https://www.kaggle.com/datasets/byildiz/amstertime?resource=download
For AmsterTime, new as ref and old as query.
After filling this form, you will be shown following command:
rsync rsync://vandaldata.polito.it/sf_xl/README.md .
You can download the README of the dataset with this bash command, and you will find instructions to download the dataset inside the README.
Following is that README file.
In this README, you have some rsync
commands. You will have to download processed/val/queries
,
## How to download
To download a directory (or file) you can run this bash command (it requires rsync, does not support scp)
`rsync -rhz --info=progress2 --ignore-existing rsync://vandaldata.polito.it/sf_xl/REL_PATH .`
for example to download the `processed/test/queries_night` folder you can run
`rsync -rhz --info=progress2 --ignore-existing rsync://vandaldata.polito.it/sf_xl/processed/test/queries_night .`
**what you need:**
rsync -rhz --info=progress2 --ignore-existing rsync://vandaldata.polito.it/sf_xl/processed/val/queries .
rsync -rhz --info=progress2 --ignore-existing rsync://vandaldata.polito.it/sf_xl/processed/val/database .
Find detailed instructions here: https://github.com/AnyLoc/Revisit-Anything/tree/main/dataloaders/msls_data_clean