Model saving and loading of sklearn

'''
 After training the model, save it locally , And reload the model for two methods of prediction 
'''
#  Method 1:pickle
from sklearn import svm
from sklearn import datasets 
import pickle 
clf = svm.SVC()
iris = datasets.load_iris() 
X,y = iris.data, iris.target 
clf.fit(X,y)
#  Save model 
with open('clf.pickle','wb') as f: 
    pickle.dump(clf,f) # Will train the good model clf Stored in variables f in , And save to local 

#  Reloading and using the model 
with open('clf.pickle','rb') as f:  
    clf_load = pickle.load(f)  # Store the model in variables clf_load in   
    print(clf_load.predict(X[0:1000])) # Call the model and predict the results 

#  Method 2:joblib Method 
from sklearn import svm 
from sklearn import datasets
import joblib
# sklearn.externals.joblib Function is used in 0.21 And previous versions of , In the latest version , This function should be deprecated and imported directly joblib
# from sklearn.externals import joblib
clf = svm.SVC()
iris = datasets.load_iris() 
X,y = iris.data, iris.target 
clf.fit(X,y) 
#  Save the trained ones clf Model  
joblib.dump(clf,'clf.pkl',compress=3) 
#  Reload the trained clf Model  
clf3 = joblib.load('clf.pkl')  
print(clf3.predict(X[0:1000]))  #  Print forecast 
#  Comparatively speaking ,joblib More efficient , but 32 position python and 64 position python It cannot be used seamlessly , You can check the situation of this machine through the following code 
import platform 
print(platform.architecture())