Component interfaces of the MAMMOth fairness toolkit.
Project description
MAMMOth-commons
Fast module development for the MAMMOth fairness toolkit. Modules refer to model loaders, dataset loaders, or metrics. The library holds common datatypes that are shared between modules, and automates the integration strategy by only needing to add a decorator. It also provides integration tests, as well as a lightweight demonstrator that is a thinned down version of the toolkit.
Some demonstration data are large and stored with LFS. Here is how to clone this repo without retrieving those files. You can pull them on demand.
Linux
GIT_LFS_SKIP_SMUDGE=1 git clone https://github.com/mammoth-eu/mammoth-commons.git
git lfs pull --include "./data/torch_model/ir50_adaface.pth" # get a large file from lfs
Windows
set set GIT_LFS_SKIP_SMUDGE=1
git clone https://github.com/mammoth-eu/mammoth-commons.git
git lfs pull --include "./data/torch_model/ir50_adaface.pth" # get a large file from lfs
:microscope: Investigate fairness
Instructions to quickly launch and install the demonstrator web application locally in your machine:
- Download this repository.
- Create a virtual environment. This is optional but recommended.
- Install dependencies with
pip install -r requirements[test].txt
. This can take a bit of time to download and install everything, but you will be able to run all modules and interface with most popular data types. - Launch the local app server with
python demonstrator/app.py
. When everything is ready, this script will also open a browser window to the app's serving page athttp://localhost:5050
.
:clipboard: Catalogue
Find a catalogue of modules implemented by the MAMMOth consortium
here. These modules are developed by and
depend on datatypes found in the main commons library that resides
under this repo's mammoth/
directory.
:thumbsup: Contributing
Instructions on how to add new modules are here. Use the GitHub issue tracker to ask questions, request features/improvements for the core library or modules, or report bugs.
Project details
Release history Release notifications | RSS feed
Download files
Download the file for your platform. If you're not sure which to choose, learn more about installing packages.
Source Distributions
Built Distribution
File details
Details for the file MAMMOth_commons-0.0.30-py3-none-any.whl
.
File metadata
- Download URL: MAMMOth_commons-0.0.30-py3-none-any.whl
- Upload date:
- Size: 42.6 kB
- Tags: Python 3
- Uploaded using Trusted Publishing? No
- Uploaded via: twine/5.1.1 CPython/3.12.4
File hashes
Algorithm | Hash digest | |
---|---|---|
SHA256 | 34350929736ad401e0330b218716f9f27b6a459f1b819dcc6bec7261fec568a6 |
|
MD5 | 923ac51c30c204b4b3a4118c4505e6c1 |
|
BLAKE2b-256 | ca25ea3abefd6cc92662f465f662f793d3505ac6d6456f49059cfc12cb9866d9 |