Samuel Lynn-Evans@QbitAI
![3a814edc668b22b1e7e7b2efd51d4af0.png](https://i-blog.csdnimg.cn/blog_migrate/89d423818a98746259457bcd6e591c23.png)
差分学习率(Differential Learning rates)意味着在训练时变换网络层比提高网络深度更重要。
基于已有模型来训练深度学习网络,这是一种被验证过很可靠的方法,可以在计算机视觉任务中得到更好的效果。
大部分已有网络(如Resnet、VGG和Inception等)都是在ImageNet数据集训练的,因此我们要根据所用数据集与ImageNet图像的相似性,来适当改变网络权重。
在修改这些权重时,我们通常要对模型的最后几层进行修改,因为这些层被用于检测基本特征(如边缘和轮廓),不同数据集有着不同基本特征。
首先,要使用Fast.ai库来获得预训练的模型,代码如下:
from fastai.conv_learner import *
# import library for creating learning object for convolutional #networks
model = VVG16()
# assign model to resnet, vgg, or even your own custom model
PATH = './folder_containing_images'
data = ImageClassifierData.from_paths(PATH)
# create fast ai data object, in this method we use from_paths where
# inside PATH each image class is separated into different folders
learn = ConvLearner.pretrained(model, data, precompute=True)
# create a learn object to quickly utilise state of the art
# techniques from the fast ai library
创建学习对象之后(learn object),通过快速冻结前面网络层并微调后面网络层来解决问题:
learn.freeze()
# freeze layers up to the last one, so weights will not be updated.
learning_rate = 0.1
learn.fit(learning_rate, epochs=3)
# train only the last layer for a few epochs
当后面网络层产生了良好效果,我们会应用差分学习率来改变前面网络层。在实际中,一般将学习率的缩小倍数设置为10倍:
learn.unfreeze()
# set requires_grads to be True for all layers, so they can be updated
learning_rate = [0.001, 0.01, 0.1]
# learning rate is set so that deepest third of layers have a rate of 0.001, # middle layers have a rate of 0.01,