# Differentiable Spearman in PyTorch (Optimize for CORR directly)

**URL:** <https://forum.numer.ai/t/differentiable-spearman-in-pytorch-optimize-for-corr-directly/2287>\
**Category:** Data Science\
**Created:** [March 11, 2021, 4:21am UTC](https://forum.numer.ai/t/differentiable-spearman-in-pytorch-optimize-for-corr-directly/2287 "2021-03-11T04:21:12Z")\
**Posts on this page:** 1\
**Showing post:** 18

<div class="post-metadata">

**Author:** ![paulito](https://avatars.discourse-cdn.com/v4/letter/p/e19adc/32.png) [@paulito](https://forum.numer.ai/u/paulito)\
**Post date:** [May 5, 2021, 10:05am UTC](https://forum.numer.ai/t/differentiable-spearman-in-pytorch-optimize-for-corr-directly/2287/18 "2021-05-05T10:05:08Z")

</div>

Look at the source code. It already has a tensorflow implementation of fast\_sort\_soft. [GitHub - google-research/fast-soft-sort: Fast Differentiable Sorting and Ranking](https://github.com/google-research/fast-soft-sort/)

---

_[View the full topic](https://forum.numer.ai/t/differentiable-spearman-in-pytorch-optimize-for-corr-directly/2287)._