Datasets:

Modalities:
Image
Text
Size:
< 1K
ArXiv:
Libraries:
Datasets
License:
dreambooth / README.md
Linoy Tsaban
Update README.md
dc09c98
metadata
configs:
  - config_name: default
    data_files:
      - split: train
        path: dataset/backpack/*.jpg
  - config_name: backpack
    data_files:
      - split: train
        path: dataset/backpack/*.jpg
  - config_name: backpack_dog
    data_files:
      - split: train
        path: dataset/backpack_dog/*.jpg
  - config_name: bear_plushie
    data_files:
      - split: train
        path: dataset/bear_plushie/*.jpg
  - config_name: berry_bowl
    data_files:
      - split: train
        path: dataset/berry_bowl/*.jpg
  - config_name: can
    data_files:
      - split: train
        path: dataset/can/*.jpg
  - config_name: candle
    data_files:
      - split: train
        path: dataset/candle/*.jpg
  - config_name: cat
    data_files:
      - split: train
        path: dataset/cat/*.jpg
  - config_name: cat2
    data_files:
      - split: train
        path: dataset/cat2/*.jpg
  - config_name: clock
    data_files:
      - split: train
        path: dataset/clock/*.jpg
  - config_name: colorful_sneaker
    data_files:
      - split: train
        path: dataset/colorful_sneaker/*.jpg
  - config_name: dog
    data_files:
      - split: train
        path: dataset/dog/*.jpg
  - config_name: dog2
    data_files:
      - split: train
        path: dataset/dog2/*.jpg
  - config_name: dog3
    data_files:
      - split: train
        path: dataset/dog3/*.jpg
  - config_name: dog5
    data_files:
      - split: train
        path: dataset/dog5/*.jpg
  - config_name: dog6
    data_files:
      - split: train
        path: dataset/dog6/*.jpg
  - config_name: dog7
    data_files:
      - split: train
        path: dataset/dog7/*.jpg
  - config_name: dog8
    data_files:
      - split: train
        path: dataset/dog8/*.jpg
  - config_name: duck_toy
    data_files:
      - split: train
        path: dataset/duck_toy/*.jpg
  - config_name: fancy_boot
    data_files:
      - split: train
        path: dataset/fancy_boot/*.jpg
  - config_name: grey_sloth_plushie
    data_files:
      - split: train
        path: dataset/grey_sloth_plushie/*.jpg
  - config_name: monster_toy
    data_files:
      - split: train
        path: dataset/monster_toy/*.jpg
  - config_name: pink_sunglasses
    data_files:
      - split: train
        path: dataset/pink_sunglasses/*.jpg
  - config_name: poop_emoji
    data_files:
      - split: train
        path: dataset/poop_emoji/*.jpg
  - config_name: rc_car
    data_files:
      - split: train
        path: dataset/rc_car/*.jpg
  - config_name: red_cartoon
    data_files:
      - split: train
        path: dataset/red_cartoon/*.jpg
  - config_name: robot_toy
    data_files:
      - split: train
        path: dataset/robot_toy/*.jpg
  - config_name: shiny_sneaker
    data_files:
      - split: train
        path: dataset/shiny_sneaker/*.jpg
  - config_name: teapot
    data_files:
      - split: train
        path: dataset/teapot/*.jpg
  - config_name: vase
    data_files:
      - split: train
        path: dataset/vase/*.jpg
  - config_name: wolf_plushie
    data_files:
      - split: train
        path: dataset/wolf_plushie/*.jpg
license: cc-by-4.0

Dataset Card for "dreambooth"

Dataset of the Google paper DreamBooth: Fine Tuning Text-to-Image Diffusion Models for Subject-Driven Generation

The dataset includes 30 subjects of 15 different classes. 9 out of these subjects are live subjects (dogs and cats) and 21 are objects. The dataset contains a variable number of images per subject (4-6). Images of the subjects are usually captured in different conditions, environments and under different angles.

We include a file dataset/prompts_and_classes.txt which contains all of the prompts used in the paper for live subjects and objects, as well as the class name used for the subjects.

The images have either been captured by the paper authors, or sourced from www.unsplash.com

The dataset/references_and_licenses.txt file contains a list of all the reference links to the images in www.unsplash.com - and attribution to the photographer, along with the license of the image.

project page | arxiv

Academic Citation

If you use this work please cite:

@inproceedings{ruiz2023dreambooth,
  title={Dreambooth: Fine tuning text-to-image diffusion models for subject-driven generation},
  author={Ruiz, Nataniel and Li, Yuanzhen and Jampani, Varun and Pritch, Yael and Rubinstein, Michael and Aberman, Kfir},
  booktitle={Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition},
  year={2023}
}

Disclaimer

This is not an officially supported Google product.