a) Fair Generative Model Via Transfer Learning

Model Base codes were adopted from Choi et.al https://github.com/ermongroup/fairgen Preprocessing of UTKFace and some data pre-processing steps were adopted from Um et.al https://openreview.net/pdf?id=F1Z3QH-VjZE Use Files in "./BIGGAN"

#Training FairGAN for Validation (Read me is abstracted from Choi etal source code and some modifications have been made for our purpose)

1) Data setup:

(a) Download the CelebA dataset here (http://mmlab.ie.cuhk.edu.hk/projects/CelebA.html) into the data/ directory (if elsewhere, note the path for step b). Of the download links provided, choose Align&Cropped Images and download Img/img_align_celeba/ folder, Anno/list_attr_celeba.txt, and Eval/list_eval_partition.txt to data/. (b) Download the UTKFace dataset here (https://susanqq.github.io/UTKFace/) into the data/ directory (c) Preprocess the CelebA dataset for faster training:

python preprocess_celeba.py --data_dir=/path/to/downloaded/dataset/celeba/ --out_dir=../data --partition=train

You should run this script for --partition=[train, val, test] to cache all the necessary data. The preprocessed files will then be saved in data/. (d) Preprocess the UTKFace dataset for fast training

python3 preprocess_UTKFace.py --data_dir=/path/to/downloaded/dataset/UTKFace/ --out_dir=../data/UTKFace

e)FID: For CelebA, we have provided unbiased FID statistics in the source directory. For Multi CelebA run ./BIGGAN/notebook/multi-attribute data and unbiased FID splits.ipynb . For UTKFace run ./BIGGAN/notebook/UTKFace unbiased FID splits.ipynb f)

#Preprocessing files for the given LA & Training classifier

2) For training of Resnet Classifier

a) Train attribute classifier with CelebA (Single Attribute)

python train_attribute_clf.py celeba ./results/celeba/attr_clf

a) Train attribute classifier with CelebA (Multi Attribute)

python train_attribute_clf.py celeba ./results/celeba/multi_clf -- multi=True

c) Train attribute classifier with UTKFace

python train_attribute_clf.py UTKFace ./results/UTKFace/attr_clf

#Prepare data-split

3) Generate the various Perc and bias spilits

The density ratio classifier should be trained for the appropriate bias and perc setting, which can be adjusted in the script below:

python get_density_ratios.py [celeba UTKFace] [ResNet18 CNN5 CNN3] --perc=[0.1, 0.25, 0.5, 1.0] --bias=[90_10, multi]

By employing --ncf argument, you can control the complexity of CNN classifiers. Note that the best density ratio classifier will be saved in its corresponding directory under ./data/[celeba UTKFace FairFace], you should transfer this to ./data.

#2) To Train the Baseline Model

Use ./BIGGAN/src/Base_imp_weight for baseline and ./BIGGAN/src/Tfl_2D_LP_FT for our proposed work

Use --reweight 1 for Choi et.al Importance-weights Select the --perc setting e.g. {0.25,0.1,0.05,0.025} Select the --bias settings e.g. {90_10,multi}, if you are testing multi you have to append --multi 1 for the model to reference the correct FD classifier

python train.py --shuffle --batch_size 128 --parallel --num_G_accumulations 1 --num_D_accumulations 1 --num_D_steps 4 --G_lr 2e-4 --D_lr 2e-4 --dataset CA64 --data_root ../../data --G_ortho 0.0 --G_attn 0 --D_attn 0 --G_init N02 --D_init N02 --ema --use_ema --ema_start 1000 --save_every 1000 --test_every 1000 --num_best_copies 2 --num_save_copies 1 --loss_type hinge --seed 777 --num_epochs 200 --start_eval 40 --reweight 1 --alpha 1.0 --bias 90_10 --perc 1.0 --name_suffix celeba_90_10_perc1.0_Baseline

#3) To Train the Pre-trained Model

Use ./BIGGAN/src/FairGAN++ for baseline and ./BIGGAN/src/Tfl_2D_LP_FT for our proposed work

Use --reweight 0 for regular traing on the reference + bias data Select the perc setting e.g. {0.25,0.1,0.05,0.025} Select the --bias settings e.g. {90_10,multi}, if you are testing multi you have to append --multi 1 for the model to reference the correct FD classifier

python train.py --shuffle --batch_size 128 --parallel --num_G_accumulations 1 --num_D_accumulations 1 --num_D_steps 2 --G_lr 5e-4 --D_lr 2e-4 --dataset CA64 --data_root ../../data --G_ortho 0.0 --G_attn 0 --D_attn 0 --G_init N02 --D_init N02 --ema --use_ema --ema_start 1000 --save_every 1000 --test_every 1000 --num_best_copies 2 --num_save_copies 1 --loss_type hinge --seed 777 --num_epochs 200 --start_eval 40 --reweight 0 --alpha 1.0 --bias 90_10 --perc 1.0 --name_suffix celeba_90_10_perc1.0_pretrained

#4) FairTL

To perform fairTL we select the new "./weights/###_copy0" pre-trained weights and copy them into a new file of your own naming choice e.g., celeba_90_10_perc1.0_pretrained_Linear_Prob Include --dummy 2 to train on the uniform D_ref dataset only

python train.py --shuffle --batch_size 8 --parallel --num_G_accumulations 1 --num_D_accumulations 1 --num_D_steps 2 --G_lr 5e-4 --D_lr 2e-4 --dataset CA64 --data_root ../../data --G_ortho 0.0 --G_attn 0 --D_attn 0 --G_init N02 --D_init N02 --ema --use_ema --ema_start 1000 --save_every 1000 --test_every 1000 --num_best_copies 2 --num_save_copies 1 --loss_type hinge --seed 777 --num_epochs 300 --start_eval 40 --reweight 0 --alpha 1.0 --bias 60_40 --perc 1.0 --name_suffix celeba_90_10_perc1.0_pretrained --resumePaused --load_weight copy0 --dummy 2

#5) (FairTL++) Perform linear probing

To perform linear probing we select the new "./weights/###_copy0" pre-trained weights and copy them into a new file of your own naming choice e.g., celeba_90_10_perc1.0_pretrained_Linear_Prob Include --dummy 2 to train on the uniform D_ref dataset only

python train_LP.py --shuffle --batch_size 8 --parallel --num_G_accumulations 1 --num_D_accumulations 1 --num_D_steps 2 --G_lr 5e-4 --D_lr 2e-4 --dataset CA64 --data_root ../../data --G_ortho 0.0 --G_attn 0 --D_attn 0 --G_init N02 --D_init N02 --ema --use_ema --ema_start 1000 --save_every 1000 --test_every 1000 --num_best_copies 2 --num_save_copies 1 --loss_type hinge --seed 777 --num_epochs 200 --start_eval 40 --reweight 0 --alpha 1.0 --bias 90_10 --perc 1.0 --name_suffix celeba_90_10_perc1.0_pretrained --resumePaused --load_weights=copy0 --dummy 2

#6) (FairTL++) Perform Fine Tuning

Similarly select the new "./weights/###_copy0" weights and copy them into a new file e.g., celeba_90_10_perc1.0_pretrained_Linear_Prob_Fine_tuning Include --dummy 2 to train on the uniform D_ref dataset only

python train_FT.py --shuffle --batch_size 8 --parallel --num_G_accumulations 1 --num_D_accumulations 1 --num_D_steps 2 --G_lr 5e-4 --D_lr 2e-4 --dataset CA64 --data_root ../../data --G_ortho 0.0 --G_attn 0 --D_attn 0 --G_init N02 --D_init N02 --ema --use_ema --ema_start 1000 --save_every 1000 --test_every 1000 --num_best_copies 2 --num_save_copies 1 --loss_type hinge --seed 777 --num_epochs 200 --start_eval 40 --reweight 0 --alpha 1.0 --bias 90_10 --perc 1.0 --name_suffix celeba_90_10_perc1.0_pretrained --resumePaused --load_weights=copy0 --dummy 2

b) Fair Generative Model Via Transfer Learning From StyleGANv2

#1) Preprocessing Raw data into the respective SA splits

Base codes were adopted from karras et.al https://github.com/NVlabs/stylegan2-ada-pytorch and labels from https://github.com/DCGM/ffhq-features-dataset Download the FFHQ dataset from https://github.com/NVlabs/ffhq-dataset and place it in ./StyleGANv2/Prepared_data/data Run the following pre-processing script in ./StyleGANv2/Prepared_data. This will output 3 types of files 1) File for training the GAN e.g., ./training/Gender/GenderTrainSamples_0.025/ 2)FID reference files e.g., ./training/Gender/GenderFIDSamples/ 3)Gender_FFHQ_Clf_Training_data.pt and Gender_FFHQ_Clf_Training_labels.pt

python sort_data.py

#2) Preprocessing the data into the zip file StyleGAN2 code requires

zip the file for training Select the source data i.e., ./StyleGANv2/stylegan2-ada-pytorch-main/Prepare_data/training/Gender/GenderTrainSamples_0.025/ Select the output dataset name (this can be anything you want)

python dataset_tool.py --source=*source of Dref* --dest=~/datasets/*dataset_name*.zip

#2) Train attribute classifier

Use the directiory ./StyleGANv2/stylegan2-ada-pytorch-main//metrics/FD Then copy the data from 3)Gender_FFHQ_Clf_Training_data.pt and Gender_FFHQ_Clf_Training_labels.pt into ./StyleGANv2/stylegan2-ada-pytorch-main//metrics/FD/data/ and post process the data to train/Test/split with:

python Prepare_training_data.py

then train the classifier with (Fill sensitive attribute(SA) name):

python train_attribute_clf.py -attribute *SA*

#3)Resume sensitive adaptation with styleGANv2

Training the pre-trained styleGANv2 on the new Dref data with linear probing Fill training datafile with output file from 2) Fill network pickle file

python train_2D_LP.py --outdir=~/training-runs --data=~/mydataset/*training datafile* --gpus=1 --network=https://api.ngc.nvidia.com/v2/models/nvidia/research/stylegan2/versions/1/files/stylegan2-ffhq-256x256.pkl --aug=noaug --kimg 200

followed by fine-tuining:

network pickle file is the output of the LP file

python train_2D.py --outdir=~/training-runs --data=~/mydataset/*training datafile* --gpus=1 --network=*network pickle file* --aug=noaug --kimg 800

#4)Measuring Fairness and FID

Select FID ref file from 2)FID reference files e.g., .`/training/Gender/GenderFIDSamples/ Select saved network from the training file Select SA to evaluate

python cal_FD --attribute *SA* --network *saved network*

python calc_metrics.py --metrics=fid50k_full --data=~/datasets/*FID ref file* --mirror=1 --network *saved network*

c) Fair Generative Model Via Transfer Learning From ProGAN

#1) Prepare training data

Use the directory ./ProGAN/prepareData Ammend food in selectedKey=keys[foo] with SA index, see keys=list(labelsDf.keys()) for the respective index select mode="perc" for selection of |Dref|/|Dbias| and mode="min" for even dataset for training classifier Change bar to the respective percentage in perc="bar" Ensure that dataPath points to the folder containing .jpg of CelebA-HQ

python sortData

#2) train proGAN

Use the directory ./ProGAN/pytorch_GAN_zoo-main Use either train_fairTL.py (for transfer learning) or train_fairTL++.py for (fairTL++) Amend the SA parameters for different Sensitive attriviute e.g., {Black_Hair,Male}

python train_fairTL.py

python train_fairTL++.py

#3) Measure Fairness discrepency (FD)

Train attribute classifier first Amend foo to the respective classname (same as the data preperation in #1)

python train_attribute_clf.py --class_name=foo

Generate data (use directory ./ProGAN/pytorch_GAN_zoo-main) Change pathPrefix to point at location of save state_dict Change stateDictNum to select save state dict index

python generate_samples.py

Measure FD Change stateDictPath to point at the classifier's weights Change dataPath to point at generated data

python FD_measure.py

#4) Measuring FID

Use the directory ./ProGAN/prepareData Prepare data first, select foo as per step 1 in selectedKey=keys[foo]

python Pre_process_FID_2.py

Use directory ./ProGAN/FID Ammend refPath to point at reference data above Ammend samplePath to point at the generated data

python measure_FID.py