diff --git a/lightning_examples/warp-drive/.meta.yml b/lightning_examples/warp-drive/.meta.yml
index 9a65f6545..988bfde35 100644
--- a/lightning_examples/warp-drive/.meta.yml
+++ b/lightning_examples/warp-drive/.meta.yml
@@ -21,5 +21,9 @@ description: This notebook introduces multi-agent reinforcement learning (MARL)
 requirements:
   - rl-warp-drive>=1.6.5
   - ffmpeg-python
+  # todo: after merging #155 we will relax this just to `torch<=1.10` and drop TV, TT, etc.
+  - torch==1.10.*
+  - torchvision==0.11.*
+  - torchtext==0.11.*
 accelerator:
   - GPU
diff --git a/lightning_examples/warp-drive/multi_agent_rl.py b/lightning_examples/warp-drive/multi_agent_rl.py
index d4c5ec8c6..753dd2836 100644
--- a/lightning_examples/warp-drive/multi_agent_rl.py
+++ b/lightning_examples/warp-drive/multi_agent_rl.py
@@ -225,6 +225,8 @@
     devices=num_gpus,
     callbacks=[cuda_callback, perf_stats_callback],
     max_epochs=num_epochs,
+    log_every_n_steps=1,
+    reload_dataloaders_every_n_epochs=1,
 )
 
 # %%