Something about baseline

[ToxicDoubleH]

Hi, nice job!
I am curious about how to implement the baseline.
for dataset in datasets: for index in range(iters): loss = cls + tri + cent
Is the baseline written like this?
And in your Table 4, line 2, as follows

Did you use unify the classification loss of the label space of all datasets ,triplet ans centerloss?
Looking forward to your reply.