[RLlib] Fix test case multi-agent Pendulum PPO (--num-agents=2 arg mi…

…ssing in BUILD). (#45820)
ray-project · Jun 10, 2024 · 829ac71 · 829ac71
1 parent 13940c9
commit 829ac71
Show file tree

Hide file tree

Showing 2 changed files with 8 additions and 17 deletions.
diff --git a/rllib/BUILD b/rllib/BUILD
@@ -417,7 +417,7 @@ py_test(
     tags = ["team:rllib", "exclusive", "learning_tests", "learning_tests_pendulum", "learning_tests_continuous", "torch_only"],
     size = "large", # bazel may complain about it being too long sometimes - large is on purpose as some frameworks take longer
     srcs = ["tuned_examples/ppo/multi_agent_pendulum_ppo.py"],
-    args = ["--as-test", "--enable-new-api-stack"]
+    args = ["--enable-new-api-stack", "--num-agents=2", "--as-test"]
 )
 
 #@OldAPIStack

diff --git a/rllib/algorithms/bc/tests/test_bc.py b/rllib/algorithms/bc/tests/test_bc.py
@@ -48,7 +48,7 @@ def test_bc_compilation_and_learning_from_offline_file(self):
             .offline_data(input_=[data_file])
         )
         num_iterations = 350
-        min_reward = 75.0
+        min_return_to_reach = 75.0
 
         # Test for RLModule API and ModelV2.
         for rl_modules in [True, False]:
@@ -79,29 +79,20 @@ def test_bc_compilation_and_learning_from_offline_file(self):
 
                         eval_results = results.get("evaluation")
                         if eval_results:
-                            print(
-                                "iter={} R={}".format(
-                                    i,
-                                    eval_results[
-                                        f"{ENV_RUNNER_RESULTS}/{EPISODE_RETURN_MEAN}"
-                                    ],
-                                )
-                            )
+                            mean_return = eval_results[ENV_RUNNER_RESULTS][
+                                EPISODE_RETURN_MEAN
+                            ]
+                            print("iter={} R={}".format(i, mean_return))
                             # Learn until good reward is reached in the actual env.
-                            if (
-                                eval_results[
-                                    f"{ENV_RUNNER_RESULTS}/{EPISODE_RETURN_MEAN}"
-                                ]
-                                > min_reward
-                            ):
+                            if mean_return > min_return_to_reach:
                                 print("learnt!")
                                 learnt = True
                                 break
 
                     if not learnt:
                         raise ValueError(
                             "`BC` did not reach {} reward from expert offline "
-                            "data!".format(min_reward)
+                            "data!".format(min_return_to_reach)
                         )
 
                     check_compute_single_action(algo, include_prev_action_reward=True)