Fix conversion of scalar constants into numpy arrays in the evaluator

onnxscript/evaluator.py

@@ -376,6 +376,8 @@
         return v
     if isinstance(v, np.ndarray):
         return v
+    if isinstance(v, np.number):
+        return np.array(v)
     raise TypeError(
         f"Unexpected onnxscript value type '{type(v)}'."
         "Valid value types are 'Tensor | list[Tensor] | None | np.ndarray | list[np.ndarray]'"

tests/issues_2024_test.py

@@ -0,0 +1,94 @@
+# Copyright (c) Microsoft Corporation.
+# Licensed under the MIT License.
+
+import unittest
+
+import numpy as np
+import onnx
+import onnxruntime as ort
+
+from onnxscript import FLOAT, script
+from onnxscript import opset20 as op
+from tests.common import testutils
+
+
+class TestIssues2024(testutils.TestBase):
+    def test_issue_1969(self):
+        # https://github.com/microsoft/onnxscript/issues/1969
+
+        x = np.array([1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13], dtype=np.float32)
+        y = np.array(x[len(x) - 3 : len(x)], dtype=np.float32)
+        z = np.array(x[len(x) - 8 : len(x) - 3], dtype=np.float32)
+
+        x_coef = np.array(
+            [1.1, 1.5, 1.7, 1.9, 1.2, 3.1, 4.5, 5.2, 8.5, 9.0, 11.0], dtype=np.float32
+        )
+        y_coef = np.array([8.5, 9.0, 11.0], dtype=np.float32)
+        z_coef = np.array([5, 7, 9, 9, 10, 11], dtype=np.float32)
+
+        t1 = len(x)
+        t2 = len(y)
+        t3 = len(z)
+
+        const_term = 2.37698
+        h1 = 1.5689
+        h2 = 1.799
+
+        @script()
+        def conv(a1: FLOAT, a2: FLOAT, a3: FLOAT, a4: FLOAT, a5: FLOAT, a6: FLOAT) -> FLOAT:
+            const = op.Constant(value_float=const_term)
+            # Define constants
+            coeff1 = op.Constant(value_float=h1)
+            coeff2 = op.Constant(value_float=h2)
+            s = coeff1 * a1 + coeff2 * a2 - const
+
+            for inx in range(t1):
+                if a4 == x[inx]:
+                    s = s + x_coef[inx]
+            for inx2 in range(t2):
+                if a5 == y[inx2]:
+                    s = s + y_coef[inx2]
+            for inx3 in range(t3):
+                if a6 == z[inx3]:
+                    s = s + z_coef[inx3]
+
+            return op.Exp(s)
+
+        onx = conv.to_model_proto()
+        # To save to model and visualize it.
+        # onnx.save(onx, "test_issue_1969.onnx")
+        onnx.checker.check_model(onx)
+
+        sess = ort.InferenceSession(onx.SerializeToString())
+        feeds = {f"a{i}": np.array([i], dtype=np.float32) for i in range(1, 7)}
+        got = sess.run(None, feeds)
+
+        def conv_np(a1, a2, a3, a4, a5, a6):
+            const = const_term
+            # Define constants
+            coeff1 = h1
+            coeff2 = h2
+            s = coeff1 * a1 + coeff2 * a2 - const
+
+            for inx in range(t1):
+                if a4 == x[inx]:
+                    s = s + x_coef[inx]
+            for inx2 in range(t2):
+                if a5 == y[inx2]:
+                    s = s + y_coef[inx2]
+            for inx3 in range(t3):
+                if a6 == z[inx3]:
+                    s = s + z_coef[inx3]
+
+            return np.exp(s)
+
+        expected = conv_np(**feeds)
+        np.allclose(expected, got[0], rtol=1e-6)
+        # Unexpected onnxscript value type '<class 'numpy.float32'>'.
+        # Valid value types are 'Tensor | list[Tensor] | None | np.ndarray | list[np.ndarray]'
+        expected2 = conv(**feeds)
+        np.allclose(expected2, got[0], rtol=1e-6)
+
+
+if __name__ == "__main__":
+    unittest.main(verbosity=2)