# Use AMD ROCm for TotalSegmentator

**URL:** <https://discourse.slicer.org/t/use-amd-rocm-for-totalsegmentator/27270>\
**Category:** Support\
**Tags:** segmentation, ai\
**Created:** [January 16, 2023, 12:57pm UTC](https://discourse.slicer.org/t/use-amd-rocm-for-totalsegmentator/27270 "2023-01-16T12:57:41Z")\
**Posts on this page:** 3\
**Page:** 1

<div class="post-metadata">

**Author:** ![Alex\_Vergara](https://sea2.discourse-cdn.com/flex002/user_avatar/discourse.slicer.org/alex_vergara/32/15205_2.png) [@Alex\_Vergara](https://discourse.slicer.org/u/Alex_Vergara)\
**Post date:** [January 16, 2023, 12:57pm UTC](https://discourse.slicer.org/t/use-amd-rocm-for-totalsegmentator/27270/1 "2023-01-16T12:57:41Z")

</div>

Is there a way to use AMD cards with ROCM? I am thinking of supported cards with the [AUR (en) - python-pytorch-rocm (archlinux.org)](https://aur.archlinux.org/packages/python-pytorch-rocm) package or card specific wheels like [xuhuisheng/rocm-gfx803 (github.com)](https://github.com/xuhuisheng/rocm-gfx803)

---

<div class="post-metadata">

**Author:** ![muratmaga](https://sea2.discourse-cdn.com/flex002/user_avatar/discourse.slicer.org/muratmaga/32/83011_2.png) [@muratmaga](https://discourse.slicer.org/u/muratmaga)\
**Post date:** [January 16, 2023, 5:21pm UTC](https://discourse.slicer.org/t/use-amd-rocm-for-totalsegmentator/27270/2 "2023-01-16T17:21:12Z")

</div>

Looks like auto-installation with the light the torch is only for the CUDA based GPUs. [light-the-torch · PyPI](https://pypi.org/project/light-the-torch/)

Torch already supports those GPUs, so you should be able to manually install those.

> **[PyTorch](https://pytorch.org/blog/pytorch-for-amd-rocm-platform-now-available-as-python-package/)**
>
> An open source machine learning framework that accelerates the path from research prototyping to production deployment.

---

<div class="post-metadata">

**Author:** ![lassoan](https://sea2.discourse-cdn.com/flex002/user_avatar/discourse.slicer.org/lassoan/32/13_2.png) [@lassoan](https://discourse.slicer.org/u/lassoan)\
**Post date:** [January 16, 2023, 6:58pm UTC](https://discourse.slicer.org/t/use-amd-rocm-for-totalsegmentator/27270/3 "2023-01-16T18:58:42Z")

</div>

Light-the-torch supports CPU, CUDA, ROCm backends. Therefore, all ROCm versions that pytorch supports should work well in Slicer.