From e05af29c219c38b6e1add544d66841cc0eb4abb6 Mon Sep 17 00:00:00 2001
From: Matteo Bunino <48362942+matbun@users.noreply.github.com>
Date: Thu, 19 Sep 2024 17:49:09 +0200
Subject: [PATCH] Fix containers tutorial (#218)

* ADD NV container toolkit

* RM RST experiment
---
 .../distributed-ml/torch-tutorial-containers/README.md      | 6 ++++++
 1 file changed, 6 insertions(+)

diff --git a/tutorials/distributed-ml/torch-tutorial-containers/README.md b/tutorials/distributed-ml/torch-tutorial-containers/README.md
index 9400f607..9f90025a 100644
--- a/tutorials/distributed-ml/torch-tutorial-containers/README.md
+++ b/tutorials/distributed-ml/torch-tutorial-containers/README.md
@@ -40,6 +40,12 @@ in this folder:
 itwinai exec-pipeline --config config.yaml --pipe-key training_pipeline
 ```
 
+> [!WARNING]
+> When using Docker, if your container does not recognizes the GPUs of your VM
+> you may need to install the
+> [Nvidia container toolkit](https://docs.nvidia.com/datacenter/cloud-native/container-toolkit/latest/index.html)
+> , if not already installed.
+
 ## Singularity (HPC environments)
 
 With singularity there is no need to explicitly bind mount the current working directory (CWD) in the container