hi,
first of all, very impressive work!
i'm trying to run the given model on the cvusa data set. i didn't find a simple query method in your code (something like - input:ground image, output: matching sat image), so i tried to make one based on your "validate" section in train.py. but i were unable to recreate the results your got (~70% for the top 1%).
can you upload the script the got your these results?
hi,
first of all, very impressive work!
i'm trying to run the given model on the cvusa data set. i didn't find a simple query method in your code (something like - input:ground image, output: matching sat image), so i tried to make one based on your "validate" section in train.py. but i were unable to recreate the results your got (~70% for the top 1%).
can you upload the script the got your these results?