FAST - API for ML models

For a Normal Machine learning Developers it is easy to develop API for there Machine learning model , what if they need to complete the task with UI to test there api's.

They can create api with the help of Flask Service,which is super easy and light weight but we can test only via POSTMAN.

In this post i am going to say about FASTAPI,with this framework we can able to build Api fastly and we can test with UI too. that was quite easy and helpful

Fast Api

The most two widely created api in a life of Machine learning developer is to test there ML model via single value test and to predict the model by passing a test file

This blog is to whom who wanted to test there api with some help of UI

Prerequesties

1)Pickled Ml Model

2)fast Api Installed

I am going to show both of them in this tutorial:

Lets Dive in:

These are the following requirements for

#These are the dependency of Fastapi
pip3 install fastapi
pip3 install multipart
pip3 install uvicorn


1) The first post request is to test the api with single value data 
     if you need to pre process the data you need to do it after getting the value or converting
     it into a data frame will easy for pre processing so i have done that ,you can do it in many 
     ways 
2) The second post request is to test the model with a test file , make sure that the target column
    is removed in the test dataset.
    Here we are using the Fast API methods called UploadFile and File which helps to get the file
    which ever file is uploaded 

At last to run this app ,make sure you have installed uvicorn in your device and can run using 
 uvicorn.run where you need to pass the ip address and the port number in which yout Fastapi  
   has to run similar to flask->run method





        The above tells how the UI will be exactly to test the first API

The above tells how the UI will be exactly to test the second API
see the full code in my Github if you find it useful please to upvote and share the link.
Do follow me in linkedin