refactor(mge/test): make test_parampack more stable

GitOrigin-RevId: d82230ea07b49dd08a9e968974de6950e1987a08

python_module/test/integration/test_parampack.py

@@ -105,9 +105,15 @@ def test_static_graph_parampack():
 
     assert np.mean(losses[-100:]) < 0.1, "Final training Loss must be low enough"
 
-    data, _ = next(train_dataset)
+    ngrid = 10
+    x = np.linspace(-1.0, 1.0, ngrid)
+    xx, yy = np.meshgrid(x, x)
+    xx = xx.reshape((ngrid * ngrid, 1))
+    yy = yy.reshape((ngrid * ngrid, 1))
+    data = np.concatenate((xx, yy), axis=1).astype(np.float32)
+
     pred = infer(data).numpy()
-    assert calculate_precision(data, pred) > 0.95, "Test precision must be high enough"
+    assert calculate_precision(data, pred) == 1.0, "Test precision must be high enough"
 
 
 @pytest.mark.slow
@@ -140,9 +146,15 @@ def test_nopack_parampack():
         losses.append(loss.numpy())
     assert np.mean(losses[-100:]) < 0.1, "Final training Loss must be low enough"
 
-    data, _ = next(train_dataset)
+    ngrid = 10
+    x = np.linspace(-1.0, 1.0, ngrid)
+    xx, yy = np.meshgrid(x, x)
+    xx = xx.reshape((ngrid * ngrid, 1))
+    yy = yy.reshape((ngrid * ngrid, 1))
+    data = np.concatenate((xx, yy), axis=1).astype(np.float32)
+
     pred = infer(data).numpy()
-    assert calculate_precision(data, pred) > 0.95, "Test precision must be high enough"
+    assert calculate_precision(data, pred) == 1.0, "Test precision must be high enough"
 
 
 @pytest.mark.slow
@@ -178,9 +190,15 @@ def test_dynamic_graph_parampack():
 
     assert np.mean(losses[-100:]) < 0.1, "Final training Loss must be low enough"
 
-    data, _ = next(train_dataset)
+    ngrid = 10
+    x = np.linspace(-1.0, 1.0, ngrid)
+    xx, yy = np.meshgrid(x, x)
+    xx = xx.reshape((ngrid * ngrid, 1))
+    yy = yy.reshape((ngrid * ngrid, 1))
+    data = np.concatenate((xx, yy), axis=1).astype(np.float32)
+
     pred = infer(data).numpy()
-    assert calculate_precision(data, pred) > 0.95, "Test precision must be high enough"
+    assert calculate_precision(data, pred) == 1.0, "Test precision must be high enough"
 
 
 @pytest.mark.slow