A system to manage machine learning models
Project description
A basic python client for working with ModelDB machine learning management system.
Visit the original project at : https://github.com/mitdbg/modeldb
This project goal is to try isolate a basic client to operate within model database, where main features will be:
Sync projects/models from file (json, yaml)
Sync projects/models by using basic mechanism (without sklearn dependencies)
Python 2.7 and 3.{5,6} compatibility.
Quick start
Install
Remotely
The most widely known way to install a python package is by pip command. The python package is available at [pypi repository](https://pypi.org/project/modeldb-basic/) (legacy repo [here](https://pypi.python.org/pypi/modeldb-basic)).
Just type this pip command to install it from pypi package repository:
pip install modeldb-basic
Alternatively it’s possible to install it by using any of these URLs:
pip install git+https://github.com/engapa/modeldb-basic[@<git_ref>]#egg=modeldb-basic
pip install <release_file>
Where [@<git_ref>] is an optional reference to a git reference (i.e: @master, @v0.1.6) and <release_file> is the URL of one release file at https://github.com/engapa/modeldb-basic/releases
Locally
Previously downloaded in your host, somebody may install the package by typing:
pip install .
or:
python setup.py install
