[英]Transfer learning bad accuracy
I have a task to classify seeds depending on the defect. 我有一项任务是根据缺陷对种子进行分类。 I have around 14k images in 7 classes (they are not equal size, some classes have more photos, some have less). 我在7个班级中有大约14k图像(它们的大小不同,有些班级有更多的照片,有些班级有更少的照片)。 I tried to train Inception V3 from scratch and I've got around 90% accuracy. 我试图从头开始训练初始V3,我的准确率大约为90%。 Then I tried transfer learning using pre-trained model with ImageNet weights. 然后我尝试使用具有ImageNet权重的预训练模型进行转移学习。 I imported inception_v3
from applications
without top fc layers, then added my own like in documentation. 我从没有顶级fc层的applications
导入了inception_v3
,然后在文档中添加了我自己的。 I ended with the following code: 我以下面的代码结束:
# Setting dimensions
img_width = 454
img_height = 227
###########################
# PART 1 - Creating Model #
###########################
# Creating InceptionV3 model without Fully-Connected layers
base_model = InceptionV3(weights='imagenet', include_top=False, input_shape = (img_height, img_width, 3))
# Adding layers which will be fine-tunned
x = base_model.output
x = GlobalAveragePooling2D()(x)
x = Dense(1024, activation='relu')(x)
predictions = Dense(7, activation='softmax')(x)
# Creating final model
model = Model(inputs=base_model.input, outputs=predictions)
# Plotting model
plot_model(model, to_file='inceptionV3.png')
# Freezing Convolutional layers
for layer in base_model.layers:
layer.trainable = False
# Summarizing layers
print(model.summary())
# Compiling the CNN
model.compile(optimizer = 'adam', loss = 'categorical_crossentropy', metrics = ['accuracy'])
##############################################
# PART 2 - Images Preproccessing and Fitting #
##############################################
# Fitting the CNN to the images
train_datagen = ImageDataGenerator(rescale = 1./255,
rotation_range=30,
width_shift_range=0.2,
height_shift_range=0.2,
shear_range = 0.2,
zoom_range = 0.2,
horizontal_flip = True,
preprocessing_function=preprocess_input,)
valid_datagen = ImageDataGenerator(rescale = 1./255,
preprocessing_function=preprocess_input,)
train_generator = train_datagen.flow_from_directory("dataset/training_set",
target_size=(img_height, img_width),
batch_size = 4,
class_mode = "categorical",
shuffle = True,
seed = 42)
valid_generator = valid_datagen.flow_from_directory("dataset/validation_set",
target_size=(img_height, img_width),
batch_size = 4,
class_mode = "categorical",
shuffle = True,
seed = 42)
STEP_SIZE_TRAIN = train_generator.n//train_generator.batch_size
STEP_SIZE_VALID = valid_generator.n//valid_generator.batch_size
# Save the model according to the conditions
checkpoint = ModelCheckpoint("inception_v3_1.h5", monitor='val_acc', verbose=1, save_best_only=True, save_weights_only=False, mode='auto', period=1)
early = EarlyStopping(monitor='val_acc', min_delta=0, patience=10, verbose=1, mode='auto')
#Training the model
history = model.fit_generator(generator=train_generator,
steps_per_epoch=STEP_SIZE_TRAIN,
validation_data=valid_generator,
validation_steps=STEP_SIZE_VALID,
epochs=25,
callbacks = [checkpoint, early])
But I've got terrible results: 45% accuracy. 但是我得到了可怕的结果:45%的准确率。 I thought it should be better. 我认为应该会更好。 I have some hypothesis what could go wrong: 我有一些假设可能会出错:
preprocessing_function=preprocess_input
(found article on the web that it is extremely important, so I decided to add that). 虽然转移学习我使用了preprocessing_function=preprocess_input
(在网上发现这篇文章非常重要,所以我决定添加它)。 rotation_range=30
, width_shift_range=0.2
, height_shift_range=0.2
, and horizontal_flip = True
while transfer learning to augment data even more. 添加了rotation_range=30
, width_shift_range=0.2
, height_shift_range=0.2
和horizontal_flip = True
同时传输学习更加增强数据。 Or did I failed something else? 或者我失败了什么?
EDIT: I post a plot of training history. 编辑:我发布了一段训练历史。 Maybe it contains valuable information: 也许它包含有价值的信息:
EDIT2: With changing parameters of InceptionV3: EDIT2:改变InceptionV3的参数:
VGG16 for comparison: VGG16进行比较:
@today, I found a problem. @today,我发现了一个问题。 It is because of some changes in Batch Normalisation layers and its behavior while freezing them. 这是因为Batch Normalization图层中的一些更改及其冻结时的行为。 Mr. Chollet gave a workaround, but I used a Keras fork made by datumbox, which solved my problem. Chollet先生给出了一个解决方法,但我使用了由datumbox制作的Keras前叉,这解决了我的问题。 The main problem is described here: 主要问题在这里描述:
https://github.com/keras-team/keras/pull/9965 https://github.com/keras-team/keras/pull/9965
Now I get ~85% accuracy and am trying to raise it. 现在我的准确率达到了85%,我正在努力提高它。
If you want to preprocess the input using the preprocess_input
method from Keras, then remove the rescale=1./255
argument. 如果要使用Keras中的preprocess_input
方法预处理输入,请删除rescale=1./255
参数。 Otherwise, keep the rescale
argument and remove the preprocessing_function
argument. 否则,请保留rescale
参数并删除preprocessing_function
参数。 Plus, try a lower learning rate like 1e-4 or 3e-5 or 1e-5 (the default learning rate of Adam optimizer is 1e-3) if loss does not decrease: 另外,如果损失不减少,请尝试较低的学习率,如1e-4或3e-5或1e-5(Adam优化器的默认学习率为1e-3):
from keras.optimizers import Adam
model.compile(optimizer = Adam(lr=learning_rate), ...)
Edit: After adding the training plot, you can see that is it overfitting on the training set. 编辑:添加训练图后,您可以看到它在训练集上过度拟合。 You can: 您可以:
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.