I don’t understand what could be causing this. It could be an issue in the implementation of the logic. I would appreciate if someone could help me out with it. I understand that debugging this will be much easier with the code, so if anyone can help me with this, I am happy to send the code to them privately
your triplet loss score is not matching with the expected output.
use search engine with triplet loss, you will find many similar threads
make sure your inputs have correctly written v2 and then v1
I did go through the threads, and altered parts of my code accordingly, for example, multiplying v2 by transpose of v1. But this was the closest I got my triplet loss to be to the expected output. Before the changes, the value was somewhere around 2.4.
Is there anything else that could be causing this?
what about triplet loss1 and tripletloss2 calculation, did you tuple them correctly?
you can share screenshot of the code by personal DM, in case you are not Able to find the issue.
Done! Thank you so much!
hi @N_Adhikari
for the code line instruction
the second mask to extract elements in the negative_zero_on_duplicate matrix that are larger than the elements in the diagonal
you have subtracted negative_zero_on_duplicate -positive which you don’t need to do.
only check if negative_zero_on_duplicate> elements in the diagonal.
All other codes are correct
Thank you so much for your help!
1 Like
did you resolve your other query?
The one for the classify function? No, I still haven’t figured that one out
I have replied to that post, please refer and reply there for that query