ml-api's Introduction

Deploying your Machine Learning Model

Code and instructions for techniques to quickly and easily deploy your machine learning model.

In this repo, you will find the code and instructions for this article. It is advised to read through the article whilst coding along using the jupyter notebooks in the notebooks folder.

This repo and the corresponding article describe several methods for creating a production-ready API:

FastAPI + Uvicorn
OpenAPI (Swagger) + ReDoc
Docker
Heroku

ml-api's People

Contributors

Stargazers

Watchers

ml-api's Issues

Bad request when using curl

First thanks for your great article in Medium!
I have a problem when implementing your code.

I use your test code as test_api.py:

import requests
to_predict_dict = {'satisfaction_level': 0.38,
                   'last_evaluation': 0.53,
                   'number_project': 2,
                   'average_montly_hours': 157,
                   'time_spend_company': 3,
                   'Work_accident': 0,
                   'promotion_last_5years': 0,
                   'sales': 'support',
                   'salary': 'low'}

url = 'http://127.0.0.1:8000/predict'
r = requests.post(url, json=to_predict_dict)
print(r.json())

It return prediction after excuting.

$ python test_api.py
{'prediction': 1}

And the server log seems work well

INFO:     127.0.0.1:49703 - "POST /predict HTTP/1.1" 200 OK

But when I use the Swagger interface, it returns {"detail":"There was an error parsing the body"}。

Using the curl is the same error.

$ curl -X POST "http://127.0.0.1:8000/predict" -H "accept: application/json" -H "Content-Type: application/json; charset=UTF-8" -d "{'satisfaction_level': 0.38, 'last_evaluation': 0.53, 'number_project': 2, 'average_montly_hours': 157, 'time_spend_company': 3, 'Work_accident': 0, 'promotion_last_5years': 0, 'sales': 'support', 'salary': 'low'}"
{"detail":"There was an error parsing the body"}%

Where is the problem and how to solve it?
Thanks!

Recommend Projects