Open PeterKim1 opened 2 years ago
Hello.
Thanks for your great codes.
I have a question about --anomaly_scorer_num_nn.
In your paper, when the anomaly score is calculated, consider b nearest patch-features in memory bank.
is --anomaly_scorer_num_nn the same as b ??
And another question is, if i want to reproduce paper's performance, is it okay to set --anomaly_scorer_num_nn as 1?
Or need to set another value? is it different by subdataset??
Thanks.
Hello.
Thanks for your great codes.
I have a question about --anomaly_scorer_num_nn.
In your paper, when the anomaly score is calculated, consider b nearest patch-features in memory bank.
is --anomaly_scorer_num_nn the same as b ??
And another question is, if i want to reproduce paper's performance, is it okay to set --anomaly_scorer_num_nn as 1?
Or need to set another value? is it different by subdataset??
Thanks.