visionml/pytracking

tomp

ChuangGe66 opened this issue · 2 comments

ChuangGe66 commented

Hello, your team's work is excellent. Here I would like to ask one question. After tomp training, there are 300 weights, which weight should I use for testing. Looking forward to your team's reply.

2006pmach commented

You can use the most recent (300) or the one that achieves the smallest validation loss. I usually just use the most recent.

ChuangGe66 commented

I'm so excited you took the time to answer me