python保存对象至本地磁盘pickle和joblib速度对比

太长不看版：半斤八两。

测试的文件3.4G:

保存至本地和加载回来的时间对比：

保存至本地joblib更快一点，加载回来pickle更快一点。

 

本地的文件大小：

 

 

代码：

import _pickle as cPickle
import joblib

joblib.dump(train ,'preprocess/train_joblib.pkl')
train_joblib=joblib.load('preprocess/train_joblib.pkl')

cPickle.dump(train,open("preprocess/train_pickle.pkl","wb")) 
train_pickle = cPickle.load(open("preprocess/train_pickle.pkl","rb"))