diff --git a/.github/checkgroup.yml b/.github/checkgroup.yml
index 923694c31687d..71c1d875abb60 100644
--- a/.github/checkgroup.yml
+++ b/.github/checkgroup.yml
@@ -246,7 +246,7 @@ subprojects:
       - ".github/workflows/ci-app-examples.yml"
       - "src/lightning_app/**"
       - "tests/tests_app_examples/**"
-      - "examples/app_*"
+      - "examples/app_*/**"
       - "requirements/app/**"
       - "setup.py"
       - ".actions/**"
diff --git a/.github/workflows/ci-app-examples.yml b/.github/workflows/ci-app-examples.yml
index 88eadcfd920f8..9646efa27db66 100644
--- a/.github/workflows/ci-app-examples.yml
+++ b/.github/workflows/ci-app-examples.yml
@@ -11,7 +11,7 @@ on:
       - ".github/workflows/ci-app-examples.yml"
       - "src/lightning_app/**"
       - "tests/tests_app_examples/**"
-      - "examples/app_*"
+      - "examples/app_*/**"
       - "requirements/app/**"
       - "setup.py"
       - ".actions/**"
diff --git a/docs/source-app/levels/basic/hello_components/pt_multinode.py b/docs/source-app/levels/basic/hello_components/pt_multinode.py
index 86bd7da10c6ff..8b39c74cdcc2e 100644
--- a/docs/source-app/levels/basic/hello_components/pt_multinode.py
+++ b/docs/source-app/levels/basic/hello_components/pt_multinode.py
@@ -22,7 +22,7 @@ def distributed_train(local_rank: int, main_address: str, main_port: int, num_no
     # 2. PREPARE DISTRIBUTED MODEL
     model = torch.nn.Linear(32, 2)
     device = torch.device(f"cuda:{local_rank}") if torch.cuda.is_available() else torch.device("cpu")
-    model = DistributedDataParallel(model, device_ids=[local_rank]).to(device)
+    model = DistributedDataParallel(model, device_ids=[local_rank] if torch.cuda.is_available() else None).to(device)
 
     # 3. SETUP LOSS AND OPTIMIZER
     criterion = torch.nn.MSELoss()
diff --git a/examples/app_multi_node/train_pytorch.py b/examples/app_multi_node/train_pytorch.py
index 9599bce5bbd85..cc9e84297c151 100644
--- a/examples/app_multi_node/train_pytorch.py
+++ b/examples/app_multi_node/train_pytorch.py
@@ -23,7 +23,7 @@ def distributed_train(local_rank: int, main_address: str, main_port: int, num_no
     # 2. PREPARE DISTRIBUTED MODEL
     model = torch.nn.Linear(32, 2)
     device = torch.device(f"cuda:{local_rank}") if torch.cuda.is_available() else torch.device("cpu")
-    model = DistributedDataParallel(model, device_ids=[local_rank]).to(device)
+    model = DistributedDataParallel(model, device_ids=[local_rank] if torch.cuda.is_available() else None).to(device)
 
     # 3. SETUP LOSS AND OPTIMIZER
     criterion = torch.nn.MSELoss()
@@ -55,7 +55,7 @@ def run(self, main_address: str, main_port: int, num_nodes: int, node_rank: int)
         )
 
 
-# 32 GPUs: (8 nodes x 4 v 100)
+# 8 GPUs: (2 nodes x 4 v 100)
 compute = L.CloudCompute("gpu-fast-multi")  # 4xV100
 component = MultiNode(PyTorchDistributed, num_nodes=2, cloud_compute=compute)
 app = L.LightningApp(component)