机器学习模型的保存和调用_机器学习模型如何保存和调用

1.使用python自带的pickle

from sklearn.ensemble import RandomForestClassifier
from sklearn import datasets
import pickle

#方法一:python自带的pickle
(X,y) = datasets.load_iris(return_X_y=True)
rfc = RandomForestClassifier(n_estimators=100,max_depth=100)
rfc.fit(X,y)
print(rfc.predict(X[0:1,:]))
#save model
f = open('saved_model/rfc.pickle','wb')
pickle.dump(rfc,f)
f.close()
#load model
f = open('saved_model/rfc.pickle','rb')
rfc1 = pickle.load(f)
f.close()
print(rfc1.predict(X[0:1,:]))
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18

2.使用sklearn中的模块joblib

使用joblib模块更加的简单了，核心代码就两行

from sklearn.ensemble import RandomForestClassifier
from sklearn import datasets
from sklearn.externals import joblib
#方法二：使用sklearn中的模块joblib
(X,y) = datasets.load_iris(return_X_y=True)
rfc = RandomForestClassifier(n_estimators=100,max_depth=100)
rfc.fit(X,y)
print(rfc.predict(X[0:1,:]))
#save model
joblib.dump(rfc, 'saved_model/rfc.pkl')
#load model
rfc2 = joblib.load('saved_model/rfc.pkl')
print(rfc2.predict(X[0:1,:]))
1
2
3
4
5
6
7
8
9
10
11
12
13

调用

#save model
joblib.dump(rfc, 'saved_model.pkl')
#load model
rfc2 = joblib.load('saved_model.pkl')
1
2
3
4

rfc2继续训练预测