metadata
license: cc-by-4.0
UMG-41M is a component of our ECCV 2024 paper, UMG-CLIP. It includes tag and caption annotations at both the image level and pixel level, as well as segmentation masks for both the foreground and background of images from six public datasets: CC3M, CC12M, SBU, VG, YFCC15M, and IN21K.
UMG-41M can be used in the following ways:
- Unzip the zip packages of the individual subdatasets, each of which contains multiple subfolders. Each subfolder contains the JSON annotations corresponding to the images.
- Download the data for the individual subdatasets from their respective websites, and place the corresponding PNG images in the same folders as the JSON annotations.
- Use the WebDataset library to access this data in the code. Our code for this is included in the repository at https://github.com/lygsbw/UMG-CLIP.