minor comments and refactors

ServiceNow · Oct 8, 2024 · 8891d50 · 8891d50
1 parent 04b44f7
commit 8891d50
Show file tree

Hide file tree

Showing 2 changed files with 8 additions and 8 deletions.
diff --git a/browsergym/experiments/src/browsergym/experiments/benchmark.py b/browsergym/experiments/src/browsergym/experiments/benchmark.py
@@ -212,24 +212,24 @@ def task_list_from_csv(
             seeds_rng=np.random.RandomState(42),
         ),
     ),
-    "workarena_l2_agent_curriculum": lambda: Benchmark(
-        name="workarena_l2_agent_curriculum",
+    "workarena_l2_agent_curriculum_eval": lambda: Benchmark(
+        name="workarena_l2_agent_curriculum_eval",
         high_level_action_set_args=DEFAULT_HIGHLEVEL_ACTION_SET_ARGS["workarena"],
         env_args_list=_make_env_args_list_from_workarena_curriculum(
             level="l2",
             task_category_filter=None,
-            meta_seed=42,
+            meta_seed=42,  # meta seed for evaluation curriculum
             max_steps=50,
             curriculum_type="agent",
         ),
     ),
-    "workarena_l3_agent_curriculum": lambda: Benchmark(
-        name="workarena_l3_agent_curriculum",
+    "workarena_l3_agent_curriculum_eval": lambda: Benchmark(
+        name="workarena_l3_agent_curriculum_eval",
         high_level_action_set_args=DEFAULT_HIGHLEVEL_ACTION_SET_ARGS["workarena"],
         env_args_list=_make_env_args_list_from_workarena_curriculum(
             level="l3",
             task_category_filter=None,
-            meta_seed=42,
+            meta_seed=42,  # meta seed for evaluation curriculum
             max_steps=50,
             curriculum_type="agent",
         ),

diff --git a/tests/experiments/test_benchmark.py b/tests/experiments/test_benchmark.py
@@ -60,8 +60,8 @@ def test_build_benchmarks():
         "visualwebarena": 910,
         "workarena_l1": 33 * 10,
         "workarena_l1_sort": 6 * 10,
-        "workarena_l2_agent_curriculum": 235,
-        "workarena_l3_agent_curriculum": 235,
+        "workarena_l2_agent_curriculum_eval": 235,
+        "workarena_l3_agent_curriculum_eval": 235,
     }
     for name, benchmark_builder in BENCHMARKS.items():
         benchmark = benchmark_builder()