RevoScalePy and MicrosoftML examples
Switch branches/tags
Nothing to show
Clone or download
Latest commit ef00418 Dec 11, 2017

readme.md

Samples for using RevoScalePy and MicrosoftML packages

revoscalepy and microsoftml are machine learning libraries provided by Microsoft. They contain many battled tested and high performance machine learning algorithms.

This gallery will showcase how to use 'revoscalepy" and 'microsoftml' for predictive analytics. There are 2 samples in this gallery:

  1. Wine quality prediction with revoscalepy functions (Notebook file" 'revoscalepy_wine_prediction.ipynb'; Python script: revoscalepy_wine_prediction.py)
  2. Adult census analysis with microsoftml functions (Notebook file: microsoftml_adult_census.ipynb; Python script: adult_census.py)

Please note:

  1. 'revoscalepy' and 'microsoftml' don't support Mac OS yet, therefore these samples won't work in Mac OS.
  2. 'microsoftml' is not bundled with Workbench installer, to run Adult census analysis sample, please first install the package.
- For Windows: 'pip install https://rserverdistribution.azureedge.net/production/revoscalepy/9.2.1/wb/1033/d282048eb04046999211535f7368a0a4/windows/microsoftml-1.5.0-py3-none-any.whl'
- For Linux (used in Docker): 'pip install https://rserverdistribution.azureedge.net/production/revoscalepy/9.2.1/wb/1033/d282048eb04046999211535f7368a0a4/linux/microsoftml-1.5.0-py3-none-any.whl'

Wine quality prediction sample using revoscalepy package

Wine quality prediction

Please pip install matplotlib before try this sample, if it has not been installed.

QuickStart

Navigate to the "Notebooks", open revoscalepy_wine_prediction.ipynb, click Start the Notebook Server button, and execute the script to run through the sample.

Quick CLI references

If you want to try exercising this sample from the command line, here are some things to try:

First, launch the Command Prompt or Powershell from the File menu.

Run in local Python environment.

$ az ml experiment submit -c local revoscalepy_wine_prediction.py

Run in a local Docker container. Please ensure your Docker engine allows at least 4 GB or RAM in order for this sample to run in Docker.

$ az ml experiment submit -c docker revoscalepy_wine_prediction.py

Create myvm run configuration to point to a Docker container on a remote VM

$ az ml computetarget attach --name myvm --address <ip address or FQDN> --username <username> --password <pwd> --type remotedocker

# prepare the environment
$ az ml experiment prepare -c myvm

Run revoscalepy_wine_prediction.py script in a Docker container in a remote VM:

$ az ml experiment submit -c myvm revoscalepy_wine_prediction.py

Adult census sample using microsoftml package

Adult census analytis

microsoftml is not bundled with Workbench installer, to run this sample, please first install the package.

  • For Windows:
pip install https://rserverdistribution.azureedge.net/production/revoscalepy/9.2.1/wb/1033/d282048eb04046999211535f7368a0a4/windows/microsoftml-1.5.0-py3-none-any.whl
  • For Linux (used in Docker):
pip install https://rserverdistribution.azureedge.net/production/revoscalepy/9.2.1/wb/1033/d282048eb04046999211535f7368a0a4/linux/microsoftml-1.5.0-py3-none-any.whl

Please pip install matplotlib before try this sample, if it has not been installed.

QuickStart

Navigate to the "Notebooks", open microsoftml_adult_census.ipynb, click Start the Notebook Server button, and execute the script to run through the sample.

Quick CLI references

If you want to try exercising this sample from the command line, here are some things to try:

First, launch the Command Prompt or Powershell from the File menu.

Run in local Python environment.

$ az ml experiment submit -c local adult_census.py

Run in a local Docker container. Please ensure your Docker engine allows at least 4 GB or RAM in order for this sample to run in Docker.

$ az ml experiment submit -c docker-python adult_census.py

Run in a Remove VM Create myvm run configuration to point to a Docker container on a remote VM

$ az ml computetarget attach --name myvm --address <ip address or FQDN> --username <username> --password <pwd> --type remotedocker

# prepare the environment
$ az ml experiment prepare -c myvm

Run adult_census.py script in a Docker container in a remote VM:

$ az ml experiment submit -c myvm adult_census.py