StephenHausler

Hi @gmberton, We hadn't planned to release it, but it is something we can look into. It is quite a large dataset, I'll investigate over the next couple weeks and...

Hi @gmberton and @MAX-OTW, sorry this has taken a while to complete. We have now created a public copy of the dataset, which is available at: https://cloudstor.aarnet.edu.au/plus/s/2LtwUtLUFpUiUC8 (this contains all...

Hi @cazhang, FYI msls.py was written by Mapillary, with a line or two changed by us. See: https://github.com/mapillary/mapillary_sls 1. Line 225 takes a query image id and returns a unique...

Closing now, feel free to reopen or email me if you have any further questions.

Hi @wzxuhuaiyuan, thanks for your interest in our work! We will have the training code released by the end of this month at the very latest.

Hi, I can try and offer some tips that might help. For a custom dataset, a lot of the training code will stay the same as in msls.py. You can...

FYI, since cloudstor has been decommissioned the Nordland dataset can now be downloaded from: https://huggingface.co/datasets/Somayeh-h/Nordland

I've taken a look, first I'm not sure our code (at least in the matching stage) will actually use all four GPUs. Can you please try using just one GPU...