神经网络的学习的目的是找到使损失函数的值尽可能小的参数。这是寻找最优参数的问题,解决这个问题的过程称为最优化(optimization)。
使用参数的梯度,沿梯度方向更新参数,并重复这个步骤多次,从而逐渐靠近最优参数,这个过程称为随机梯度下降法(stochastic gradient descent),简称SGD。
探险家虽然看不到周围的情况,但是能够知道当前所在位置的坡度(通过脚底感受地面的倾斜状况)。于是,朝着当前所在位置的坡度最大的方向前进,就是SGD的策略。
SGD实现:
class SGD:
"""随机梯度下降法(Stochastic Gradient Descent)"""
#进行初始化时的参数lr表示learning rate(学习率)
def __init__(self, lr=0.01):
self.lr = lr
#该方法在SGD中会被反复调用。
def update(self, params, grads):
for key in params.keys():
params[key] -= self.lr * grads[key]
SGD呈“之”字形移动。这是一个相当低效的路径。也就是说,SGD 的缺点是,如果函数的形状非均向(anisotropic),比如呈延伸状,搜索的路径就会非常低效。因此,我们需要比单纯朝梯度方向前进的SGD更聪明的方法。SGD低效的根本原因是,梯度的方向并没有指向最小值的方向。
Momentum 是“动量”的意思,和物理有关。Momentum方法给人的感觉就像是小球在地面上滚动。在物体不受任何力时,担使物体逐渐减
速的任务
实现
class Momentum:
"""Momentum SGD"""
def __init__(self, lr=0.01, momentum=0.9):
self.lr = lr
self.momentum = momentum
#实例变量v会保存物体的速度。
self.v = None
def update(self, params, grads):
if self.v is None:
self.v = {}
for key, val in params.items():
#v会以字典型变量的形式保存与参数结构相同的数据
self.v[key] = np.zeros_like(val)
for key in params.keys():
self.v[key] = self.momentum*self.v[key] - self.lr*grads[key]
params[key] += self.v[key]
和SGD相比,我们发现“之”字形的“程度”减轻了。这是因为虽然x轴方向上受到的力非常小,但是一直在同一方向上受力,所以朝同一个方向会有一定的加速。反过来,虽然y轴方向上受到的力很大,但是因为交互地受到正方向和反方向的力,它们会互相抵消,所以y轴方向上的速度不稳定。因此,和SGD时的情形相比,
可以更快地朝x轴方向靠近,减弱“之”字形的变动程度。
在关于学习率的有效技巧中,有一种被称为学习率衰减(learning rate decay)的方法,即随着学习的进行,使学习率逐渐减小。实际上,一开始“多”学,然后逐渐“少”学的方法,在神经网络的学习中经常被使用。
逐渐减小学习率的想法,相当于将“全体”参数的学习率值一起降低。而AdaGrad进一步发展了这个想法,针对“一个一个”的参数,赋予其“定制”的值。表达式如下:
可以按参数的元素进行学习率衰减,使变动大的参数的学习率逐渐减小。
融合了Momentum和AdaGrad的方法。
完整代码如下:
创建文件optimizer.py,实现各种优化器 添加代码如下:
# coding: utf-8
import numpy as np
class SGD:
"""随机梯度下降法(Stochastic Gradient Descent)"""
def __init__(self, lr=0.01):
self.lr = lr
def update(self, params, grads):
for key in params.keys():
params[key] -= self.lr * grads[key]
class Momentum:
"""Momentum SGD"""
def __init__(self, lr=0.01, momentum=0.9):
self.lr = lr
self.momentum = momentum
self.v = None
def update(self, params, grads):
if self.v is None:
self.v = {}
for key, val in params.items():
self.v[key] = np.zeros_like(val)
for key in params.keys():
self.v[key] = self.momentum*self.v[key] - self.lr*grads[key]
params[key] += self.v[key]
class Nesterov:
"""Nesterov's Accelerated Gradient (http://arxiv.org/abs/1212.0901)"""
def __init__(self, lr=0.01, momentum=0.9):
self.lr = lr
self.momentum = momentum
self.v = None
def update(self, params, grads):
if self.v is None:
self.v = {}
for key, val in params.items():
self.v[key] = np.zeros_like(val)
for key in params.keys():
self.v[key] *= self.momentum
self.v[key] -= self.lr * grads[key]
params[key] += self.momentum * self.momentum * self.v[key]
params[key] -= (1 + self.momentum) * self.lr * grads[key]
class AdaGrad:
"""AdaGrad"""
def __init__(self, lr=0.01):
self.lr = lr
self.h = None
def update(self, params, grads):
if self.h is None:
self.h = {}
for key, val in params.items():
self.h[key] = np.zeros_like(val)
for key in params.keys():
self.h[key] += grads[key] * grads[key]
params[key] -= self.lr * grads[key] / (np.sqrt(self.h[key]) + 1e-7)
class RMSprop:
"""RMSprop"""
def __init__(self, lr=0.01, decay_rate = 0.99):
self.lr = lr
self.decay_rate = decay_rate
self.h = None
def update(self, params, grads):
if self.h is None:
self.h = {}
for key, val in params.items():
self.h[key] = np.zeros_like(val)
for key in params.keys():
self.h[key] *= self.decay_rate
self.h[key] += (1 - self.decay_rate) * grads[key] * grads[key]
params[key] -= self.lr * grads[key] / (np.sqrt(self.h[key]) + 1e-7)
class Adam:
"""Adam (http://arxiv.org/abs/1412.6980v8)"""
def __init__(self, lr=0.001, beta1=0.9, beta2=0.999):
self.lr = lr
self.beta1 = beta1
self.beta2 = beta2
self.iter = 0
self.m = None
self.v = None
def update(self, params, grads):
if self.m is None:
self.m, self.v = {}, {}
for key, val in params.items():
self.m[key] = np.zeros_like(val)
self.v[key] = np.zeros_like(val)
self.iter += 1
lr_t = self.lr * np.sqrt(1.0 - self.beta2**self.iter) / (1.0 - self.beta1**self.iter)
for key in params.keys():
#self.m[key] = self.beta1*self.m[key] + (1-self.beta1)*grads[key]
#self.v[key] = self.beta2*self.v[key] + (1-self.beta2)*(grads[key]**2)
self.m[key] += (1 - self.beta1) * (grads[key] - self.m[key])
self.v[key] += (1 - self.beta2) * (grads[key]**2 - self.v[key])
params[key] -= lr_t * self.m[key] / (np.sqrt(self.v[key]) + 1e-7)
#unbias_m += (1 - self.beta1) * (grads[key] - self.m[key]) # correct bias
#unbisa_b += (1 - self.beta2) * (grads[key]*grads[key] - self.v[key]) # correct bias
#params[key] += self.lr * unbias_m / (np.sqrt(unbisa_b) + 1e-7)
创建文件optimizer_compare_naive.py,引入优化器
添加代码如下:
# coding: utf-8
import sys, os
sys.path.append(os.pardir) # 为了导入父目录的文件而进行的设定
import numpy as np
import matplotlib.pyplot as plt
from collections import OrderedDict
from optimizer import *
def f(x, y):
return x**2 / 20.0 + y**2
def df(x, y):
return x / 10.0, 2.0*y
init_pos = (-7.0, 2.0)
params = {}
params['x'], params['y'] = init_pos[0], init_pos[1]
grads = {}
grads['x'], grads['y'] = 0, 0
optimizers = OrderedDict()
optimizers["SGD"] = SGD(lr=0.95)
optimizers["Momentum"] = Momentum(lr=0.1)
optimizers["AdaGrad"] = AdaGrad(lr=1.5)
optimizers["Adam"] = Adam(lr=0.3)
idx = 1
for key in optimizers:
optimizer = optimizers[key]
x_history = []
y_history = []
params['x'], params['y'] = init_pos[0], init_pos[1]
for i in range(30):
x_history.append(params['x'])
y_history.append(params['y'])
grads['x'], grads['y'] = df(params['x'], params['y'])
optimizer.update(params, grads)
x = np.arange(-10, 10, 0.01)
y = np.arange(-5, 5, 0.01)
X, Y = np.meshgrid(x, y)
Z = f(X, Y)
# for simple contour line
mask = Z > 7
Z[mask] = 0
# plot
plt.subplot(2, 2, idx)
idx += 1
plt.plot(x_history, y_history, 'o-', color="red")
plt.contour(X, Y, Z)
plt.ylim(-10, 10)
plt.xlim(-10, 10)
plt.plot(0, 0, '+')
#colorbar()
#spring()
plt.title(key)
plt.xlabel("x")
plt.ylabel("y")
plt.show()
运行结果:
根据使用的方法不同,参数更新的路径也不同。只看这个图的话,AdaGrad似乎是最好的,不过也要注意,结果会根据要解决的问题而变。并且,很显然,超参数(学习率等)的设定值不同,结果也会发生变化。
并不存在能在所有问题中都表现良好的方法。这4种方法各有各的特点,都有各自擅长解决的问题和不擅长解决的问题。
很多研究中至今仍在使用SGD。Momentum和AdaGrad也是值得一试的方法。最近,很多研究人员和技术人员都喜欢用Adam。
完整代码如下:
创建文件夹common,移入optimizer.py,
在文件夹common下创建文件util.py
添加代码如下:
# coding: utf-8
import numpy as np
def smooth_curve(x):
"""用于使损失函数的图形变圆滑
参考:http://glowingpython.blogspot.jp/2012/02/convolution-with-numpy.html
"""
window_len = 11
s = np.r_[x[window_len-1:0:-1], x, x[-1:-window_len:-1]]
w = np.kaiser(window_len, 2)
y = np.convolve(w/w.sum(), s, mode='valid')
return y[5:len(y)-5]
def shuffle_dataset(x, t):
"""打乱数据集
Parameters
----------
x : 训练数据
t : 监督数据
Returns
-------
x, t : 打乱的训练数据和监督数据
"""
permutation = np.random.permutation(x.shape[0])
x = x[permutation,:] if x.ndim == 2 else x[permutation,:,:,:]
t = t[permutation]
return x, t
def conv_output_size(input_size, filter_size, stride=1, pad=0):
return (input_size + 2*pad - filter_size) / stride + 1
def im2col(input_data, filter_h, filter_w, stride=1, pad=0):
"""
Parameters
----------
input_data : 由(数据量, 通道, 高, 长)的4维数组构成的输入数据
filter_h : 滤波器的高
filter_w : 滤波器的长
stride : 步幅
pad : 填充
Returns
-------
col : 2维数组
"""
N, C, H, W = input_data.shape
out_h = (H + 2*pad - filter_h)//stride + 1
out_w = (W + 2*pad - filter_w)//stride + 1
img = np.pad(input_data, [(0,0), (0,0), (pad, pad), (pad, pad)], 'constant')
col = np.zeros((N, C, filter_h, filter_w, out_h, out_w))
for y in range(filter_h):
y_max = y + stride*out_h
for x in range(filter_w):
x_max = x + stride*out_w
col[:, :, y, x, :, :] = img[:, :, y:y_max:stride, x:x_max:stride]
col = col.transpose(0, 4, 5, 1, 2, 3).reshape(N*out_h*out_w, -1)
return col
def col2im(col, input_shape, filter_h, filter_w, stride=1, pad=0):
"""
Parameters
----------
col :
input_shape : 输入数据的形状(例:(10, 1, 28, 28))
filter_h :
filter_w
stride
pad
Returns
-------
"""
N, C, H, W = input_shape
out_h = (H + 2*pad - filter_h)//stride + 1
out_w = (W + 2*pad - filter_w)//stride + 1
col = col.reshape(N, out_h, out_w, C, filter_h, filter_w).transpose(0, 3, 4, 5, 1, 2)
img = np.zeros((N, C, H + 2*pad + stride - 1, W + 2*pad + stride - 1))
for y in range(filter_h):
y_max = y + stride*out_h
for x in range(filter_w):
x_max = x + stride*out_w
img[:, :, y:y_max:stride, x:x_max:stride] += col[:, :, y, x, :, :]
return img[:, :, pad:H + pad, pad:W + pad]
在文件夹common下创建文件multi_layer_net.py
添加代码如下:
# coding: utf-8
import sys, os
sys.path.append(os.pardir) # 为了导入父目录的文件而进行的设定
import numpy as np
from collections import OrderedDict
from common.layers import *
from common.gradient import numerical_gradient
class MultiLayerNet:
"""全连接的多层神经网络
Parameters
----------
input_size : 输入大小(MNIST的情况下为784)
hidden_size_list : 隐藏层的神经元数量的列表(e.g. [100, 100, 100])
output_size : 输出大小(MNIST的情况下为10)
activation : 'relu' or 'sigmoid'
weight_init_std : 指定权重的标准差(e.g. 0.01)
指定'relu'或'he'的情况下设定“He的初始值”
指定'sigmoid'或'xavier'的情况下设定“Xavier的初始值”
weight_decay_lambda : Weight Decay(L2范数)的强度
"""
def __init__(self, input_size, hidden_size_list, output_size,
activation='relu', weight_init_std='relu', weight_decay_lambda=0):
self.input_size = input_size
self.output_size = output_size
self.hidden_size_list = hidden_size_list
self.hidden_layer_num = len(hidden_size_list)
self.weight_decay_lambda = weight_decay_lambda
self.params = {}
# 初始化权重
self.__init_weight(weight_init_std)
# 生成层
activation_layer = {'sigmoid': Sigmoid, 'relu': Relu}
self.layers = OrderedDict()
for idx in range(1, self.hidden_layer_num+1):
self.layers['Affine' + str(idx)] = Affine(self.params['W' + str(idx)],
self.params['b' + str(idx)])
self.layers['Activation_function' + str(idx)] = activation_layer[activation]()
idx = self.hidden_layer_num + 1
self.layers['Affine' + str(idx)] = Affine(self.params['W' + str(idx)],
self.params['b' + str(idx)])
self.last_layer = SoftmaxWithLoss()
def __init_weight(self, weight_init_std):
"""设定权重的初始值
Parameters
----------
weight_init_std : 指定权重的标准差(e.g. 0.01)
指定'relu'或'he'的情况下设定“He的初始值”
指定'sigmoid'或'xavier'的情况下设定“Xavier的初始值”
"""
all_size_list = [self.input_size] + self.hidden_size_list + [self.output_size]
for idx in range(1, len(all_size_list)):
scale = weight_init_std
if str(weight_init_std).lower() in ('relu', 'he'):
scale = np.sqrt(2.0 / all_size_list[idx - 1]) # 使用ReLU的情况下推荐的初始值
elif str(weight_init_std).lower() in ('sigmoid', 'xavier'):
scale = np.sqrt(1.0 / all_size_list[idx - 1]) # 使用sigmoid的情况下推荐的初始值
self.params['W' + str(idx)] = scale * np.random.randn(all_size_list[idx-1], all_size_list[idx])
self.params['b' + str(idx)] = np.zeros(all_size_list[idx])
def predict(self, x):
for layer in self.layers.values():
x = layer.forward(x)
return x
def loss(self, x, t):
"""求损失函数
Parameters
----------
x : 输入数据
t : 教师标签
Returns
-------
损失函数的值
"""
y = self.predict(x)
weight_decay = 0
for idx in range(1, self.hidden_layer_num + 2):
W = self.params['W' + str(idx)]
weight_decay += 0.5 * self.weight_decay_lambda * np.sum(W ** 2)
return self.last_layer.forward(y, t) + weight_decay
def accuracy(self, x, t):
y = self.predict(x)
y = np.argmax(y, axis=1)
if t.ndim != 1 : t = np.argmax(t, axis=1)
accuracy = np.sum(y == t) / float(x.shape[0])
return accuracy
def numerical_gradient(self, x, t):
"""求梯度(数值微分)
Parameters
----------
x : 输入数据
t : 教师标签
Returns
-------
具有各层的梯度的字典变量
grads['W1']、grads['W2']、...是各层的权重
grads['b1']、grads['b2']、...是各层的偏置
"""
loss_W = lambda W: self.loss(x, t)
grads = {}
for idx in range(1, self.hidden_layer_num+2):
grads['W' + str(idx)] = numerical_gradient(loss_W, self.params['W' + str(idx)])
grads['b' + str(idx)] = numerical_gradient(loss_W, self.params['b' + str(idx)])
return grads
def gradient(self, x, t):
"""求梯度(误差反向传播法)
Parameters
----------
x : 输入数据
t : 教师标签
Returns
-------
具有各层的梯度的字典变量
grads['W1']、grads['W2']、...是各层的权重
grads['b1']、grads['b2']、...是各层的偏置
"""
# forward
self.loss(x, t)
# backward
dout = 1
dout = self.last_layer.backward(dout)
layers = list(self.layers.values())
layers.reverse()
for layer in layers:
dout = layer.backward(dout)
# 设定
grads = {}
for idx in range(1, self.hidden_layer_num+2):
grads['W' + str(idx)] = self.layers['Affine' + str(idx)].dW + self.weight_decay_lambda * self.layers['Affine' + str(idx)].W
grads['b' + str(idx)] = self.layers['Affine' + str(idx)].db
return grads
在文件夹common下创建文件optimizer_compare_mnist.py
添加代码如下:
# coding: utf-8
import os
import sys
sys.path.append(os.pardir) # 为了导入父目录的文件而进行的设定
import matplotlib.pyplot as plt
from dataset.mnist import load_mnist
from common.util import smooth_curve
from common.multi_layer_net import MultiLayerNet
from common.optimizer import *
# 0:读入MNIST数据==========
(x_train, t_train), (x_test, t_test) = load_mnist(normalize=True)
train_size = x_train.shape[0]
batch_size = 128
max_iterations = 2000
# 1:进行实验的设置==========
optimizers = {}
optimizers['SGD'] = SGD()
optimizers['Momentum'] = Momentum()
optimizers['AdaGrad'] = AdaGrad()
optimizers['Adam'] = Adam()
#optimizers['RMSprop'] = RMSprop()
networks = {}
train_loss = {}
for key in optimizers.keys():
networks[key] = MultiLayerNet(
input_size=784, hidden_size_list=[100, 100, 100, 100],
output_size=10)
train_loss[key] = []
# 2:开始训练==========
for i in range(max_iterations):
batch_mask = np.random.choice(train_size, batch_size)
x_batch = x_train[batch_mask]
t_batch = t_train[batch_mask]
for key in optimizers.keys():
grads = networks[key].gradient(x_batch, t_batch)
optimizers[key].update(networks[key].params, grads)
loss = networks[key].loss(x_batch, t_batch)
train_loss[key].append(loss)
if i % 100 == 0:
print( "===========" + "iteration:" + str(i) + "===========")
for key in optimizers.keys():
loss = networks[key].loss(x_batch, t_batch)
print(key + ":" + str(loss))
# 3.绘制图形==========
markers = {"SGD": "o", "Momentum": "x", "AdaGrad": "s", "Adam": "D"}
x = np.arange(max_iterations)
for key in optimizers.keys():
plt.plot(x, smooth_curve(train_loss[key]), marker=markers[key], markevery=100, label=key)
plt.xlabel("iterations")
plt.ylabel("loss")
plt.ylim(0, 1)
plt.legend()
plt.show()
与SGD相比,其他3种方法学习得更快,而且速度基本相同,仔细看的话,AdaGrad的学习进行得稍微快一点。这个实验需要注意的地方是,实验结果会随学习率等超参数、神经网络的结构(几层深等)的不同而发生变化。不过,一般而言,与SGD相比,其他3种方法可以学习得更快,有时最终的识别精度也更高。
权值衰减就是一种提高泛化能力的技巧,以减小权重参数的值为目的进行学习的方法。通过减小权重参数的值来抑制过拟合的发生。
如果想减小权重的值,一开始就将初始值设为较小的值才是正途。
将权重初始值设为0不是一个好主意。事实上,将权重初始值设为0 的话,将无法正确进行学习。
这是因为在误差反向传播法中,所有的权重值都会进行相同的更新。比如,在2层神经网络中,假设第1层和第2层的权重为0。这样一来,正向传播时,因为输入层的权重为0,所以第2层的神经元全部会被传递相同的值。
将激活函数的输出数据称为“激活值”,但是有的文献中会将在层之间流动的数据也称为“激活值”。
实现:
# coding: utf-8
import numpy as np
import matplotlib.pyplot as plt
def sigmoid(x):
return 1 / (1 + np.exp(-x))
def ReLU(x):
return np.maximum(0, x)
def tanh(x):
return np.tanh(x)
#高斯分布随机生成1000个数据作为输入数据,并把它们传给5层神经网络。
input_data = np.random.randn(1000, 100) # 1000个数据
node_num = 100 # 各隐藏层的节点(神经元)数
hidden_layer_size = 5 # 隐藏层有5层
activations = {} # 激活值的结果保存在这里
x = input_data
for i in range(hidden_layer_size):
if i != 0:
x = activations[i-1]
# 改变初始值进行实验!
w = np.random.randn(node_num, node_num) * 1
# w = np.random.randn(node_num, node_num) * 0.01
# w = np.random.randn(node_num, node_num) * np.sqrt(1.0 / node_num)
# w = np.random.randn(node_num, node_num) * np.sqrt(2.0 / node_num)
a = np.dot(x, w)
# 将激活函数的种类也改变,来进行实验!
z = sigmoid(a)
# z = ReLU(a)
# z = tanh(a)
activations[i] = z
# 绘制直方图
for i, a in activations.items():
plt.subplot(1, len(activations), i+1)
plt.title(str(i+1) + "-layer")
if i != 0: plt.yticks([], [])
# plt.xlim(0.1, 1)
# plt.ylim(0, 7000)
plt.hist(a.flatten(), 30, range=(0,1))
plt.show()
将权重的标准差设为1时,结果显示各层的激活值呈偏向0和1的分布。偏向0和1的数据分布会造成反向传播中梯度的值不断变小,最后消失。这个问题称为梯度消失(gradient vanishing)。层次加深的深度学习中,梯度消失的问题可能会更加严重。
将权重的标准差设为0.01时,这次呈集中在0.5附近的分布。因为不像刚才的例子那样偏向0和1,所以不会发生梯度消失的问题。
因此,激活值在分布上有所偏向会出现“表现力受限”的问题。
尝试使用Xavier Glorot等人的论文中推荐的权重初始值(俗称“Xavier初始值”)。Xavier 的论文中,为了使各层的激活值呈现出具有相同广度的分布,推导了合适的权重尺度。推导出的结论是,如果前一层的节点数为n,则初始值使用标准差为
1 / n 1/ \sqrt{n} 1/n
的分布。结果表明后面的层的分布呈稍微歪斜的形状。如果用tanh函数(双曲线函数)代替sigmoid函数,这个稍微歪斜的问题就能得到改善。实际上,使用tanh函数后,会呈漂亮的吊钟型分布。tanh函数和sigmoid函数同是S型曲线函数,但tanh函数是关于原点(0, 0)对称的S型曲线,而sigmoid函数是关于(x, y)=(0, 0.5)对称的S型曲线。众所周知,用作激活函数的函数最好具有关于原点对称的性质。
当激活函数使用ReLU时,一般推荐使用ReLU专用的初始值,也就是Kaiming He等人推荐的初始值,也称为He初始值。
在这种情况下,随着层的加深,偏向一点点变大。实际上,层加深后,激活值的偏向变大,学习时会出现梯度消失的问题。而当初始值为He初始值时,各层中分布的广度相同。由于即便层加深,数据的广度也能保持不变,因此逆向传播时,也会传递合适的值。
总结一下,当激活函数使用ReLU时,权重初始值使用He初始值,当激活函数为sigmoid或tanh等S型曲线函数时,初始值使用Xavier初始值。这是目前的最佳实践。
创建文件weight_init_compare.py
添加代码如下:
# coding: utf-8
import os
import sys
sys.path.append(os.pardir) # 为了导入父目录的文件而进行的设定
import numpy as np
import matplotlib.pyplot as plt
from dataset.mnist import load_mnist
from common.util import smooth_curve
from common.multi_layer_net import MultiLayerNet
from common.optimizer import SGD
# 0:读入MNIST数据==========
(x_train, t_train), (x_test, t_test) = load_mnist(normalize=True)
train_size = x_train.shape[0]
batch_size = 128
max_iterations = 2000
# 1:进行实验的设置==========
weight_init_types = {'std=0.01': 0.01, 'Xavier': 'sigmoid', 'He': 'relu'}
optimizer = SGD(lr=0.01)
networks = {}
train_loss = {}
for key, weight_type in weight_init_types.items():
networks[key] = MultiLayerNet(input_size=784, hidden_size_list=[100, 100, 100, 100],
output_size=10, weight_init_std=weight_type)
train_loss[key] = []
# 2:开始训练==========
for i in range(max_iterations):
batch_mask = np.random.choice(train_size, batch_size)
x_batch = x_train[batch_mask]
t_batch = t_train[batch_mask]
for key in weight_init_types.keys():
grads = networks[key].gradient(x_batch, t_batch)
optimizer.update(networks[key].params, grads)
loss = networks[key].loss(x_batch, t_batch)
train_loss[key].append(loss)
if i % 100 == 0:
print("===========" + "iteration:" + str(i) + "===========")
for key in weight_init_types.keys():
loss = networks[key].loss(x_batch, t_batch)
print(key + ":" + str(loss))
# 3.绘制图形==========
markers = {'std=0.01': 'o', 'Xavier': 's', 'He': 'D'}
x = np.arange(max_iterations)
for key in weight_init_types.keys():
plt.plot(x, smooth_curve(train_loss[key]), marker=markers[key], markevery=100, label=key)
plt.xlabel("iterations")
plt.ylabel("loss")
plt.ylim(0, 2.5)
plt.legend()
plt.show()
std = 0.01时完全无法进行学习。这和刚才观察到的激活值的分布一样,是因为正向传播中传递的值很小(集中在0附近的数据)。因此,逆向传播时求到的梯度也很小,权重几乎不进行更新。相反,当权重初始值为Xavier初始值和He初始值时,学习进行得很顺利。
并且,我们发现He初始值时的学习进度更快一些。
Batch Normalization可以“强制性”地调整激活值的分布
优点:
• 可以使学习快速进行(可以增大学习率)。
• 不那么依赖初始值(对于初始值不用那么神经质)。
• 抑制过拟合(降低Dropout等的必要性)。
Batch Norm的思路是调整各层的激活值分布使其拥有适当的广度。为此,要向神经网络中插入对数据分布进行正规化的层,即Batch Normalization 层
使用 MNIST 数据集,观察使用Batch Norm层和不使用Batch Norm层时学习的过程
实现:
# coding: utf-8
import sys, os
sys.path.append(os.pardir) # 为了导入父目录的文件而进行的设定
import numpy as np
import matplotlib.pyplot as plt
from dataset.mnist import load_mnist
from common.multi_layer_net_extend import MultiLayerNetExtend
from common.optimizer import SGD, Adam
(x_train, t_train), (x_test, t_test) = load_mnist(normalize=True)
# 减少学习数据
x_train = x_train[:1000]
t_train = t_train[:1000]
max_epochs = 20
train_size = x_train.shape[0]
batch_size = 100
learning_rate = 0.01
def __train(weight_init_std):
bn_network = MultiLayerNetExtend(input_size=784, hidden_size_list=[100, 100, 100, 100, 100], output_size=10,
weight_init_std=weight_init_std, use_batchnorm=True)
network = MultiLayerNetExtend(input_size=784, hidden_size_list=[100, 100, 100, 100, 100], output_size=10,
weight_init_std=weight_init_std)
optimizer = SGD(lr=learning_rate)
train_acc_list = []
bn_train_acc_list = []
iter_per_epoch = max(train_size / batch_size, 1)
epoch_cnt = 0
for i in range(1000000000):
batch_mask = np.random.choice(train_size, batch_size)
x_batch = x_train[batch_mask]
t_batch = t_train[batch_mask]
for _network in (bn_network, network):
grads = _network.gradient(x_batch, t_batch)
optimizer.update(_network.params, grads)
if i % iter_per_epoch == 0:
train_acc = network.accuracy(x_train, t_train)
bn_train_acc = bn_network.accuracy(x_train, t_train)
train_acc_list.append(train_acc)
bn_train_acc_list.append(bn_train_acc)
print("epoch:" + str(epoch_cnt) + " | " + str(train_acc) + " - " + str(bn_train_acc))
epoch_cnt += 1
if epoch_cnt >= max_epochs:
break
return train_acc_list, bn_train_acc_list
# 3.绘制图形==========
weight_scale_list = np.logspace(0, -4, num=16)
x = np.arange(max_epochs)
for i, w in enumerate(weight_scale_list):
print( "============== " + str(i+1) + "/16" + " ==============")
train_acc_list, bn_train_acc_list = __train(w)
plt.subplot(4,4,i+1)
plt.title("W:" + str(w))
if i == 15:
plt.plot(x, bn_train_acc_list, label='Batch Normalization', markevery=2)
plt.plot(x, train_acc_list, linestyle = "--", label='Normal(without BatchNorm)', markevery=2)
else:
plt.plot(x, bn_train_acc_list, markevery=2)
plt.plot(x, train_acc_list, linestyle="--", markevery=2)
plt.ylim(0, 1.0)
if i % 4:
plt.yticks([])
else:
plt.ylabel("accuracy")
if i < 12:
plt.xticks([])
else:
plt.xlabel("epochs")
plt.legend(loc='lower right')
plt.show()
使用Batch Norm后,学习进行得更快了。
发生过拟合的原因,主要有以下两个。
• 模型拥有大量参数、表现力强。
• 训练数据少。
过拟合代码实现,
创建文件overfit_weight_decay.py
添加代码如下:
# coding: utf-8
import os
import sys
sys.path.append(os.pardir) # 为了导入父目录的文件而进行的设定
import numpy as np
import matplotlib.pyplot as plt
from dataset.mnist import load_mnist
from common.multi_layer_net import MultiLayerNet
from common.optimizer import SGD
(x_train, t_train), (x_test, t_test) = load_mnist(normalize=True)
# 为了再现过拟合,减少学习数据
x_train = x_train[:300]
t_train = t_train[:300]
# weight decay(权值衰减)的设定 =======================
#weight_decay_lambda = 0 # 不使用权值衰减的情况
weight_decay_lambda = 0.1
# ====================================================
network = MultiLayerNet(input_size=784, hidden_size_list=[100, 100, 100, 100, 100, 100], output_size=10,
weight_decay_lambda=weight_decay_lambda)
optimizer = SGD(lr=0.01)
max_epochs = 201
train_size = x_train.shape[0]
batch_size = 100
train_loss_list = []
train_acc_list = []
test_acc_list = []
iter_per_epoch = max(train_size / batch_size, 1)
epoch_cnt = 0
for i in range(1000000000):
batch_mask = np.random.choice(train_size, batch_size)
x_batch = x_train[batch_mask]
t_batch = t_train[batch_mask]
grads = network.gradient(x_batch, t_batch)
optimizer.update(network.params, grads)
if i % iter_per_epoch == 0:
train_acc = network.accuracy(x_train, t_train)
test_acc = network.accuracy(x_test, t_test)
train_acc_list.append(train_acc)
test_acc_list.append(test_acc)
print("epoch:" + str(epoch_cnt) + ", train acc:" + str(train_acc) + ", test acc:" + str(test_acc))
epoch_cnt += 1
if epoch_cnt >= max_epochs:
break
# 3.绘制图形==========
markers = {'train': 'o', 'test': 's'}
x = np.arange(max_epochs)
plt.plot(x, train_acc_list, marker='o', label='train', markevery=10)
plt.plot(x, test_acc_list, marker='s', label='test', markevery=10)
plt.xlabel("epochs")
plt.ylabel("accuracy")
plt.ylim(0, 1.0)
plt.legend(loc='lower right')
plt.show()
过了 100 个 epoch 左右后,用训练数据测量到的识别精度几乎都为100%。但是,对于测试数据,离 100% 的识别精度还有较大的差距。如此大的识别精度差距,是只拟合了训练数据的结果。从图中可知,模型对训练时没有使用的一般数据(测试数据)拟合得不是很好。
权值衰减是一直以来经常被使用的一种抑制过拟合的方法。该方法通过在学习的过程中对大的权重进行惩罚,来抑制过拟合。很多过拟合原本就是因为权重参数取值过大才发生的。
对于刚刚进行的实验,应用λ = 0.1的权值衰减。
虽然训练数据的识别精度和测试数据的识别精度之间有差距,但是与没有使用权值衰减相比,差距变小了。
Dropout 是一种在学习的过程中随机删除神经元的方法。训练时,随机选出隐藏层的神经元,然后将其删除。
在common文件夹下创建文件multi_layer_net_extend.py
添加代码如下:
# coding: utf-8
import sys, os
sys.path.append(os.pardir) # 为了导入父目录的文件而进行的设定
import numpy as np
from collections import OrderedDict
from common.layers import *
from common.gradient import numerical_gradient
class MultiLayerNetExtend:
"""扩展版的全连接的多层神经网络
具有Weiht Decay、Dropout、Batch Normalization的功能
Parameters
----------
input_size : 输入大小(MNIST的情况下为784)
hidden_size_list : 隐藏层的神经元数量的列表(e.g. [100, 100, 100])
output_size : 输出大小(MNIST的情况下为10)
activation : 'relu' or 'sigmoid'
weight_init_std : 指定权重的标准差(e.g. 0.01)
指定'relu'或'he'的情况下设定“He的初始值”
指定'sigmoid'或'xavier'的情况下设定“Xavier的初始值”
weight_decay_lambda : Weight Decay(L2范数)的强度
use_dropout: 是否使用Dropout
dropout_ration : Dropout的比例
use_batchNorm: 是否使用Batch Normalization
"""
def __init__(self, input_size, hidden_size_list, output_size,
activation='relu', weight_init_std='relu', weight_decay_lambda=0,
use_dropout = False, dropout_ration = 0.5, use_batchnorm=False):
self.input_size = input_size
self.output_size = output_size
self.hidden_size_list = hidden_size_list
self.hidden_layer_num = len(hidden_size_list)
self.use_dropout = use_dropout
self.weight_decay_lambda = weight_decay_lambda
self.use_batchnorm = use_batchnorm
self.params = {}
# 初始化权重
self.__init_weight(weight_init_std)
# 生成层
activation_layer = {'sigmoid': Sigmoid, 'relu': Relu}
self.layers = OrderedDict()
for idx in range(1, self.hidden_layer_num+1):
self.layers['Affine' + str(idx)] = Affine(self.params['W' + str(idx)],
self.params['b' + str(idx)])
if self.use_batchnorm:
self.params['gamma' + str(idx)] = np.ones(hidden_size_list[idx-1])
self.params['beta' + str(idx)] = np.zeros(hidden_size_list[idx-1])
self.layers['BatchNorm' + str(idx)] = BatchNormalization(self.params['gamma' + str(idx)], self.params['beta' + str(idx)])
self.layers['Activation_function' + str(idx)] = activation_layer[activation]()
if self.use_dropout:
self.layers['Dropout' + str(idx)] = Dropout(dropout_ration)
idx = self.hidden_layer_num + 1
self.layers['Affine' + str(idx)] = Affine(self.params['W' + str(idx)], self.params['b' + str(idx)])
self.last_layer = SoftmaxWithLoss()
def __init_weight(self, weight_init_std):
"""设定权重的初始值
Parameters
----------
weight_init_std : 指定权重的标准差(e.g. 0.01)
指定'relu'或'he'的情况下设定“He的初始值”
指定'sigmoid'或'xavier'的情况下设定“Xavier的初始值”
"""
all_size_list = [self.input_size] + self.hidden_size_list + [self.output_size]
for idx in range(1, len(all_size_list)):
scale = weight_init_std
if str(weight_init_std).lower() in ('relu', 'he'):
scale = np.sqrt(2.0 / all_size_list[idx - 1]) # 使用ReLU的情况下推荐的初始值
elif str(weight_init_std).lower() in ('sigmoid', 'xavier'):
scale = np.sqrt(1.0 / all_size_list[idx - 1]) # 使用sigmoid的情况下推荐的初始值
self.params['W' + str(idx)] = scale * np.random.randn(all_size_list[idx-1], all_size_list[idx])
self.params['b' + str(idx)] = np.zeros(all_size_list[idx])
def predict(self, x, train_flg=False):
for key, layer in self.layers.items():
if "Dropout" in key or "BatchNorm" in key:
x = layer.forward(x, train_flg)
else:
x = layer.forward(x)
return x
def loss(self, x, t, train_flg=False):
"""求损失函数
参数x是输入数据,t是教师标签
"""
y = self.predict(x, train_flg)
weight_decay = 0
for idx in range(1, self.hidden_layer_num + 2):
W = self.params['W' + str(idx)]
weight_decay += 0.5 * self.weight_decay_lambda * np.sum(W**2)
return self.last_layer.forward(y, t) + weight_decay
def accuracy(self, X, T):
Y = self.predict(X, train_flg=False)
Y = np.argmax(Y, axis=1)
if T.ndim != 1 : T = np.argmax(T, axis=1)
accuracy = np.sum(Y == T) / float(X.shape[0])
return accuracy
def numerical_gradient(self, X, T):
"""求梯度(数值微分)
Parameters
----------
X : 输入数据
T : 教师标签
Returns
-------
具有各层的梯度的字典变量
grads['W1']、grads['W2']、...是各层的权重
grads['b1']、grads['b2']、...是各层的偏置
"""
loss_W = lambda W: self.loss(X, T, train_flg=True)
grads = {}
for idx in range(1, self.hidden_layer_num+2):
grads['W' + str(idx)] = numerical_gradient(loss_W, self.params['W' + str(idx)])
grads['b' + str(idx)] = numerical_gradient(loss_W, self.params['b' + str(idx)])
if self.use_batchnorm and idx != self.hidden_layer_num+1:
grads['gamma' + str(idx)] = numerical_gradient(loss_W, self.params['gamma' + str(idx)])
grads['beta' + str(idx)] = numerical_gradient(loss_W, self.params['beta' + str(idx)])
return grads
def gradient(self, x, t):
# forward
self.loss(x, t, train_flg=True)
# backward
dout = 1
dout = self.last_layer.backward(dout)
layers = list(self.layers.values())
layers.reverse()
for layer in layers:
dout = layer.backward(dout)
# 设定
grads = {}
for idx in range(1, self.hidden_layer_num+2):
grads['W' + str(idx)] = self.layers['Affine' + str(idx)].dW + self.weight_decay_lambda * self.params['W' + str(idx)]
grads['b' + str(idx)] = self.layers['Affine' + str(idx)].db
if self.use_batchnorm and idx != self.hidden_layer_num+1:
grads['gamma' + str(idx)] = self.layers['BatchNorm' + str(idx)].dgamma
grads['beta' + str(idx)] = self.layers['BatchNorm' + str(idx)].dbeta
return grads
在common文件夹下创建文件layers.py
添加Dropout类
完整layers.py如下:
# coding: utf-8
import numpy as np
from common.functions import *
from common.util import im2col, col2im
class Relu:
def __init__(self):
self.mask = None
def forward(self, x):
self.mask = (x <= 0)
out = x.copy()
out[self.mask] = 0
return out
def backward(self, dout):
dout[self.mask] = 0
dx = dout
return dx
class Sigmoid:
def __init__(self):
self.out = None
def forward(self, x):
out = sigmoid(x)
self.out = out
return out
def backward(self, dout):
dx = dout * (1.0 - self.out) * self.out
return dx
class Affine:
def __init__(self, W, b):
self.W =W
self.b = b
self.x = None
self.original_x_shape = None
# 权重和偏置参数的导数
self.dW = None
self.db = None
def forward(self, x):
# 对应张量
self.original_x_shape = x.shape
x = x.reshape(x.shape[0], -1)
self.x = x
out = np.dot(self.x, self.W) + self.b
return out
def backward(self, dout):
dx = np.dot(dout, self.W.T)
self.dW = np.dot(self.x.T, dout)
self.db = np.sum(dout, axis=0)
dx = dx.reshape(*self.original_x_shape) # 还原输入数据的形状(对应张量)
return dx
class SoftmaxWithLoss:
def __init__(self):
self.loss = None
self.y = None # softmax的输出
self.t = None # 监督数据
def forward(self, x, t):
self.t = t
self.y = softmax(x)
self.loss = cross_entropy_error(self.y, self.t)
return self.loss
def backward(self, dout=1):
batch_size = self.t.shape[0]
if self.t.size == self.y.size: # 监督数据是one-hot-vector的情况
dx = (self.y - self.t) / batch_size
else:
dx = self.y.copy()
dx[np.arange(batch_size), self.t] -= 1
dx = dx / batch_size
return dx
class Dropout:
"""
http://arxiv.org/abs/1207.0580
"""
def __init__(self, dropout_ratio=0.5):
self.dropout_ratio = dropout_ratio
self.mask = None
def forward(self, x, train_flg=True):
if train_flg:
self.mask = np.random.rand(*x.shape) > self.dropout_ratio
return x * self.mask
else:
return x * (1.0 - self.dropout_ratio)
def backward(self, dout):
return dout * self.mask
class BatchNormalization:
"""
http://arxiv.org/abs/1502.03167
"""
def __init__(self, gamma, beta, momentum=0.9, running_mean=None, running_var=None):
self.gamma = gamma
self.beta = beta
self.momentum = momentum
self.input_shape = None # Conv层的情况下为4维,全连接层的情况下为2维
# 测试时使用的平均值和方差
self.running_mean = running_mean
self.running_var = running_var
# backward时使用的中间数据
self.batch_size = None
self.xc = None
self.std = None
self.dgamma = None
self.dbeta = None
def forward(self, x, train_flg=True):
self.input_shape = x.shape
if x.ndim != 2:
N, C, H, W = x.shape
x = x.reshape(N, -1)
out = self.__forward(x, train_flg)
return out.reshape(*self.input_shape)
def __forward(self, x, train_flg):
if self.running_mean is None:
N, D = x.shape
self.running_mean = np.zeros(D)
self.running_var = np.zeros(D)
if train_flg:
mu = x.mean(axis=0)
xc = x - mu
var = np.mean(xc**2, axis=0)
std = np.sqrt(var + 10e-7)
xn = xc / std
self.batch_size = x.shape[0]
self.xc = xc
self.xn = xn
self.std = std
self.running_mean = self.momentum * self.running_mean + (1-self.momentum) * mu
self.running_var = self.momentum * self.running_var + (1-self.momentum) * var
else:
xc = x - self.running_mean
xn = xc / ((np.sqrt(self.running_var + 10e-7)))
out = self.gamma * xn + self.beta
return out
def backward(self, dout):
if dout.ndim != 2:
N, C, H, W = dout.shape
dout = dout.reshape(N, -1)
dx = self.__backward(dout)
dx = dx.reshape(*self.input_shape)
return dx
def __backward(self, dout):
dbeta = dout.sum(axis=0)
dgamma = np.sum(self.xn * dout, axis=0)
dxn = self.gamma * dout
dxc = dxn / self.std
dstd = -np.sum((dxn * self.xc) / (self.std * self.std), axis=0)
dvar = 0.5 * dstd / self.std
dxc += (2.0 / self.batch_size) * self.xc * dvar
dmu = np.sum(dxc, axis=0)
dx = dxc - dmu / self.batch_size
self.dgamma = dgamma
self.dbeta = dbeta
return dx
class Convolution:
def __init__(self, W, b, stride=1, pad=0):
self.W = W
self.b = b
self.stride = stride
self.pad = pad
# 中间数据(backward时使用)
self.x = None
self.col = None
self.col_W = None
# 权重和偏置参数的梯度
self.dW = None
self.db = None
def forward(self, x):
FN, C, FH, FW = self.W.shape
N, C, H, W = x.shape
out_h = 1 + int((H + 2*self.pad - FH) / self.stride)
out_w = 1 + int((W + 2*self.pad - FW) / self.stride)
col = im2col(x, FH, FW, self.stride, self.pad)
col_W = self.W.reshape(FN, -1).T
out = np.dot(col, col_W) + self.b
out = out.reshape(N, out_h, out_w, -1).transpose(0, 3, 1, 2)
self.x = x
self.col = col
self.col_W = col_W
return out
def backward(self, dout):
FN, C, FH, FW = self.W.shape
dout = dout.transpose(0,2,3,1).reshape(-1, FN)
self.db = np.sum(dout, axis=0)
self.dW = np.dot(self.col.T, dout)
self.dW = self.dW.transpose(1, 0).reshape(FN, C, FH, FW)
dcol = np.dot(dout, self.col_W.T)
dx = col2im(dcol, self.x.shape, FH, FW, self.stride, self.pad)
return dx
class Pooling:
def __init__(self, pool_h, pool_w, stride=1, pad=0):
self.pool_h = pool_h
self.pool_w = pool_w
self.stride = stride
self.pad = pad
self.x = None
self.arg_max = None
def forward(self, x):
N, C, H, W = x.shape
out_h = int(1 + (H - self.pool_h) / self.stride)
out_w = int(1 + (W - self.pool_w) / self.stride)
col = im2col(x, self.pool_h, self.pool_w, self.stride, self.pad)
col = col.reshape(-1, self.pool_h*self.pool_w)
arg_max = np.argmax(col, axis=1)
out = np.max(col, axis=1)
out = out.reshape(N, out_h, out_w, C).transpose(0, 3, 1, 2)
self.x = x
self.arg_max = arg_max
return out
def backward(self, dout):
dout = dout.transpose(0, 2, 3, 1)
pool_size = self.pool_h * self.pool_w
dmax = np.zeros((dout.size, pool_size))
dmax[np.arange(self.arg_max.size), self.arg_max.flatten()] = dout.flatten()
dmax = dmax.reshape(dout.shape + (pool_size,))
dcol = dmax.reshape(dmax.shape[0] * dmax.shape[1] * dmax.shape[2], -1)
dx = col2im(dcol, self.x.shape, self.pool_h, self.pool_w, self.stride, self.pad)
return dx
common下创建文件trainer.py
添加代码如下:
# coding: utf-8
import sys, os
sys.path.append(os.pardir) # 为了导入父目录的文件而进行的设定
import numpy as np
from common.optimizer import *
class Trainer:
"""进行神经网络的训练的类
"""
def __init__(self, network, x_train, t_train, x_test, t_test,
epochs=20, mini_batch_size=100,
optimizer='SGD', optimizer_param={'lr':0.01},
evaluate_sample_num_per_epoch=None, verbose=True):
self.network = network
self.verbose = verbose
self.x_train = x_train
self.t_train = t_train
self.x_test = x_test
self.t_test = t_test
self.epochs = epochs
self.batch_size = mini_batch_size
self.evaluate_sample_num_per_epoch = evaluate_sample_num_per_epoch
# optimzer
optimizer_class_dict = {'sgd':SGD, 'momentum':Momentum, 'nesterov':Nesterov,
'adagrad':AdaGrad, 'rmsprpo':RMSprop, 'adam':Adam}
self.optimizer = optimizer_class_dict[optimizer.lower()](**optimizer_param)
self.train_size = x_train.shape[0]
self.iter_per_epoch = max(self.train_size / mini_batch_size, 1)
self.max_iter = int(epochs * self.iter_per_epoch)
self.current_iter = 0
self.current_epoch = 0
self.train_loss_list = []
self.train_acc_list = []
self.test_acc_list = []
def train_step(self):
batch_mask = np.random.choice(self.train_size, self.batch_size)
x_batch = self.x_train[batch_mask]
t_batch = self.t_train[batch_mask]
grads = self.network.gradient(x_batch, t_batch)
self.optimizer.update(self.network.params, grads)
loss = self.network.loss(x_batch, t_batch)
self.train_loss_list.append(loss)
if self.verbose: print("train loss:" + str(loss))
if self.current_iter % self.iter_per_epoch == 0:
self.current_epoch += 1
x_train_sample, t_train_sample = self.x_train, self.t_train
x_test_sample, t_test_sample = self.x_test, self.t_test
if not self.evaluate_sample_num_per_epoch is None:
t = self.evaluate_sample_num_per_epoch
x_train_sample, t_train_sample = self.x_train[:t], self.t_train[:t]
x_test_sample, t_test_sample = self.x_test[:t], self.t_test[:t]
train_acc = self.network.accuracy(x_train_sample, t_train_sample)
test_acc = self.network.accuracy(x_test_sample, t_test_sample)
self.train_acc_list.append(train_acc)
self.test_acc_list.append(test_acc)
if self.verbose: print("=== epoch:" + str(self.current_epoch) + ", train acc:" + str(train_acc) + ", test acc:" + str(test_acc) + " ===")
self.current_iter += 1
def train(self):
for i in range(self.max_iter):
self.train_step()
test_acc = self.network.accuracy(self.x_test, self.t_test)
if self.verbose:
print("=============== Final Test Accuracy ===============")
print("test acc:" + str(test_acc))
正向传播时传递了信号的神经元,反向传播时按原样传递信号;正向传播时没有传递信号的神经元,反向传播时信号将停在那里。
使用MNIST数据集进行验证
创建文件overfit_dropout.py
添加代码如下:
# coding: utf-8
import os
import sys
sys.path.append(os.pardir) # 为了导入父目录的文件而进行的设定
import numpy as np
import matplotlib.pyplot as plt
from dataset.mnist import load_mnist
from common.multi_layer_net_extend import MultiLayerNetExtend
from common.trainer import Trainer
(x_train, t_train), (x_test, t_test) = load_mnist(normalize=True)
# 为了再现过拟合,减少学习数据
x_train = x_train[:300]
t_train = t_train[:300]
# 设定是否使用Dropuout,以及比例 ========================
use_dropout = True # 不使用Dropout的情况下为False
dropout_ratio = 0.2
# ====================================================
network = MultiLayerNetExtend(input_size=784, hidden_size_list=[100, 100, 100, 100, 100, 100],
output_size=10, use_dropout=use_dropout, dropout_ration=dropout_ratio)
trainer = Trainer(network, x_train, t_train, x_test, t_test,
epochs=301, mini_batch_size=100,
optimizer='sgd', optimizer_param={'lr': 0.01}, verbose=True)
trainer.train()
train_acc_list, test_acc_list = trainer.train_acc_list, trainer.test_acc_list
# 绘制图形==========
markers = {'train': 'o', 'test': 's'}
x = np.arange(len(train_acc_list))
plt.plot(x, train_acc_list, marker='o', label='train', markevery=10)
plt.plot(x, test_acc_list, marker='s', label='test', markevery=10)
plt.xlabel("epochs")
plt.ylabel("accuracy")
plt.ylim(0, 1.0)
plt.legend(loc='lower right')
plt.show()
通过使用Dropout,训练数据和测试数据的识别精度的差距变小了。并且,训练数据也没有到达100%的识别精度。像这样,通过使用Dropout,即便是表现力强的网络,也可以抑制过拟合。
机器学习中经常使用集成学习。所谓集成学习,就是让多个模型单独进行学习,推理时再取多个模型的输出的平均值。用神经网络的语境来说,比如,准备5个结构相同(或者类似)的网络,分别进行学习,测试时,以这5个网络的输出的平均值作为答案。
通过进行集成学习,神经网络的识别精度可以提高好几个百分点。
超参数是指,比如各层的神经元数量、batch大小、参数更新时的学习率或权值衰减等。如果这些超参数没有设置合适的值,模型的性能就会很差。虽然超参数的取值非常重要,但是在决定超参数的过程中一般会伴随很多的试错。
不能使用测试数据评估超参数的性能。
因为如果使用测试数据调整超参数,超参数的值会对测试数据发生过拟合。换句话说,用测试数据确认超参数的值的“好坏”,就会导致超参数的值被调整为只拟合测试数据。这样的话,可能就会得到不能拟合其他数据、泛化能力低的模型。
因此,调整超参数时,必须使用超参数专用的确认数据。用于调整超参数的数据,一般称为验证数据
训练数据用于参数(权重和偏置)的学习,验证数据用于超参数的性能评估。
进行超参数的最优化时,逐渐缩小超参数的“好值”的存在范围非常重要。所谓逐渐缩小范围,是指一开始先大致设定一个范围,从这个范围中随机选出一个超参数(采样),用这个采样到的值进行识别精度的评估;然后,多次重复该操作,观察识别精度的结果,根据这个结果缩小超参数的“好值”的范围。通过重复这一操作,就可以逐渐确定超参数的合适范围。
步骤0
设定超参数的范围。
步骤1
从设定的超参数范围中随机采样。
步骤2
使用步骤1中采样到的超参数的值进行学习,通过验证数据评估识别精度(但是要将epoch设置得很小)。
步骤3
重复步骤1和步骤2(100次等),根据它们的识别精度的结果,缩小超参数的范围。
在超参数的最优化中,如果需要更精炼的方法,可以使用贝叶斯最优化
创建文件hyperparameter_optimization.py
添加代码如下:
# coding: utf-8
import sys, os
sys.path.append(os.pardir) # 为了导入父目录的文件而进行的设定
import numpy as np
import matplotlib.pyplot as plt
from dataset.mnist import load_mnist
from common.multi_layer_net import MultiLayerNet
from common.util import shuffle_dataset
from common.trainer import Trainer
(x_train, t_train), (x_test, t_test) = load_mnist(normalize=True)
# 为了实现高速化,减少训练数据
x_train = x_train[:500]
t_train = t_train[:500]
# 分割验证数据
validation_rate = 0.20
validation_num = x_train.shape[0] * validation_rate
x_train, t_train = shuffle_dataset(x_train, t_train)
# x_val = x_train[:validation_num]
# t_val = t_train[:validation_num]
# x_train = x_train[validation_num:]
# t_train = t_train[validation_num:]
x_val = x_train[:int(validation_num)]
t_val = t_train[:int(validation_num)]
x_train = x_train[int(validation_num):]
t_train = t_train[int(validation_num):]
def __train(lr, weight_decay, epocs=50):
network = MultiLayerNet(input_size=784, hidden_size_list=[100, 100, 100, 100, 100, 100],
output_size=10, weight_decay_lambda=weight_decay)
trainer = Trainer(network, x_train, t_train, x_val, t_val,
epochs=epocs, mini_batch_size=100,
optimizer='sgd', optimizer_param={'lr': lr}, verbose=False)
trainer.train()
return trainer.test_acc_list, trainer.train_acc_list
# 超参数的随机搜索======================================
optimization_trial = 100
results_val = {}
results_train = {}
for _ in range(optimization_trial):
# 指定搜索的超参数的范围===============
weight_decay = 10 ** np.random.uniform(-8, -4)
lr = 10 ** np.random.uniform(-6, -2)
# ================================================
val_acc_list, train_acc_list = __train(lr, weight_decay)
print("val acc:" + str(val_acc_list[-1]) + " | lr:" + str(lr) + ", weight decay:" + str(weight_decay))
key = "lr:" + str(lr) + ", weight decay:" + str(weight_decay)
results_val[key] = val_acc_list
results_train[key] = train_acc_list
# 绘制图形========================================================
print("=========== Hyper-Parameter Optimization Result ===========")
graph_draw_num = 20
col_num = 5
row_num = int(np.ceil(graph_draw_num / col_num))
i = 0
for key, val_acc_list in sorted(results_val.items(), key=lambda x:x[1][-1], reverse=True):
print("Best-" + str(i+1) + "(val acc:" + str(val_acc_list[-1]) + ") | " + key)
plt.subplot(row_num, col_num, i+1)
plt.title("Best-" + str(i+1))
plt.ylim(0.0, 1.0)
if i % 5: plt.yticks([])
plt.xticks([])
x = np.arange(len(val_acc_list))
plt.plot(x, val_acc_list)
plt.plot(x, results_train[key], "--")
i += 1
if i >= graph_draw_num:
break
plt.show()
学习率在0.001到0.01、权值衰减系数在10−8到10−6之间时,学习可以顺利进行。像这样,观察可以使学习顺利进行的超参数的范围,从而缩小值的范围。然后,在这个缩小的范围中重复相同的操作。学习率在0.001到0.01、权值衰减系数在10−8到10−6之间时,学习可以顺利进行。像这样,观察可以使学习顺利进行的超参数的范围,从而缩小值的范围。然后,在这个缩小的范围中重复相同的操作。