# Distribution Implementations

**URL:** <https://discuss.pytorch.org/t/distribution-implementations/400>\
**Category:** Uncategorized\
**Created:** [February 8, 2017, 9:40pm UTC](https://discuss.pytorch.org/t/distribution-implementations/400 "2017-02-08T21:40:07Z")\
**Posts on this page:** 1\
**Showing post:** 7

<div class="post-metadata">

**Author:** ![junpenglao](https://discuss.pytorch.org/user_avatar/discuss.pytorch.org/junpenglao/32/188_2.png) [@junpenglao](https://discuss.pytorch.org/u/junpenglao)\
**Post date:** [March 3, 2017, 3:20pm UTC](https://discuss.pytorch.org/t/distribution-implementations/400/7 "2017-03-03T15:20:05Z")

</div>

Actually, I do want to sample from it, as I am interested in using them for [Bayesian inference](https://discuss.pytorch.org/t/bayesian-computation-in-pytorch/755/5). In that case, maybe it is better to wrap it in a `Function` class?

---

_[View the full topic](https://discuss.pytorch.org/t/distribution-implementations/400)._