|
--- |
|
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. |
|
|
|
An example for the annotations in UMG-41M: |
|
|
|
<div align="center"> |
|
<img src="example.pdf" alt="Example" align=center /> |
|
</div> |
|
|
|
UMG-41M can be used in the following ways: |
|
|
|
1. Unzip the zip packages of the individual subdatasets, each of which contains multiple subfolders. Each subfolder contains the JSON annotations corresponding to the images. |
|
2. 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. |
|
3. Use the WebDataset library to access this data in the code. |
|
|
|
Our code for UMG-CLIP will be available at https://github.com/lygsbw/UMG-CLIP. |
