[英]CNN with keras, accuracy not improving
I have started with Machine Learning recently, I am learning CNN, I planned to write an application for Car Damage severity detection, with the help of this Keras blog and this github repo .我最近开始学习机器学习,我正在学习 CNN,我计划在这个Keras 博客和这个github repo的帮助下编写一个用于汽车损坏严重程度检测的应用程序。
This is how car data-set looks like:这就是汽车数据集的样子:
F:\WORKSPACE\ML\CAR_DAMAGE_DETECTOR\DATASET\DATA3A
├───training (979 Images for all 3 categories of training set)
│ ├───01-minor
│ ├───02-moderate
│ └───03-severe
└───validation (171 Images for all 3 categories of validation set)
├───01-minor
├───02-moderate
└───03-severe
Following code gives me only 32% of accuracy.下面的代码只给了我 32% 的准确率。
from keras.preprocessing.image import ImageDataGenerator
from keras.models import Sequential
from keras.layers import Conv2D, MaxPooling2D
from keras.layers import Activation, Dropout, Flatten, Dense
from keras import backend as K
# dimensions of our images.
img_width, img_height = 150, 150
train_data_dir = 'dataset/data3a/training'
validation_data_dir = 'dataset/data3a/validation'
nb_train_samples = 979
nb_validation_samples = 171
epochs = 10
batch_size = 16
if K.image_data_format() == 'channels_first':
input_shape = (3, img_width, img_height)
else:
input_shape = (img_width, img_height, 3)
model = Sequential()
model.add(Conv2D(32, (3, 3), input_shape=input_shape))
model.add(Activation('relu'))
model.add(MaxPooling2D(pool_size=(2, 2)))
model.add(Conv2D(32, (3, 3)))
model.add(Activation('relu'))
model.add(MaxPooling2D(pool_size=(2, 2)))
model.add(Conv2D(64, (3, 3)))
model.add(Activation('relu'))
model.add(MaxPooling2D(pool_size=(2, 2)))
model.add(Flatten())
model.add(Dense(64))
model.add(Activation('relu'))
model.add(Dropout(0.5))
model.add(Dense(1))
model.add(Activation('sigmoid'))
model.compile(loss='binary_crossentropy',
optimizer='rmsprop',
metrics=['accuracy'])
# this is the augmentation configuration we will use for training
train_datagen = ImageDataGenerator(
rescale=1. / 255,
shear_range=0.2,
zoom_range=0.2,
horizontal_flip=True)
# this is the augmentation configuration we will use for testing:
# only rescaling
test_datagen = ImageDataGenerator(rescale=1. / 255)
train_generator = train_datagen.flow_from_directory(
train_data_dir,
target_size=(img_width, img_height),
batch_size=batch_size,
class_mode='binary')
validation_generator = test_datagen.flow_from_directory(
validation_data_dir,
target_size=(img_width, img_height),
batch_size=batch_size,
class_mode='binary')
model.fit_generator(
train_generator,
steps_per_epoch=nb_train_samples // batch_size,
epochs=epochs,
validation_data=validation_generator,
validation_steps=nb_validation_samples // batch_size)
model.save_weights('first_try.h5')
I tried:我试过:
Conv2D
layerConv2D
层中的过滤器大小Conv2D
layer, MaxPooling
layersConv2D
层、 MaxPooling
层adam
, Sgd
, etcadam
、 Sgd
等(1,1) and (5,5)
instead of (3,3)
(1,1) and (5,5)
而不是(3,3)
(256, 256)
, (64, 64)
from (150, 150)
(150, 150)
更新为(256, 256)
, (64, 64)
(150, 150)
But no luck, every-time I'm getting accuracy up to 32% or less than that but not more.但没有运气,每次我的准确率都达到 32% 或更低,但不会更高。 Any idea what I'm missing.
知道我错过了什么。
As in the github repo we can see, it gives 72% accuracy for the same dataset (Training -979, Validation -171).正如我们在github 存储库中看到的那样,它为相同的数据集提供了 72% 的准确率(训练 -979,验证 -171)。 Why its not working for me.
为什么它不适合我。
I tried his code from the github link on my machine but it hanged up while training the dataset(I waited for more than 8 hours), so changed the approach, but still no luck so far.我在我的机器上从 github 链接尝试了他的代码,但在训练数据集时挂断了(我等了 8 个多小时),所以改变了方法,但到目前为止仍然没有运气。
Here's the Pastebin containing output of my training epochs.下面是引擎收录我的训练时期的含输出。
The issue is caused by a mis-match between the number of output classes (three) and your choice of final layer activation (sigmoid) and loss-function (binary cross entropy).该问题是由输出类的数量(三个)与您选择的最终层激活(sigmoid)和损失函数(二元交叉熵)之间的不匹配引起的。
The sigmoid function 'squashes' real values into a value between [0, 1] but it is designed for binary (two class) problems only. sigmoid 函数将实际值“压缩”为 [0, 1] 之间的值,但它仅适用于二元(两类)问题。 For multiple classes you need to use something like the softmax function.
对于多个类,您需要使用诸如 softmax 函数之类的东西。 Softmax is a generalised version of sigmoid (the two should be equivalent when you have two classes).
Softmax 是 sigmoid 的广义版本(当你有两个类时,两者应该是等价的)。
The loss value also needs to be updated to one that can handle multiple classes - categorical cross entropy will work in this case.损失值也需要更新为可以处理多个类的值——分类交叉熵在这种情况下会起作用。
In terms of code, if you modify the model definition and compilation code to the version below it should work.在代码方面,如果将模型定义和编译代码修改为以下版本应该可以工作。
model = Sequential()
model.add(Conv2D(32, (3, 3), input_shape=input_shape))
model.add(Activation('relu'))
model.add(MaxPooling2D(pool_size=(2, 2)))
model.add(Conv2D(32, (3, 3)))
model.add(Activation('relu'))
model.add(MaxPooling2D(pool_size=(2, 2)))
model.add(Conv2D(64, (3, 3)))
model.add(Activation('relu'))
model.add(MaxPooling2D(pool_size=(2, 2)))
model.add(Flatten())
model.add(Dense(64))
model.add(Activation('relu'))
model.add(Dropout(0.5))
model.add(Dense(3))
model.add(Activation('softmax'))
model.compile(loss='categorical_crossentropy',
optimizer='rmsprop',
metrics=['accuracy'])
Finally you need to specify class_mode='categorical'
in your data generators.最后,您需要在数据生成器中指定
class_mode='categorical'
。 That will ensure that the output targets are formatted as a categorical 3-column matrix that has a one in the column corresponding to the correct value and zeroes elsewhere.这将确保输出目标被格式化为一个分类的 3 列矩阵,该矩阵在对应于正确值的列中有一个 1,在其他地方有 0。 This response format is needed by the
categorical_cross_entropy
loss function. categorical_cross_entropy
损失函数需要这种响应格式。
Minor correction:小修正:
model.add(Dense(1))
Should be:应该:
model.add(Dense(3))
It has to comply with number of classes in the output.它必须符合输出中的类数。
声明:本站的技术帖子网页,遵循CC BY-SA 4.0协议,如果您需要转载,请注明本站网址或者原文地址。任何问题请咨询:yoyou2525@163.com.