[CVPR 2021] Pytorch implementation of Hijack-GAN: Unintended-Use of Pretrained, Black-Box GANs

Last update: Sep 05, 2022

Overview

Hijack-GAN: Unintended-Use of Pretrained, Black-Box GANs

In this work, we propose a framework HijackGAN, which enables non-linear latent space traversal and gain high-level controls, e.g., attributes, head poses, and landmarks, over unconditional image generation GANs in a fully black-box setting. It opens up the possibility of reusing GANs while raising concerns about unintended usage.

[Paper (CVPR 2021)][Project Page]

Prerequisites

Install required packages

pip install -r requirements.txt

Download pretrained GANs

Download the CelebAHQ pretrained weights of ProgressiveGAN [paper][code] and StyleGAN [paper][code], and then put those weights in ./models/pretrain. For example,

pretrain/
├── Pretrained_Models_Should_Be_Placed_Here
├── karras2018iclr-celebahq-1024x1024.pkl
├── karras2019stylegan-celebahq-1024x1024.pkl
├── pggan_celebahq_z.pt
├── stylegan_celebahq_z.pt
├── stylegan_headpose_z_dp.pt
└── stylegan_landmark_z.pt

Quick Start

Specify number of images to edit, a model to generate images, some parameters for editting.

LATENT_CODE_NUM=1
python edit.py \
    -m pggan_celebahq \
    -b boundaries/ \
    -n "$LATENT_CODE_NUM" \
    -o results/stylegan_celebahq_eyeglasses \
    --step_size 0.2 \
    --steps 40 \
    --attr_index 0 \
    --task attribute \
    --method ours

Usage

Important: For different given images (initial points), different step size and steps may be considered. In the following examples, we provide the parameters used in our paper. One could adjust them for better performance.

Specify Number of Samples

LATENT_CODE_NUM=1

Unconditional Modification

python edit.py \
    -m pggan_celebahq \
    -b boundaries/ \
    -n "$LATENT_CODE_NUM" \
    -o results/stylegan_celebahq_smile_editing \
    --step_size 0.2 \
    --steps 40 \
    --attr_index 0\
    --task attribute

Conditional Modification

python edit.py \
    -m pggan_celebahq \
    -b boundaries/ \
    -n "$LATENT_CODE_NUM" \
    -o results/stylegan_celebahq_smile_editing \
    --step_size 0.2 \
    --steps 40 \
    --attr_index 0\
    --condition\
    -i codes/pggan_cond/age.npy
    --task attribute

Head pose

Pitch

python edit.py \
    -m stylegan_celebahq \
    -b boundaries/ \
    -n "$LATENT_CODE_NUM" \
    -o results/ \
    --task head_pose \
    --method ours \
    --step_size 0.01 \
    --steps 2000 \
    --attr_index 1\
    --condition\
    --direction -1 \
    --demo

Yaw

python edit.py \
    -m stylegan_celebahq \
    -b boundaries/ \
    -n "$LATENT_CODE_NUM" \
    -o results/ \
    --task head_pose \
    --method ours \
    --step_size 0.1 \
    --steps 200 \
    --attr_index 0\
    --condition\
    --direction 1\
    --demo

Landmarks

Parameters for reference: (attr_index, step_size, steps) (4: 0.005 400) (5: 0.01 100), (6: 0.1 200), (8 0.1 200)

CUDA_VISIBLE_DEVICES=0 python edit.py \
    -m stylegan_celebahq \
    -b boundaries/ \
    -n "$LATENT_CODE_NUM" \
    -o results/ \
    --task landmark \
    --method ours \
    --step_size 0.1 \
    --steps 200 \
    --attr_index 6\
    --condition\
    --direction 1 \
    --demo

Generate Balanced Data

This a templeate showing how we generated balanced data for attribute manipulation (16 attributes in our internal experiments). You can modify it to fit your task better. Please first refer to here and replace YOUR_TASK_MODEL with your own classification model, and then run:

NUM=500000
CUDA_VISIBLE_DEVICES=0 python generate_balanced_data.py -m stylegan_celebahq \
    -o ./generated_data -K ./generated_data/indices.pkl -n "$NUM" -SI 0 --no_generated_imgs

[CVPR 2021] Pytorch implementation of Hijack-GAN: Unintended-Use of Pretrained, Black-Box GANs

Related tags

Overview

Hijack-GAN: Unintended-Use of Pretrained, Black-Box GANs

Prerequisites

Install required packages

Download pretrained GANs

Quick Start

Usage

Specify Number of Samples

Unconditional Modification

Conditional Modification

Head pose

Pitch

Yaw

Landmarks

Generate Balanced Data

Evaluations

TO-DO

Acknowledgment

Owner

Hui-Po Wang

Implementation of "Scaled-YOLOv4: Scaling Cross Stage Partial Network" using PyTorch framwork.

PartImageNet is a large, high-quality dataset with part segmentation annotations

My solution for the 7th place / 245 in the Umoja Hack 2022 challenge

A method to perform unsupervised cross-region adaptation of crop classifiers trained with satellite image time series.

This is code to fit per-pixel environment map with spherical Gaussian lobes, using LBFGS optimization

Qt-GUI implementation of the YOLOv5 algorithm (ver.6 and ver.5)

A generalized framework for prototyping full-stack cooperative driving automation applications under CARLA+SUMO.

Real-time VIBE: Frame by Frame Inference of VIBE (Video Inference for Human Body Pose and Shape Estimation)

[ICLR 2022] Pretraining Text Encoders with Adversarial Mixture of Training Signal Generators

BOVText: A Large-Scale, Multidimensional Multilingual Dataset for Video Text Spotting

PyTorch Implementation of the SuRP algorithm by the authors of the AISTATS 2022 paper "An Information-Theoretic Justification for Model Pruning"

A project to make Amazon Echo respond to sign language using your webcam

Calling Julia from Python - an experiment on data loading

Official pytorch implementation of paper "Image-to-image Translation via Hierarchical Style Disentanglement".

Code for the Weighted, Accelerated and Restarted Primal-dual algorithm. This algorithm achieves stable linear convergence for reconstruction from undersampled noisy measurements under an approximate sharpness condition. See the paper for details.

Code for IntraQ, PyTorch implementation of our paper under review

Neural Style and MSG-Net

Keras Image Embeddings using Contrastive Loss

DynamicViT: Efficient Vision Transformers with Dynamic Token Sparsification