https://github.com/facultyai/faculty-models
Tooling to load models from the Faculty model registry
https://github.com/facultyai/faculty-models
faculty-platform machine-learning mlflow python
Last synced: about 2 months ago
JSON representation
Tooling to load models from the Faculty model registry
- Host: GitHub
- URL: https://github.com/facultyai/faculty-models
- Owner: facultyai
- Created: 2019-09-30T18:34:08.000Z (over 6 years ago)
- Default Branch: master
- Last Pushed: 2020-07-07T08:58:40.000Z (almost 6 years ago)
- Last Synced: 2025-07-20T14:37:34.734Z (11 months ago)
- Topics: faculty-platform, machine-learning, mlflow, python
- Language: Python
- Homepage: https://faculty.ai/platform/
- Size: 27.3 KB
- Stars: 0
- Watchers: 11
- Forks: 1
- Open Issues: 0
-
Metadata Files:
- Readme: README.rst
- License: LICENSE-2.0.txt
Awesome Lists containing this project
README
faculty-models
==============
``faculty-models`` is a tool to help you use models from the model registry in
Faculty Platform.
.. warning::
This library's API is subject to change as new functionality is added to
the model registry feature in Faculty Platform.
Installation
------------
``faculty-models`` comes preinstalled in Python environments available in
Faculty platform. To use it externally, install it from PyPI with ``pip``:
.. code-block:: bash
pip install faculty-models
If you've not already done so on the computer you're using, you'll also need to
generate and store CLI credentials for the Platform. You can do this with
`the Faculty CLI
`_.
Usage
-----
The model registry in Faculty Platform includes a feature that helps you
generate the snippets you need. It will help you get the project and model IDs
you need to use ``faculty-models``.
If your model is in the `MLmodel format
`_ (likely because you used `MLflow
`_ to store it), you can load it directly back into Python
with:
.. code-block:: python
import faculty_models
model = faculty_models.load_mlmodel(
project_id="998328c3-23df-4225-a3ee-0a53d1409fbd",
model_id="c998fca9-e093-47ea-9896-8f75db695b91"
)
Otherwise, you can use the following to download the contents of the model to
the local filesystem. ``download`` returns the path of the downloaded model
files:
.. code-block:: python
import faculty_models
path = faculty_models.download(
project_id="998328c3-23df-4225-a3ee-0a53d1409fbd",
model_id="c998fca9-e093-47ea-9896-8f75db695b91"
)
The above examples always download the latest version of a model. To get a
specific verion, pass the version number when calling either function:
.. code-block:: python
import faculty_models
model = faculty_models.load_mlmodel(
project_id="998328c3-23df-4225-a3ee-0a53d1409fbd",
model_id="c998fca9-e093-47ea-9896-8f75db695b91",
version=4
)
If you only wish to download part of the model, or if you wish to load an
MLmodel that is in a subdirectory of the model, pass the path argument to
either function:
.. code-block:: python
import faculty_models
model = faculty_models.load_mlmodel(
project_id="998328c3-23df-4225-a3ee-0a53d1409fbd",
model_id="c998fca9-e093-47ea-9896-8f75db695b91",
path="sub/path"
)