[RLlib] Forward fix for failing PPO Torch RLTrainer test (#32308)

Signed-off-by: Kourosh Hakhamaneshi <kourosh@anyscale.com>
ray-project · Feb 8, 2023 · b85eb52 · b85eb52
1 parent 59c62e4
commit b85eb52
Showing 1 changed file with 5 additions and 1 deletion.
diff --git a/rllib/algorithms/ppo/tests/test_ppo_rl_trainer.py b/rllib/algorithms/ppo/tests/test_ppo_rl_trainer.py
@@ -20,6 +20,10 @@
         [[0.1, 0.2, 0.3, 0.4], [0.5, 0.6, 0.7, 0.8], [0.9, 1.0, 1.1, 1.2]],
         dtype=np.float32,
     ),
+    SampleBatch.NEXT_OBS: np.array(
+        [[0.1, 0.2, 0.3, 0.4], [0.5, 0.6, 0.7, 0.8], [0.9, 1.0, 1.1, 1.2]],
+        dtype=np.float32,
+    ),
     SampleBatch.ACTIONS: np.array([0, 1, 1]),
     SampleBatch.PREV_ACTIONS: np.array([0, 1, 1]),
     SampleBatch.REWARDS: np.array([1.0, -1.0, 0.5], dtype=np.float32),
@@ -57,7 +61,7 @@ def test_loss(self):
             .training(
                 gamma=0.99,
                 model=dict(
-                    fcnet_hiddens=[10],
+                    fcnet_hiddens=[10, 10],
                     fcnet_activation="linear",
                     vf_share_layers=False,
                 ),