Example of Microsoft Machine Learning Server usage, which shows how to deploy trained model and then use it
In order to be able to test Microsoft Machine Learning Server (MMLS) first of all you should install it
https://docs.microsoft.com/en-us/machine-learning-server/install/machine-learning-server-windows-install
Don't forget to deploy it, also remember credentials.
az ml admin bootstrap
Next step
Install keras and tensorflow to conda, which is inside of MMLS or you can use pip. In my case it was:
pip3 install --upgrade tensorflow
conda install -c conda-forge keras
Next is model.
You can download models from model folder, or you can train a new model.
MMLS communcation.
Generally we can deploy some functionality for further usage, later destroy.
In main file you will see authentication process and 3 functions: deploy, explore, destroy.
Connection
Put your login and password here. You got those after setting up MMLS.
context = (login, password)
Usage
Now simply run deploy -> explore -> destroy.
Don't be afraid of warnings and errors, MMLS produces a lot of them. :)
If everything is fine the output should be simmilar to this:
{'answer': 3}