diff --git a/.ci/docker/requirements.txt b/.ci/docker/requirements.txt
index c032835633..89dd788ae7 100644
--- a/.ci/docker/requirements.txt
+++ b/.ci/docker/requirements.txt
@@ -28,8 +28,8 @@ tensorboard
 jinja2==3.1.3
 pytorch-lightning
 torchx
-torchrl==0.6.0
-tensordict==0.6.0
+torchrl==0.7.2
+tensordict==0.7.2
 ax-platform>=0.4.0
 nbformat>=5.9.2
 datasets
diff --git a/.jenkins/validate_tutorials_built.py b/.jenkins/validate_tutorials_built.py
index ed790a1887..f78ec11e1a 100644
--- a/.jenkins/validate_tutorials_built.py
+++ b/.jenkins/validate_tutorials_built.py
@@ -50,6 +50,7 @@
     "intermediate_source/flask_rest_api_tutorial",
     "intermediate_source/text_to_speech_with_torchaudio",
     "intermediate_source/tensorboard_profiler_tutorial", # reenable after 2.0 release.
+    "advanced_source/semi_structured_sparse" # reenable after 3303 is fixed.
 ]
 
 def tutorial_source_dirs() -> List[Path]:
diff --git a/advanced_source/coding_ddpg.py b/advanced_source/coding_ddpg.py
index 3ee4ddc39b..90ea4565da 100644
--- a/advanced_source/coding_ddpg.py
+++ b/advanced_source/coding_ddpg.py
@@ -1040,7 +1040,7 @@ def ceil_div(x, y):
 
 ###############################################################################
 # let's use the TD(lambda) estimator!
-loss_module.make_value_estimator(ValueEstimators.TDLambda, gamma=gamma, lmbda=lmbda)
+loss_module.make_value_estimator(ValueEstimators.TDLambda, gamma=gamma, lmbda=lmbda, device=device)
 
 ###############################################################################
 # .. note::
diff --git a/advanced_source/semi_structured_sparse.py b/advanced_source/semi_structured_sparse.py
index 38c2c6878b..1bf3586045 100644
--- a/advanced_source/semi_structured_sparse.py
+++ b/advanced_source/semi_structured_sparse.py
@@ -210,6 +210,8 @@
 SparseSemiStructuredTensor._FORCE_CUTLASS = True
 torch.manual_seed(100)
 
+# Set default device to "cuda:0"
+torch.set_default_device(torch.device("cuda:0" if torch.cuda.is_available() else "cpu"))
 
 ######################################################################
 # We’ll also need to define some helper functions that are specific to the
diff --git a/intermediate_source/reinforcement_ppo.py b/intermediate_source/reinforcement_ppo.py
index b25a6f8c8a..44d5c05a8a 100644
--- a/intermediate_source/reinforcement_ppo.py
+++ b/intermediate_source/reinforcement_ppo.py
@@ -551,7 +551,7 @@
 #
 
 advantage_module = GAE(
-    gamma=gamma, lmbda=lmbda, value_network=value_module, average_gae=True
+    gamma=gamma, lmbda=lmbda, value_network=value_module, average_gae=True, device=device,
 )
 
 loss_module = ClipPPOLoss(