Merge pull request #26 from ChrisMRuss/workflow

pypi installer

README.md

@@ -4,11 +4,11 @@ OxonFair is an expressive toolkit designed to enforce a wide-range of fairness d
 The toolkit is designed to overcome a range of shortcomings in existing fairness toolkits for high-capacity models that overfit to the training data.
 It is designed and works for computer vision and NLP problems alongside tabular data.
 
-For low-capacity models (, linear regression over a small number of variables, and decision-trees of limited depth), we recommend [fairlearn](https://github.com/fairlearn/fairlearn).
+For low-capacity models (e.g., logistic regression over a small number of variables, and decision-trees of limited depth), we recommend [fairlearn](https://github.com/fairlearn/fairlearn).
 
 We support a range of complex classifiers including [pytorch](https://pytorch.org/), [scikit learn](https://scikit-learn.org/stable/), and ensembles provided by [autogluon](https://auto.gluon.ai/stable/index.html).
 
-It is a modified version of [autogluon.fair](https://github.com/autogluon/autogluon-fair) and actively maintained.
+OxonFair is a modified version of [autogluon.fair](https://github.com/autogluon/autogluon-fair) and actively maintained.
 
 ## Source install
 

examples/training_a_two_head_model/README.md

@@ -1,22 +1,20 @@
-## Setup for Training a Two-Head Model
+# Setup for Training a Two-Head Model
 
-1) Download img_align_celeba.zip and list_attr_celeba.txt from http://mmlab.ie.cuhk.edu.hk/projects/CelebA.html
-and move it to data/celeba/.
+This directory contains files for training a two-headed model on the celbA dataset.
 
-2) Create a virtual environment with python=3.8, and activate it. 
+1) Download img_align_celeba.zip and list_attr_celeba.txt from <http://mmlab.ie.cuhk.edu.hk/projects/CelebA.html>
+and move it to data/celeba/ inside this directory.
 
-3) Install your torch and torchvision. We used torch==1.12.1+cu116 and torchvision==0.13.1+cu116.
+2) Create a virtual environment with python=3.8, and activate it.
 
-````commandline
-pip install torch==1.12.1+cu116 torchvision==0.13.1+cu116 --extra-index-url https://download.pytorch.org/whl/cu116
-````
+3) Install torch and torchvision. We used torch==1.12.1+cu116 and torchvision==0.13.1+cu116.
+
+    pip install torch==1.12.1+cu116 torchvision==0.13.1+cu116 --extra-index-url https://download.pytorch.org/whl/cu116
 
 4) Install other minimal requirements.
-   ````
+
    python -m pip install -r requirements.txt
-   ````
 
 5) Run the python script.
-    ````
+
    python two_head_model_demo.py --target_class 9 --protected_class 20
-   

examples/training_a_two_head_model/two_head_model_demo.py

@@ -205,4 +205,3 @@ def validation_step(self, batch, batch_idx):
     loaded_model = LitTwoHead.load_from_checkpoint(checkpoint_path=checkpoint_callback.best_model_path, model=model)
     loaded_model.eval()
     trainer.test(model=loaded_model, dataloaders=test_loader)
-

pyproject.toml

@@ -0,0 +1,11 @@
+[build-system]
+requires = ["setuptools>=68.2.2"]
+build-backend = "setuptools.build_meta"
+
+[project]
+name = "oxonfair"
+dynamic = ["optional-dependencies",'license','classifiers','version','authors','dependencies','readme','description','requires-python'] 
+
+[project.urls]
+Homepage =  "https://github.com/ChrisMRuss/oxon-fair/"
+Issues = "https://github.com/ChrisMRuss/oxon-fair/issues"

setup.py

@@ -57,8 +57,8 @@ def default_setup_args(*, version):
         license="Apache-2.0",
         license_files=("LICENSE", "NOTICE"),
         # Package info
-        packages=find_packages("src"),
         package_dir={"": "src"},
+        packages=find_packages("src"),
         namespace_packages=[],
         zip_safe=True,
         include_package_data=True,