VIVID-10M

[project page] | [Paper] | [arXiv]

VIVID-10M is the first large-scale hybrid image-video local editing dataset aimed at reducing data construction and model training costs, comprising 9.7M samples that encompass a wide range of video editing tasks.

Data Index

The data index is located at four .csv files:

vivid-image-change.csv
vivid-image-remove.csv
vivid-video-change.csv
vivid-video-remove.csv

VIVID-Video splits contains the columns:

local_caption, # caption of masked object
source_video_path, # ground-truth video path
crop_video_path, # cropped video path (need to synthesize)
mask_path, # masked video path
editing_mode # change or remove

VIVID-Image splits contains the columns:

local_caption, # caption of masked object
source_image_path, # ground-truth image path
crop_image_path, # cropped image path (need to synthesize)
mask_path, # masked image path
editing_mode # change or remove

Get started

1. Download all files from this repository.
2. Merge split files.

cat vivid-video.tar.part-* > vivid-video.tar
cat vivid-image.tar.part-* > vivid-image.tar

3. Expand the .tar file.

tar -xvf vivid-video.tar
tar -xvf vivid-image.tar

4. (Optional) Synthesize cropped data.

python get_crop_data.py