decay\_epoch * ceil(current\_epoch / decay\_epoch)
Args:
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): Number of epochs to decay over.
power (float): The power of polynomial. It 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`.
Raises:
TypeError: If `learning_rate` or `end_learning_rate` or `power` is not a float.
TypeError: If `total_step` or `step_per_epoch` or `decay_epoch` is not an int.
TypeError: If `update_decay_epoch` is not a bool.
ValueError: If `learning_rate` or `power` is not greater than 0.
Supported Platforms:
``Ascend`` ``GPU`` ``CPU``
Examples:
>>> import mindspore.nn as nn
>>>
>>> lr = 0.1
>>> end_learning_rate = 0.01
>>> total_step = 6
>>> step_per_epoch = 2
>>> decay_epoch = 2
>>> power = 0.5
>>> lr = nn.polynomial_decay_lr(lr, end_learning_rate, total_step, step_per_epoch, decay_epoch, power)
>>> net = nn.Dense(2, 3)
>>> optim = nn.SGD(net.trainable_params(), learning_rate=lr)
"""
validator.check_positive_float(learning_rate, 'learning_rate')
validator.check_is_float(learning_rate, 'learning_rate')
if not isinstance(end_learning_rate, float):
raise TypeError(f"For 'polynomial_decay_lr', the argument 'end_learning_rate' must be type of float, "
f"but got 'end_learning_rate' type: {type(end_learning_rate)}.")
validator.check_non_negative_float(end_learning_rate, "end_learning_rate", None)
validator.check_positive_float(power, 'power')
validator.check_is_float(power, 'power')
validator.check_positive_int(total_step, 'total_step')
validator.check_positive_int(step_per_epoch, 'step_per_epoch')
validator.check_positive_int(decay_epoch, 'decay_epoch')
validator.check_value_type('update_decay_epoch', update_decay_epoch, [bool])
origin_decay_epoch = decay_epoch
function = lambda x, y: (x, min(x, y))
if update_decay_epoch:
function = lambda x, y: (origin_decay_epoch * max(math.ceil(y / origin_decay_epoch), 1), y)
lr = []
delta = learning_rate - end_learning_rate
for i in range(total_step):
current_epoch = math.floor(i / step_per_epoch)
decay_epoch, tmp_epoch = function(decay_epoch, current_epoch)
lr.append(delta * (1 - tmp_epoch / decay_epoch) ** power + end_learning_rate)
return lr
[文档]def warmup_lr(learning_rate, total_step, step_per_epoch, warmup_epoch):
r"""
Gets learning rate warming up. The learning rate for each step will be stored in a list.
For the i-th step, the formula of computing warmup_learning_rate[i] is:
.. math::
warmup\_learning\_rate[i] = learning\_rate * tmp\_epoch / warmup\_epoch
Where :math:`tmp\_epoch= \min(current\_epoch, warmup\_epoch),\ current\_epoch=floor(\frac{i}{step\_per\_epoch})`
Args:
learning_rate (float): The initial value of learning rate.
total_step (int): The total number of steps.
step_per_epoch (int): The number of steps in per epoch.
warmup_epoch (int): A value that determines the epochs of the learning rate is warmed up.
Returns:
list[float]. The size of list is `total_step`.
Raises:
TypeError: If `learning_rate` is not a float.
TypeError: If `total_step` or `step_per_epoch` or `decay_epoch` is not an int.
ValueError: If `learning_rate` is less than 0.
Supported Platforms:
``Ascend`` ``GPU`` ``CPU``
Examples:
>>> import mindspore.nn as nn
>>>
>>> learning_rate = 0.1
>>> total_step = 6
>>> step_per_epoch = 2
>>> warmup_epoch = 2
>>> lr = nn.warmup_lr(learning_rate, total_step, step_per_epoch, warmup_epoch)
>>> net = nn.Dense(2, 3)
>>> optim = nn.SGD(net.trainable_params(), learning_rate=lr)
"""
if not isinstance(learning_rate, float):
raise TypeError(f"For 'warmup_lr', the argument 'learning_rate' must be type of float, "
f"but got 'learning_rate' type: {type(learning_rate)}.")
validator.check_non_negative_float(learning_rate, "learning_rate", None)
validator.check_positive_int(warmup_epoch, 'warmup_epoch')
validator.check_positive_int(total_step, 'total_step')
validator.check_positive_int(step_per_epoch, 'step_per_epoch')
function = lambda x, y: (x, min(x, y))
lr = []
for i in range(total_step):
current_epoch = math.floor(i / step_per_epoch)
warmup_epoch, tmp_epoch = function(warmup_epoch, current_epoch)
lr.append(learning_rate * tmp_epoch / warmup_epoch)
return lr
__all__ = [
'piecewise_constant_lr',
'exponential_decay_lr',
'natural_exp_decay_lr',
'inverse_decay_lr',
'cosine_decay_lr',
'polynomial_decay_lr',
'warmup_lr'
]