Make torch custom operator test deterministic

test/test_torch.py

@@ -160,9 +160,10 @@ class TestTorch(unittest.TestCase):
         )
 
     def test_feature_selection_custom_operators(self):
-        X = pd.DataFrame({f"k{i}": np.random.randn(1000) for i in range(10, 21)})
+        rstate = np.random.RandomState(0)
+        X = pd.DataFrame({f"k{i}": rstate.randn(2000) for i in range(10, 21)})
         cos_approx = lambda x: 1 - (x**2) / 2 + (x**4) / 24 + (x**6) / 720
-        y = X["k15"] ** 2 + cos_approx(X["k20"])
+        y = X["k15"] ** 2 + 2 * cos_approx(X["k20"])
 
         model = PySRRegressor(
             progress=False,
@@ -172,7 +173,12 @@ class TestTorch(unittest.TestCase):
             early_stop_condition=1e-5,
             extra_sympy_mappings={"cos_approx": cos_approx},
             extra_torch_mappings={"cos_approx": cos_approx},
+            random_state=0,
+            deterministic=True,
+            procs=0,
+            multithreading=False,
         )
+        np.random.seed(0)
         model.fit(X.values, y.values)
         torch_module = model.pytorch()
 
@@ -180,4 +186,5 @@ class TestTorch(unittest.TestCase):
 
         torch_output = torch_module(torch.tensor(X.values)).detach().numpy()
 
-        np.testing.assert_almost_equal(np_output, torch_output, decimal=4)
+        np.testing.assert_almost_equal(y.values, np_output, decimal=4)
+        np.testing.assert_almost_equal(y.values, torch_output, decimal=4)