You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
When I ran the aachen pipeline with SP+SG configuration, I got the results close to the SP+SG baseline (88.6 / 95.3 / 98.7 86.7 / 92.9 / 100.0). But when I ran with SP+NN configuration, the results were far below the baseline. These changes are made:
NetVLAD retrieval top-30
NN with mutual check, distance threshold 0.7
covisibility clustering = True
When I run the NetVLAD by myself, the results are 85.7 / 93.1 / 97.1 66.3 / 81.6 / 93.9.
When I use the loc_pairs provided in pairs/aachen, the results are 86.5 / 93.4 / 97.2 70.4 / 84.7 / 92.9.
Are such performance differences within an acceptable range? Did I misconfigure something or miss something? Thanks!
The text was updated successfully, but these errors were encountered:
When I ran the aachen pipeline with SP+SG configuration, I got the results close to the SP+SG baseline (88.6 / 95.3 / 98.7 86.7 / 92.9 / 100.0). But when I ran with SP+NN configuration, the results were far below the baseline. These changes are made:
When I run the NetVLAD by myself, the results are 85.7 / 93.1 / 97.1 66.3 / 81.6 / 93.9.
When I use the loc_pairs provided in pairs/aachen, the results are 86.5 / 93.4 / 97.2 70.4 / 84.7 / 92.9.
Are such performance differences within an acceptable range? Did I misconfigure something or miss something? Thanks!
The text was updated successfully, but these errors were encountered: