daam - GitLite

KO | EN

gitlite — search

#javascript #hacktoberfest #go #android #machine-learning #deep-learning #linux #macos #nodejs #kubernetes #vue #awesome-list #security #ios

daam

(★ 803)

Diffusion attentive attribution maps for interpreting Stable Diffusion.

파일 탐색기

최종 버전 다운로드 (.zip)

- - __init__.py
  - demo.py
  - evaluate.py
  - generate.py
- __init__.py
- _version.py
- evaluate.py
- experiment.py
- heatmap.py
- hook.py
- trace.py
- utils.py
- vocab-small.tsv
- vocab.tsv
- - daam.rst.txt
  - daam.run.rst.txt
  - index.rst.txt
  - modules.rst.txt
- - alabaster.css
  - basic.css
  - custom.css
  - doctools.js
  - documentation_options.js
  - file.png
  - jquery-3.5.1.js
  - jquery.js
  - language_data.js
  - minus.png
  - plus.png
  - pygments.css
  - searchtools.js
  - underscore-1.13.1.js
  - underscore.js
- .buildinfo
- .nojekyll
- conf.py
- daam.html
- daam.rst
- daam.run.html
- daam.run.rst
- genindex.html
- index.html
- index.rst
- make.sh
- Makefile
- modules.html
- modules.rst
- objects.inv
- py-modindex.html
- search.html
- searchindex.js
- 0-setup.ipynb
- 1-visuosyntactic-analyses.ipynb
- 2-visuosemantic-analyses.ipynb
.gitignore
example.jpg
LICENSE
MANIFEST.in
README.md
requirements.txt
scrollbar.css
setup.py

🚀 설치 가이드

1. 코드 내려받기

git clone https://github.com/castorini/daam

깃허브에서 프로젝트 코드 전체를 내 컴퓨터로 내려받습니다.

cd daam

방금 내려받은 프로젝트 폴더 안으로 이동합니다.

2. Python

쉬움 추천

사전 준비물

Git GitHub에서 프로젝트 코드를 내려받으려면 필요합니다.
Python 3 설치 시 'Add Python to PATH' 옵션을 꼭 체크하세요 (Windows).

pip install -r requirements.txt

requirements.txt 등에 명시된 파이썬 라이브러리를 설치합니다.

jupyter notebook

브라우저에서 노트북(.ipynb) 파일들을 열람하고 실행할 수 있는 Jupyter 화면을 켭니다.

✅ 에러 메시지 없이 실행되고 터미널에 안내 문구가 출력되면 정상입니다.

// repository documentation

Was this content helpful?

★ 0 (0 ratings)

# What the DAAM: Interpreting Stable Diffusion Using Cross Attention

[![HF Spaces](https://img.shields.io/badge/HuggingFace%20Space-online-green.svg)](https://huggingface.co/spaces/tetrisd/Diffusion-Attentive-Attribution-Maps) [![Citation](https://img.shields.io/badge/Citation-ACL-orange.svg)](https://gist.github.com/daemon/639de6fea584d7df1a62f04a2ea0cdad) [![PyPi version](https://badgen.net/pypi/v/daam?color=blue)](https://pypi.org/project/daam) [![Downloads](https://static.pepy.tech/badge/daam)](https://pepy.tech/project/daam)

![example image](https://raw.githubusercontent.com/castorini/daam/main/example.jpg)

### Updated to support Stable Diffusion XL (SDXL) and Diffusers 0.21.1!

I regularly update this codebase. Please submit an issue if you have any questions.

In [our paper](https://aclanthology.org/2023.acl-long.310), we propose diffusion attentive attribution maps (DAAM), a cross attention-based approach for interpreting Stable Diffusion.
Check out our demo: https://huggingface.co/spaces/tetrisd/Diffusion-Attentive-Attribution-Maps.
See our [documentation](https://castorini.github.io/daam/), hosted by GitHub pages, and [our Colab notebook](https://colab.research.google.com/drive/1miGauqa07uHnDoe81NmbmtTtnupmlipv?usp=sharing), updated for v0.1.0.

## Getting Started
First, install [PyTorch](https://pytorch.org) for your platform.
Then, install DAAM with `pip install daam`, unless you want an editable version of the library, in which case do `git clone https://github.com/castorini/daam && pip install -e daam`.
Finally, login using `huggingface-cli login` to get many stable diffusion models -- you'll need to get a token at [HuggingFace.co](https://huggingface.co/).

### Running the Website Demo
Simply run `daam-demo` in a shell and navigate to http://localhost:8080.
The same demo as the one on HuggingFace Spaces will show up.

### Using DAAM as a CLI Utility
DAAM comes with a simple generation script for people who want to quickly try it out.
Try running
```bash
$ mkdir -p daam-test && cd daam-test
$ daam "A dog running across the field."
$ ls
a.heat_map.png    field.heat_map.png    generation.pt   output.png  seed.txt
dog.heat_map.png  running.heat_map.png  prompt.txt
```
Your current working directory will now contain the generated image as `output.png` and a DAAM map for every word, as well as some auxiliary data.
You can see more options for `daam` by running `daam -h`.
To use Stable Diffusion XL as the backend, run `daam --model xl-base-1.0 "Dog jumping"`.

### Using DAAM as a Library

Import and use DAAM as follows:

```python
from daam import trace, set_seed
from diffusers import DiffusionPipeline
from matplotlib import pyplot as plt
import torch

model_id = 'stabilityai/stable-diffusion-xl-base-1.0'
device = 'cuda'

pipe = DiffusionPipeline.from_pretrained(model_id, use_auth_token=True, torch_dtype=torch.float16, use_safetensors=True, variant='fp16')
pipe = pipe.to(device)

prompt = 'A dog runs across the field'
gen = set_seed(0)  # for reproducibility

with torch.no_grad():
    with trace(pipe) as tc:
        out = pipe(prompt, num_inference_steps=50, generator=gen)
        heat_map = tc.compute_global_heat_map()
        heat_map = heat_map.compute_word_heat_map('dog')
        heat_map.plot_overlay(out.images[0])
        plt.show()
```

You can also serialize and deserialize the DAAM maps pretty easily:

```python
from daam import GenerationExperiment, trace

with trace(pipe) as tc:
    pipe('A dog and a cat')
    exp = tc.to_experiment('experiment-dir')
    exp.save()  # experiment-dir now contains all the data and heat maps

exp = GenerationExperiment.load('experiment-dir')  # load the experiment
```

We'll continue adding docs.
In the meantime, check out the `GenerationExperiment`, `GlobalHeatMap`, and `DiffusionHeatMapHooker` classes, as well as the `daam/run/*.py` example scripts.
You can download the COCO-Gen dataset from the paper at http://ralphtang.com/coco-gen.tar.gz.
If clicking the link doesn't work on your browser, copy and paste it in a new tab, or use a CLI utility such as `wget`.

## See Also
- [DAAM-i2i](https://github.com/RishiDarkDevil/daam-i2i), an extension of DAAM to image-to-image attribution.

- [Furkan's video](https://www.youtube.com/watch?v=XiKyEKJrTLQ) on easily getting started with DAAM.

- [1littlecoder's video](https://www.youtube.com/watch?v=J2WtkA1Xfew) for a code demonstration and Colab notebook of an older version of DAAM.

## Citation
```
@inproceedings{tang2023daam,
    title = "What the {DAAM}: Interpreting Stable Diffusion Using Cross Attention",
    author = "Tang, Raphael  and
      Liu, Linqing  and
      Pandey, Akshat  and
      Jiang, Zhiying  and
      Yang, Gefei  and
      Kumar, Karun  and
      Stenetorp, Pontus  and
      Lin, Jimmy  and
      Ture, Ferhan",
    booktitle = "Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)",
    year = "2023",
    url = "https://aclanthology.org/2023.acl-long.310",
}
```