[人工智能]python 机器学习模型模型的保存与加载实现

介绍两种保存训练好的模型方法

joblib

import joblib
joblib.dump(model, ‘model1.pkl’) #保存模型，后缀为 .pkl
pre = joblib.load(‘model1.pkl’) #加载模型
代码实现：

from matplotlib import pyplot as plot
import numpy as np
from sklearn import linear_model
import joblib 
#创建数据集生成50到60，shape=（20，1）的随机二维数组
X_train = np.random.randint(50,60,size=(20,1))
Y_train = np.random.randint(50,60,size=(20,1))
# 建立线性模型
model1 = linear_model.LinearRegression()
model1.fit(X_train, Y_train)

# 保存 model
joblib .dump(model1, 'model1.pkl')
print("模型保存成功")

# 加载 model
pre = joblib .load('model1.pkl')
Y_pred = pre.predict(X_train)  #使用predict预测

# 可视化
# 1.训练集数据
plot.scatter(X_train, Y_train, color='green')
# 2.线性预测数据
plot.plot(X_train, Y_pred, color='red')
plot.show()

结果：

pickle

import pickle
f = open(‘model2.pkl’, ‘wb’) # 保存模型，后缀为 .pkl
pickle.dump(model2, f)
f.close()

f = open(’.model2.pkl’, ‘rb’) # 加载模型
pre = pickle.load(f)
f.close()

代码示例：

from matplotlib import pyplot as plot
import numpy as np
from sklearn import linear_model
import pickle

#创建数据集生成50到60，shape=（20，1）的随机二维数组
X_train = np.random.randint(50,60,size=(20,1))
Y_train = np.random.randint(50,60,size=(20,1))
# 建立线性模型
model2 = linear_model.LinearRegression()
model2.fit(X_train, Y_train)

# 保存 model
f = open('model2.pkl', 'wb')
pickle.dump(model2, f)
f.close()

print("模型保存成功")

# 加载 model
f = open('model2.pkl', 'rb')
pre = pickle.load(f)
f.close()

Y_pred = pre.predict(X_train)

# 可视化
# 1.训练集数据
plot.scatter(X_train, Y_train, color='green')
# 2.测试数据
plot.plot(X_train, Y_pred, color='red')
plot.show()

结果：