Hello everyone, we did a brief update on the data section of the challenge and the dataset.
The official dataset can be downloaded here: https://huggingface.co/datasets/marcosv/rgb2raw/
(click on "files" in the web https://huggingface.co/datasets/marcosv/rgb2raw/tree/main)
- train_rgb2raw.zip 8.73 GB
- val_rgbs_in.zip 208 MB
Hello,
I got a question on updated data.
Before this update, each scene has one corresponding pkl file for meta data.
However, on updated dataset, each scene has two pkl file with meta data.
For example, I got pairs of (1.pkl / 01.pkl) / (2.pkl / 02.pkl) ....
And their value is different. (between pair)
Could you clarify the difference of those pkl file?
Thanks
V/R
Hello, sorry for the notation 1.pkl and 01.pkl are different because are indeed two different scenes. Each scene has a unique id. -- the images are also different.
Sorry for the confusion, just process the images differently assuming they are different scenes
(this is miss-notation comes from the way we collected data, sorry for that)
Hello, sorry for the notation 1.pkl and 01.pkl are different because are indeed two different scenes. Each scene has a unique id. -- the images are also different.
Sorry for the confusion, just process the images differently assuming they are different scenes
(this is miss-notation comes from the way we collected data, sorry for that)