https://github.com/ruchikachavhan/concept-prune

Code for the paper - ConceptPrune: Concept Editing in Diffusion Models via Skilled Neuron Pruning
https://github.com/ruchikachavhan/concept-prune

Last synced: 13 days ago
JSON representation

Code for the paper - ConceptPrune: Concept Editing in Diffusion Models via Skilled Neuron Pruning

• Host: GitHub
• URL: https://github.com/ruchikachavhan/concept-prune
• Owner: ruchikachavhan
• Created: 2024-05-29T13:02:57.000Z (6 months ago)
• Default Branch: main
• Last Pushed: 2024-08-13T15:10:34.000Z (3 months ago)
• Last Synced: 2024-08-13T18:14:33.337Z (3 months ago)
• Language: Jupyter Notebook
• Size: 1.41 MB
• Stars: 9
• Watchers: 3
• Forks: 0
• Open Issues: 0
• Metadata Files:
  • Readme: README.md

Awesome Lists containing this project

• awesome-diffusion-categorized - [Code

README

        
# ConceptPrune

Code for the paper - ConceptPrune: Concept Editing in Diffusion Models via Skilled Neuron Pruning ([arxiv preprint](https://arxiv.org/abs/2405.19237v1))

## Introduction

While large-scale text-to-image diffusion models have demonstrated impressive image-generation capabilities, there are significant concerns about their potential misuse for generating unsafe content, violating copyright, and perpetuating societal biases. Recently, the text-to-image generation community has begun addressing these concerns by editing or unlearning undesired concepts from pre-trained models. However, these methods often involve data-intensive and inefficient fine-tuning or utilize various forms of token remapping, rendering them susceptible to adversarial jailbreaks. In this paper, we present a simple and effective training-free approach, ConceptPrune, wherein we first identify critical regions within pre-trained models responsible for generating undesirable concepts, thereby facilitating straightforward concept unlearning via weight pruning. Experiments across a range of concepts including artistic styles, nudity, object erasure, and gender debiasing demonstrate that target concepts can be efficiently erased by pruning a tiny fraction, approximately 0.12% of total weights, enabling multi-concept erasure and robustness against various white-box and black-box adversarial attacks.

## Experiments

### Environment Setup

Create the environment from the ```environment.yml``` file.

```conda env create -f environment.yml```

```conda activate concept-prune```

We recommend using ```diffusers v0.29.2``` as the results may change for different versions.

### Code Structure

The file structure is as follows

  ```configs``` - Contains ```.yaml`` file for basic arguments. These arguments can be changed within the scripts using argument parsers.

  ```datasets``` - Contains txt or csv file with prompts for different concepts

  ```neuron_receivers``` - Contains classes to hook Feed Forward network (FFN) modules within the Unet to record neuron activations

  ```wanda``` - Contains scripts to calculate WANDA pruing metric introduced in [Sun et. al](https://arxiv.org/abs/2306.11695) for FFN weights

  ```utils``` - Basic helper functions

  ```benchmarking``` - Scripts to run all the benchmarks in the paper for different concepts

### Pruning the model using WANDA

To obtain a pruned model for a concept ``````, run the following - 

1. Discover skilled neurons for a concept

    ```

    python wanda/wanda.py --target  --skill_ratio 0.01

    ```

    `````` is the concept that we want to erase. Replace  ``` --skill_ratio 

    ```

3. Next, we take a union over skilled neurons for the first few timesteps.

    Run the follwing command to obtain the pruned model.

    ```

    python wanda/save_union_over_time.py --target  --timesteps  --skill_ratio 

    ```

    We provide the values of these hyper-parameters in Table 7 in the Appendix for every concept.

### Benchmarks

#### Baselines

Train concept erasure baselines - [UCE](https://github.com/rohitgandikota/unified-concept-editing), [FMN](https://github.com/SHI-Labs/Forget-Me-Not), [ESD](https://github.com/rohitgandikota/erasing), [Concept-Ablation](https://github.com/nupurkmr9/concept-ablation) using their respective repositories. n our code base, we provide code to evalaute these baselines on concept-erasure benchmarks for different concepts. In the following experiments, ausedd ```uce, esd, fmn, concept-ablation``` for `````` respectively to run the above baselines. 

### Download Checkpoints

We will provide checkpoints on Hugging Face soon!

#### Evaluate ConceptPrune

1. Artist Styles

    To evaluate artist style erasure for ```Van Gogh, Monet, Pablo Picasso, Da Vinci, Salvador Dali`` for ConceptPrune, run 

    ```

    python benchmarking/artist_erasure.py --target  --baseline concept-prune --ckpt_name 

    ```

    We created a dataset of 50 prompts using ChatGPT for different artists such that each prompt contains the painting name along with the name of the artist. These propmts are available in ```datasets/```. The script saves images  and a json files with CLIP metric reported in the paper in the ```results/``` folder.

2. Nudity 

    To evaluate nudity erasure with ConceptPrune on the I2P dataset, run

    ```

    python benchmarking/nudity_eval.py --eval_dataset i2p --baseline 'concept-prune' --gpu 0 --ckpt_name 

    ```

    To run ConceptPrune on black-box adversarial prompt datasets, [MMA](https://openaccess.thecvf.com/content/CVPR2024/papers/Yang_MMA-Diffusion_MultiModal_Attack_on_Diffusion_Models_CVPR_2024_paper.pdf) and [Ring-A-Bell](https://arxiv.org/abs/2310.10012), replace ```i2p``` with ```mma``` and ```ring-a-bell``` respectively.

    We evaluate nudity in images using the [NudeNet detector](https://pypi.org/project/nudenet/). The script saves images and a json files with NudeNet scores reported in the paper in the ```results/``` folder.

3. Object Erasing

    To evaluate object erasure with ConceptPrune, run

    ```

    python benchmarking/object_erase.py --target  --baseline concept-prune --removal_mode erase --ckpt_name 

    ```

    

    To check interference of concept removal with unrelated classes, run

    ```

    python benchmarking/object_erase.py --target  --baseline concept-prune --removal_mode keep --ckpt_name 

    ```

    where `````` is the name of a class in ImageNette classes.  he script saves images and a json files with ResNet50 accuracies reported in the paper in the ```results/``` folder.

4. Gender reversal

    To evaluate gender reversal from Female to Male, run

    ```

    python benchmarking/gender_reversal.py --target male --ckpt_name 

    ```

    Replace ```male``` with ```female``` to reverse gender from Male to Female. We calculate the success of gender reversal using CLIP to classify between males females. The script saves images in the ```results/``` folder for 250 seeds.

5. COCO evaluation

    To evaluate ConceptPrune on COCO dataset, run

    ```

    python benchmarking/eval_coco.py --target  --baseline concept-prune --ckpt_name 

    ``` 

5. Memorization

    To evaluate ConceptPrune on COCO dataset, run

    ```

    python benchmarking/inference_mem.py --target memorize_$i$ --baseline concept-prune --ckpt_name 

    ``` 

    This will save images and calculate SSCD and CLIP score and store the results in a json file. We run this script for 10 different seeds for every model and report average performance.

### Cite us!

If you find our paper useful, please consider citing our work. 

```

@article{chavhan2024conceptpruneconcepteditingdiffusion,

      title={ConceptPrune: Concept Editing in Diffusion Models via Skilled Neuron Pruning}, 

      author={Ruchika Chavhan and Da Li and Timothy Hospedales},

      year={2024},

      journal={ArXiv}

}

```

```

@article{chavhan2024conceptpruneconcepteditingdiffusion,

      title={Memorized Images in Diffusion Models share a Subspace that can be Located and Deleted}, 

      author={Ruchika Chavhan and  Ondrej Bohdal and Yongshuo Zong and Da Li and Timothy Hospedales},

      year={2024},

      journal={ArXiv}

}

```

### Contact

Please contact [email protected] for any questions!