tanujjain / deploy-ml-model

Deploying a simple machine learning model to an AWS ec2 instance using flask and docker.

Serve a Machine Learning Model as a Webservice

Serving a simple machine learning model as a webservice using flask and docker.

Getting Started

1. Use Model_training.ipynb to train a logistic regression model on the iris dataset and generate a pickled model file (iris_trained_model.pkl)
2. Use app.py to wrap the inference logic in a flask server to serve the model as a REST webservice:
   • Execute the command python app.py to run the flask app.
   • Go to the browser and hit the url 0.0.0.0:80 to get a message Hello World! displayed. NOTE: A permission error may be received at this point. In this case, change the port number to 5000 in app.run() command in app.py. (Port 80 is a privileged port, so change it to some port that isn't, eg: 5000)
   • Next, run the below command in terminal to query the flask server to get a reply 2 for the model file provided in this repo:

      curl -X POST \
      0.0.0.0:80/predict \
      -H 'Content-Type: application/json' \
      -d '[5.9,3.0,5.1,1.8]'
   

3. Run docker build -t app-iris . to build the docker image using Dockerfile. (Pay attention to the period in the docker build command)
4. Run docker run -p 80:80 app-iris to run the docker container that got generated using the app-iris docker image. (This assumes that the port in app.py is set to 80)
5. Use the below command in terminal to query the flask server to get a reply 2 for the model file provided in this repo:

       curl -X POST \
       0.0.0.0:80/predict \
       -H 'Content-Type: application/json' \
       -d '[5.9,3.0,5.1,1.8]'
   

For details on floating the containerized app on AWS ec2 instance, see the blog.

LICENSE

See LICENSE for details.

deploy-ml-model's Issues

Getting error while sending test data using curl in local server

Getting error while sending test data using curl in local server
app.py: return self.view_functionsrule.endpoint
File "app.py", line 28, in get_prediction
data = np.array(data)[np.newaxis, :] # converts shape from (4,) to (1, 4)
IndexError: too many indices for array
127.0.0.1 - - [31/Mar/2020 16:06:23] "�[35m�[1mPOST /predict HTTP/1.1�[0m" 500 -_

Curl post: _curl -X POST 127.0.0.1:5000/predict -H 'Content-Type:application/json' -d '[5.9,3.0,5.1,1.8]'

<title>500 Internal Server Error</title>

Internal Server Error

The server encountered an internal error and was unable to complete your request. Either the server is overloaded or there is an error in the application.

_

Issue with post request

Hello,

I have followed the instruction after passing the post request, where can I see the result of the request?