Devo Python ML Model Manager Client

The ML Model Manager is a service to register machine learning models on Devo platform. These models can be used through the query engine using the mlevalmodel(...) operation or through the FLOW correlation engine including in the context the MlSingleModelEval unit.

devo-mlmodelmanager provides an easy-to-use client for Devo’s ML Model Manager. Built on top of the widely used Requests library exposes a simplified interface for model management, allowing you to focus in the machine learning workflows and not worry about the integration with Devo platform.

A simple example

from devo_ml.modelmanager import create_client_from_token

url = "<model-manager-server-url>"
token = "<valid-access-token>"

client = create_client_from_token(url, token)

client.add_model(
   "pokemon_onnx_regression",          # model name
   "ONNX",                             # model engine
   "~/models/pokemon.onnx",            # model file
   description="A funny Pokemon prediction"
)

Requirements

• Python 3.7+

Install

$ pip install devo-mlmodelmanager

Documentation

Explore the documentation to learn more.