在 Azure Web 服务中而非 ML 实验中执行 python 脚本时出现错误 0085

Error 0085 while executing python script in Azure Web service but not in ML Experiment

我的工作流程 运行 在实验上是完美的,但是在部署到 Web 服务后,我在 post 时收到此错误。

Python代码:

# -*- coding: utf-8 -*-

#import sys
import pickle
import pandas as pd
from sklearn.tree import DecisionTreeClassifier
from sklearn import tree 

def azureml_main(dataframe1 = None, dataframe2 = None):
    print('input dataframe1 ',dataframe1)
    decision_tree_pkl_predictive_maint = r'.\Script Bundle\decision_tree_pkl_predictive_maint.pkl'

    #sys.path.insert(0,".\Script Bundle")
    #model = pickle.load(open(".\Script Bundle\decision_tree_pkl_predictive_maint.pkl", 'rb'))

    modle_file = open(decision_tree_pkl_predictive_maint,"rb")
    model = pickle.load(modle_file)

    #return the mode of prediciton
    result = model.predict(dataframe1)
    print(result)
    result_df = pd.DataFrame({'prediction_class':result})
    return result_df,

错误:

执行 Python 脚本 RRS:错误 0085:脚本评估期间发生以下错误,请查看输出日志以获取更多信息:------------ 错误消息开始于 Python 解释器 ---------- 执行函数时捕获异常:Traceback(最近调用最后一次):文件“\server\InvokePy.py”,第 120 行,在 executeScript outframe = mod.azureml_main(*inframes) 文件“\temp-1036260731852293620.py”,第 46 行,在 azureml_main modle_file = open(decision_tree_pkl_predictive_maint,"rb") FileNotFoundError: [ Errno 2] No such file or directory: '.\Script Bundle\decision_tree_pkl_predictive_maint.pkl' -------- 来自 Python 解释器的错误消息结束 -------- --

请指教

问题与您的文件路径有关。确保您包含了正确的路径。