# What version of PyTorch should we use with Ray Train?

**URL:** <https://discuss.ray.io/t/what-version-of-pytorch-should-we-use-with-ray-train/4663>\
**Category:** Ray Train\
**Created:** [January 11, 2022, 9:09pm UTC](https://discuss.ray.io/t/what-version-of-pytorch-should-we-use-with-ray-train/4663 "2022-01-11T21:09:17Z")\
**Posts on this page:** 2\
**Page:** 1

<div class="post-metadata">

**Author:** ![Lacruche](https://avatars.discourse-cdn.com/v4/letter/l/ecccb3/32.png) [@Lacruche](https://discuss.ray.io/u/Lacruche)\
**Post date:** [January 11, 2022, 9:09pm UTC](https://discuss.ray.io/t/what-version-of-pytorch-should-we-use-with-ray-train/4663/1 "2022-01-11T21:09:17Z")

</div>

What version of PyTorch should we use with Ray Train? (ray 1.9.1)

---

<div class="post-metadata">

**Author:** ![matthewdeng](https://sea2.discourse-cdn.com/flex020/user_avatar/discuss.ray.io/matthewdeng/32/1446_2.png) [@matthewdeng](https://discuss.ray.io/u/matthewdeng)\
**Post date:** [January 11, 2022, 11:16pm UTC](https://discuss.ray.io/t/what-version-of-pytorch-should-we-use-with-ray-train/4663/2 "2022-01-11T23:16:11Z")

</div>

There isn’t a requirement, but `ray-ml` docker images are built with `1.9.0` (or `1.9.0+cu111` for GPU) and have been tested to work!