Sebastian Raschka

I don't think this issue is related to CORAL; you may have some issues with your dataset because it cannot complete the first epoch. I suggest you iterate over your...

Good point, there was a related issue here #22 I.e., make sure that the labels start at 0 by subtracting "min(age)" from all labels during training. Then, to make predictions,...

Hm that's weird. Haven't seen this before. I just ran the script and it seems to work fine for me. The software versions were # DLIB: 19.22.0 # NumPy: 1.20.2...

I would call it "guaranteeing" monotonicity :). Yeah, it's the bias and the weight sharing together with the loss function. Regarding the changes compared to a regular network, maybe this...

Yeah most of the code here is the code used straight up what's used in the paper. So, given this code, it is expected that users label the dataset such...

We used the frontal face detector in DLIB for CACD, for example. There are probably better ones but this one seemed quite reliable and was easy to use off-the-shelf. The...

Thanks for the comment. Just to make sure we are on the same page, your question is why we used a different task importance weight than Niu et al? We...

Ordinal CNN is the method by Niu et al. 2016 you mentioned at the top of this thread ("Ordinal Regression with Multiple Output CNN for Age Estimation"). Sure. We have...

Thanks! But I see that there are two periods now: ``` c.execute("UPDATE {tn} SET {cn}='sebastian_r' WHERE {idf}=123456".\ .format(tn=table_name, idf=id_column, cn=new_column)) ``` Could you please remove the upper one?

No worries, and I think it's visually a bit misleading. I think we could leave it as is. It has a bit of those "when you see you old code...