From 5f29c4d1164bf4fa21ab4a88ba4f327baf4d72ef Mon Sep 17 00:00:00 2001
From: Jirka Borovec <6035284+Borda@users.noreply.github.com>
Date: Thu, 31 Oct 2024 17:29:46 +0100
Subject: [PATCH] caution

---
 docs/source/pages/overview.rst | 7 ++++---
 1 file changed, 4 insertions(+), 3 deletions(-)

diff --git a/docs/source/pages/overview.rst b/docs/source/pages/overview.rst
index a5d10377c92..34d0dcbd6fc 100644
--- a/docs/source/pages/overview.rst
+++ b/docs/source/pages/overview.rst
@@ -492,9 +492,10 @@ In practice this means that:
 
 A functional metric is differentiable if its corresponding modular metric is differentiable.
 
-For PyTorch versions 2.1 or higher, differentiation in DDP mode is enabled, allowing autograd graph
-propagation after the ``all_gather`` operation. This is useful for synchronizing metrics used as
-loss functions in a DDP setting.
+.. caution::
+    For PyTorch versions 2.1 or higher, differentiation in DDP mode is enabled, allowing autograd graph
+    propagation after the ``all_gather`` operation. This is useful for synchronizing metrics used as
+    loss functions in a DDP setting.
 
 ***************************************
 Metrics and hyperparameter optimization