README.md

Digit recognition for MSE

The aim of this example is to provide an app to perform digit recognition from a drawn canvas, based on a trained model, using tensorflow and keras.

Deploy your application

$ mse cloud deploy  # in same folder as mse.toml

Your application is now ready to be used.

Test it

$ mse cloud test <APP_ID>

Use it

With gradio

Using gradio, you can run locally the given gradio interface :

$ pip install gradio
$ gradio client/client.py

And access the interface at the given local url.

Using curl

Get the SSL certificate (without checking the trustworthiness of the enclave):

$ # replace $APP_DOMAIN_NAME with your own app domaine name
$ openssl s_client -showcerts -connect $APP_DOMAIN_NAME:443 </dev/null 2>/dev/null | openssl x509 -outform PEM > cert.pem

You can also get the certificate and check it using:

$ mse cloud verify "$APP_DOMAIN_NAME"

then just query your trusted microservice with the given client:

$ # launch local python server
$ python client/server.py
$ # open html page, provide the domain name of your deployed app, and draw a digit