如何在'执行python脚本'部分中将外部pickle文件内容作为数据框加载到azure ml studio中？

Question

这是我正在使用的脚本：

 import warnings
 warnings.filterwarnings('ignore') 
 import pandas as pd
 import sys
 import pickle
 def azureml_main(dataframe1 = None, dataframe2 = None):
 sys.path.append('.\\Script Bundle')
 dataframe1 = pickle.load(open(r'/Script Bundle/descript.pkl', 'rb'))
 return dataframe1,

但是当我执行它时，低于错误

 FileNotFoundError: [Errno 2] No such file or directory: '/Script 
 Bundle/descript.pkl'
 Process returned with non-zero exit code 1

Answer 1

该模型在'./Script Bundle'下可用，你不需要整个sys.path.append的东西。所以你可以使用类似下面的代码：

import pandas as pd
import pickle

def azureml_main(dataframe1 = None, dataframe2 = None):

    model = pickle.load( open( "./Script Bundle/iris.pkl", "rb" ) )
    cleanFrame = dataframe1[['sepal-length','sepal-width','petal-length','petal-width']]

    prediction = model.predict(cleanFrame)

    dataframe1['Scored Labels'] = prediction

    return dataframe1

你可以看一下this ML Studio experiment的例子。