learning_rate_scheduler

cosine_decay

paddle.fluid.layers.cosine_decay(learning_rate, step_each_epoch, epochs)[源代码]

使用 cosine decay 的衰减方式进行学习率调整。

在训练模型时,建议一边进行训练一边降低学习率。 通过使用此方法,学习速率将通过如下cosine衰减策略进行衰减:

\[decayed\_lr = learning\_rate * 0.5 * (cos(epoch * math.pi / epochs) + 1)\]
参数:
  • learning_rate (Variable | float) - 初始学习率。
  • step_each_epoch (int) - 一次迭代中的步数。
  • epochs - 总迭代次数。

代码示例

import paddle.fluid as fluid
base_lr = 0.1
lr = fluid.layers.cosine_decay( learning_rate = base_lr, step_each_epoch=10000, epochs=120)

exponential_decay

paddle.fluid.layers.exponential_decay(learning_rate, decay_steps, decay_rate, staircase=False)[源代码]

在学习率上运用指数衰减。 训练模型时,推荐在训练过程中降低学习率。每次 decay_steps 步骤中用 decay_rate 衰减学习率。

if staircase == True:
    decayed_learning_rate = learning_rate * decay_rate ^ floor(global_step / decay_steps)
else:
    decayed_learning_rate = learning_rate * decay_rate ^ (global_step / decay_steps)
参数:
  • learning_rate (Variable|float)-初始学习率
  • decay_steps (int)-见以上衰减运算
  • decay_rate (float)-衰减率。见以上衰减运算
  • staircase (Boolean)-若为True,按离散区间衰减学习率。默认:False

返回:衰减的学习率

返回类型:变量(Variable)

代码示例

import paddle.fluid as fluid
base_lr = 0.1
sgd_optimizer = fluid.optimizer.SGD(
    learning_rate=fluid.layers.exponential_decay(
        learning_rate=base_lr,
        decay_steps=10000,
        decay_rate=0.5,
        staircase=True))

inverse_time_decay

paddle.fluid.layers.inverse_time_decay(learning_rate, decay_steps, decay_rate, staircase=False)[源代码]

在初始学习率上运用逆时衰减。

训练模型时,最好在训练过程中降低学习率。通过执行该函数,将对初始学习率运用逆向衰减函数。

if staircase == True:
     decayed_learning_rate = learning_rate / (1 + decay_rate * floor(global_step / decay_step))
 else:
     decayed_learning_rate = learning_rate / (1 + decay_rate * global_step / decay_step)
参数:
  • learning_rate (Variable|float)-初始学习率
  • decay_steps (int)-见以上衰减运算
  • decay_rate (float)-衰减率。见以上衰减运算
  • staircase (Boolean)-若为True,按间隔区间衰减学习率。默认:False

返回:衰减的学习率

返回类型:变量(Variable)

示例代码:

import paddle.fluid as fluid
base_lr = 0.1
sgd_optimizer = fluid.optimizer.SGD(
    learning_rate=fluid.layers.natural_exp_decay(
        learning_rate=base_lr,
        decay_steps=10000,
        decay_rate=0.5,
        staircase=True))
sgd_optimizer.minimize(avg_cost)

linear_lr_warmup

paddle.fluid.layers.linear_lr_warmup(learning_rate, warmup_steps, start_lr, end_lr)[源代码]

在正常学习率调整之前先应用线性学习率热身(warm up)进行初步调整。

if global_step < warmup_steps:
    linear_step = end_lr - start_lr
    lr = start_lr + linear_step * (global_step / warmup_steps)
参数:
  • learning_rate (float | Variable) - 学习率,类型为float值或变量。
  • warmup_steps (int) - 进行warm up过程的步数。
  • start_lr (float) - warm up的起始学习率
  • end_lr (float) - warm up的最终学习率。

返回:进行热身衰减后的学习率。

示例代码

import paddle.fluid as fluid
boundaries = [100, 200]
lr_steps = [0.1, 0.01, 0.001]
warmup_steps = 50
start_lr = 1. / 3.
end_lr = 0.1
decayed_lr = fluid.layers.linear_lr_warmup(
    fluid.layers.piecewise_decay(boundaries, lr_steps),
    warmup_steps, start_lr, end_lr)

natural_exp_decay

paddle.fluid.layers.natural_exp_decay(learning_rate, decay_steps, decay_rate, staircase=False)[源代码]

将自然指数衰减运用到初始学习率上。

if not staircase:
    decayed_learning_rate = learning_rate * exp(- decay_rate * (global_step / decay_steps))
else:
    decayed_learning_rate = learning_rate * exp(- decay_rate * floor(global_step / decay_steps))
参数:
  • learning_rate - 标量float32值或变量。是训练过程中的初始学习率。
  • decay_steps - Python int32数
  • decay_rate - Python float数
  • staircase - Boolean.若设为true,每个decay_steps衰减学习率

返回:衰减的学习率

示例代码:

import paddle.fluid as fluid
base_lr = 0.1
sgd_optimizer = fluid.optimizer.SGD(
    learning_rate=fluid.layers.natural_exp_decay(
          learning_rate=base_lr,
          decay_steps=10000,
          decay_rate=0.5,
          staircase=True))

noam_decay

paddle.fluid.layers.noam_decay(d_model, warmup_steps)[源代码]

Noam衰减方法。noam衰减的numpy实现如下。

import padde.fluid as fluid
import numpy as np
# 设置超参数
d_model = 2
current_steps = 20
warmup_steps = 200
# 计算
lr_value = np.power(d_model, -0.5) * np.min([
                       np.power(current_steps, -0.5),
                       np.power(warmup_steps, -1.5) * current_steps])

请参照 attention is all you need

参数:
  • d_model (Variable)-模型的输入和输出维度
  • warmup_steps (Variable)-超参数

返回:衰减的学习率

代码示例

import padde.fluid as fluid
warmup_steps = 100
learning_rate = 0.01
lr = fluid.layers.learning_rate_scheduler.noam_decay(
               1/(warmup_steps *(learning_rate ** 2)),
               warmup_steps)

piecewise_decay

paddle.fluid.layers.piecewise_decay(boundaries, values)[源代码]

对初始学习率进行分段衰减。

该算法可用如下代码描述。

boundaries = [10000, 20000]
values = [1.0, 0.5, 0.1]
if step < 10000:
    learning_rate = 1.0
elif 10000 <= step < 20000:
    learning_rate = 0.5
else:
    learning_rate = 0.1
参数:
  • boundaries -一列代表步数的数字
  • values -一列学习率的值,从不同的步边界中挑选

返回:衰减的学习率

代码示例

import paddle.fluid as fluid
boundaries = [10000, 20000]
values = [1.0, 0.5, 0.1]
optimizer = fluid.optimizer.Momentum(
    momentum=0.9,
    learning_rate=fluid.layers.piecewise_decay(boundaries=boundaries, values=values),
    regularization=fluid.regularizer.L2Decay(1e-4))

polynomial_decay

paddle.fluid.layers.polynomial_decay(learning_rate, decay_steps, end_learning_rate=0.0001, power=1.0, cycle=False)[源代码]

对初始学习率使用多项式衰减

if cycle:
    decay_steps = decay_steps * ceil(global_step / decay_steps)
else:
    global_step = min(global_step, decay_steps)
    decayed_learning_rate = (learning_rate - end_learning_rate) *
        (1 - global_step / decay_steps) ^ power + end_learning_rate
参数:
  • learning_rate (Variable|float32)-标量float32值或变量。是训练过程中的初始学习率。
  • decay_steps (int32)-Python int32数
  • end_learning_rate (float)-Python float数
  • power (float)-Python float数
  • cycle (bool)-若设为true,每decay_steps衰减学习率

返回:衰减的学习率

返回类型:变量(Variable)

代码示例

import paddle.fluid as fluid
start_lr = 0.01
total_step = 5000
end_lr = 0
lr = fluid.layers.polynomial_decay(
    start_lr, total_step, end_lr, power=1)