..

tests/benchmark-models/test_petab_benchmark.py

@@ -89,6 +89,13 @@ class GradientCheckSettings:
 settings["Giordano_Nature2020"] = GradientCheckSettings(
     atol_check=1e-6, rtol_check=1e-3, rng_seed=1
 )
+settings["Weber_BMC2015"] = GradientCheckSettings(
+    atol_sim=1e-12,
+    rtol_sim=1e-12,
+    atol_check=1e-6,
+    rtol_check=1e-2,
+    rng_seed=1,
+)
 
 
 def assert_gradient_check_success(
@@ -115,8 +122,14 @@ def assert_gradient_check_success(
     df = check_result.df
     df["abs_diff"] = np.abs(df["expectation"] - df["test"])
     df["rel_diff"] = df["abs_diff"] / np.abs(df["expectation"])
+    max_adiff = df["abs_diff"].max()
+    max_rdiff = df["rel_diff"].max()
     with pd.option_context("display.max_columns", None, "display.width", None):
-        raise AssertionError(f"Gradient check failed:\n{df}")
+        raise AssertionError(
+            f"Gradient check failed:\n{df}\n\n"
+            f"Maximum absolute difference: {max_adiff}\n"
+            f"Maximum relative difference: {max_rdiff}"
+        )
 
 
 @pytest.mark.filterwarnings(
@@ -160,11 +173,7 @@ def test_benchmark_gradient(model, scale, sensitivity_method, request):
         pytest.skip()
 
     if (
-        model
-        in (
-            "Weber_BMC2015",
-            "Sneyd_PNAS2002",
-        )
+        model in ("Sneyd_PNAS2002",)
         and sensitivity_method == SensitivityMethod.forward
     ):
         # FIXME
@@ -209,7 +218,7 @@ def test_benchmark_gradient(model, scale, sensitivity_method, request):
         # cache=not debug,
         cache=False,
     )
-    noise_level = 0.1
+    noise_level = 0.05
     np.random.seed(cur_settings.rng_seed)
 
     # find a point where the derivative can be computed