Reranking network
Closed this issue · 1 comments
yejy53 commented
Thank you for the great work you do. It seems that your reranking network does not require training, and there is no corresponding loss. I used the same processing logic. It seems that the reranking network for my extracted global features is not very ideal.
ShihaoShao-GH commented
You can finetune the hyper-parameters for reranking on a small validation datasets which can further boost the performance.