salad
salad copied to clipboard
serizba
Optimal Transport Aggregation for Visual Place Recognition
Hello, I want to reproduce the experiment in the paper, but cannot find the corresponding parameters. Can you provide the training code of DINOV2-Netvlad?
Hello, I would like to reproduce the experiment of Resnet+SALAD in the paper, but cannot find the corresponding setting parameters about training a Resnet+SALAD model. So that I can compare...
Hi, can you please release the visualization code for Figure 5?
Thank you for your outstanding idea. Can you complete the 40th line of eval.py code MSLS_Test(input_transform=transform)? This result needs to be uploaded to the website for testing, but for the...
Hi, I am curious about how you computed the feature importance in figure 3 in the paper. Would you be able to share some details? Many thanks.
Hi, Thank you for this impressive work! I have a question regarding the dropout layer in SALAD. Have you conducted an ablation study on it? Is its primary purpose to...
Hi, @serizba When I tried to implement SALAD-544, I found that using the parameters in your paper could not obtain the feature dimension of 544, but instead obtained 512. When...
Hi, Thanks for releasing code. I got confuse about difference between MSLS Challenge and MSLS Val dataset. How to validate your provided CKPT file on "MSLS Challenge"? If I am...
