Priya Nagda

@r4pt0s I couldn't complete the component. I Will unassign myself.

> 2\. Few libraries like PaddlePaddle use (B, Num_classes) to compute > See [PaddlePaddle/Paddle#10717](https://github.com/PaddlePaddle/Paddle/pull/10717) > https://github.com/PaddlePaddle/Paddle/blob/develop/python/paddle/nn/functional/loss.py#L41 ``` # compute softmax over the classes axis p = F.softmax(inputs, dim=-1) p =...

Firstly, apologies for the delay, I had my end-term exams going on so had to put the PR on hold. I think this implementation should align with the requirements. I...

Hi @oke-aditya, I've added the basic tests, could you check once? Also, I have trouble understanding the failing checks, could you maybe point out where I've gone wrong?

@oke-aditya I haven't yet given it a detailed look, but I think one thing of concern while comparing the two libraries is the input dimensions, for torchvision, we follow (B,...

Results don't match because in this piece of code here we are specifying the one hot values as per the torchvision input format (B, N1, N2, .., C) (also followed...