1

无法在使用 ktrain 模型的前端 streamlit 上进行预测，请提供有关如何为预测功能提供输入的建议。

基本上，我想了解如何为我保存的 ktrain 回归模型提供输入，以便我可以将其合并到简化的 web 应用程序按钮中。

我已经尝试将数组、列表和数据框作为 .predict 函数中的参数，但似乎仍然缺少一些东西。在点击预测按钮时出现值错误。

import streamlit as st
from PIL import Image 
import pandas as pd
from tensorflow import keras
model = keras.models.load_model("predictor.h5")


st.write("This is an application to calculate Employee Mental Fatigue Score")
image = Image.open("IMG_2605.jpeg")
st.image(image, use_column_width=True)

WFH_Setup_Available = st.text_input("is work from home enabled for you?")
Designation =st.text_input("what is your designation?")
Average_hours_worked_per_day = st.text_input("how many hours you work on an average per day?")
Employee_satisfaction_score = st.text_input("Please enter your satisfaction score on scale of 10")
data = ['WFH_Setup_Available', 'Designation', 'Average_hours_worked_per_day' , 'Employee_satisfaction_score']


def mental_fatigue_score(WFH_Setup_Available, Designation, Average_hours_worked_per_day, Employee_satisfaction_score):
  prediction = model.predict([[WFH_Setup_Available, Designation, Average_hours_worked_per_day, Employee_satisfaction_score]])
  print(prediction)
  return prediction


if st.button("Predict"):
  result= mental_fatigue_score(WFH_Setup_Available, Designation, Average_hours_worked_per_day, Employee_satisfaction_score)
  st.success('The output is {}'.format(result))

在此处输入图像描述 请建议如何为 streamlit web 应用程序的 .predict 函数提供输入。我已经使用 ktrain 回归器训练了预测器。

4

1 回答 1

1

通过将ktrain模型保存为我自己解决了这个问题

predictor.save('predictor')
predictor = ktrain.load_predictor('predictor')

当我保存为预测器时，它会创建一个文件夹，其中有一个tf_mode.h5和tf_model.preproc。

这比我预期的要容易。

进一步的火车输入应该是如下数据框 -

data = {'WFH_Setup_Available':WFH_Setup_Available,'Designation':Designation, 'Company_Type':Company_Type, 
        'Average_hours_worked_per_day': Average_hours_worked_per_day, 'Employee_satisfaction_score': Employee_satisfaction_score}

data = pd.DataFrame([data])
于 2021-05-17T08:29:51.380 回答