Weijian Xu
Weijian Xu
I think with the current PyTorch library, it would be even simpler to use `torch.nn.functional.unfold` to implement the correlation function.
For MN branch, I wonder if the weights are shared between user and shop images, or you use different weights for user and shop images?
@seokhyeonSong Thank you for your code! However, I wonder if you implement the Match-Net (for retrieval) or not. Recently I have some problems in reproducing the retrieval numbers. Thank you...
Thanks anyway! Currently I am implementing the Match R-CNN but still feel confused about some details mentioned in the paper.
+1 Hope that the author can provide the coefficient of the losses.
@Dayan-Zhanchi I reproduced the match-net baseline with mask features (i.e. use the RoI features after the same RoIAlign with mask branch's). My re-implementation is based on maskrcnn-benchmark. Reported results in...
> @xwjabc Can you explain the value lambdas you used for each of the different losses? Hi simaoh, I simply use 1 as the coefficient for all losses.
> @xwjabc Could you share with me the number of epochs, batch size and lr optimizer steps? I also use two 2 x TITAN RTX GPU. Thank you! Some information...
Thank you for your great help! Besides, I have two more questions: 1. In the first version of the answer of the match network, I noticed that there are several...
Thank you for your great help! In my current implementation, I use the mask features after RoIAlign in the mask branch. However, the number of instances in the mask features...