From bff964c429a5bfc8ca85789f20f37d6bfb60b294 Mon Sep 17 00:00:00 2001
From: Marc Sun <57196510+SunMarc@users.noreply.github.com>
Date: Thu, 3 Jul 2025 11:07:11 +0200
Subject: [PATCH] Decouple device_map='auto' and tp_plan='auto'  (#38942)

* dissociate

* better place

* fix
---
 src/transformers/modeling_utils.py | 10 ++++++----
 1 file changed, 6 insertions(+), 4 deletions(-)

diff --git a/src/transformers/modeling_utils.py b/src/transformers/modeling_utils.py
index e6b7031ab3..f7f34c75ea 100644
--- a/src/transformers/modeling_utils.py
+++ b/src/transformers/modeling_utils.py
@@ -4431,10 +4431,12 @@ class PreTrainedModel(nn.Module, ModuleUtilsMixin, PushToHubMixin, PeftAdapterMi
                 "`tp_plan` and `device_map` are mutually exclusive. Choose either one for parallelization."
             )
 
-        # If torchrun was used, make sure to TP by default. This way people don't need to change tp or device map
-        if device_map == "auto" and tp_plan is None and int(os.environ.get("WORLD_SIZE", 0)):
-            tp_plan = "auto"  # device_map = "auto" in torchrun equivalent to TP plan = AUTO!
-            device_map = None
+        if device_map == "auto" and int(os.environ.get("WORLD_SIZE", 0)):
+            logger.info(
+                "You've set device_map=`auto` while triggering a distributed run with torchrun. This might lead to unexpected behavior. "
+                "If your plan is to load the model on each device, you should set device_map={"
+                ": PartialState().process_index} where PartialState comes from accelerate library"
+            )
 
         # We need to correctly dispatch the model on the current process device. The easiest way for this is to use a simple
         # `device_map` pointing to the correct device