liumouliu / deepibl Goto Github PK

Hi,

Thanks for your nice work!
I have several questions about the following table

53e0 Best epoch: 4 (out of 7)
===========================================
Recall@N      0001 0002 0003 0004 0005 0010 
===========================================
off-the-shelf 0.80 0.86 0.90 0.92 0.93 0.96 
our trained   0.89 0.94 0.95 0.96 0.97 0.98 
trained/shelf 1.12 1.09 1.06 1.05 1.05 1.02

1. Does the off-the-shelf here include (ImageNet pretrained) VGG16+VLAD+PCA?
2. What does trained/shelf mean?
3. I am now trying to implement your results by PyTorch, but the results seem similar to the raw NetVLAD. So do you have any suggestions?

Thank you very much!

Hi Liuliu,
Thanks for your amazing job.
As stated in the paper, you took pitts30k-train as the training dataset and others for the test.
But in the sample script, the code shows that pitts250k-train is taken as training dataset. Which one should we follow to reproduce the results? Or the model trained with pitts250k is used for classification?
Another question is that: are the results in table 2 of the paper gotten after implementing pca whitening with dim 4096?