Fix bug in gradient of `set_subtensor`

pytensor/tensor/subtensor.py

@@ -1902,7 +1902,7 @@ def _sum_grad_over_bcasted_dims(x, gx):
     if gx.broadcastable != x.broadcastable:
         x_dim_added = gx.ndim - x.ndim
         x_broad = (True,) * x_dim_added + x.broadcastable
-        assert sum(gx.broadcastable) < sum(x_broad)
+        assert sum(gx.broadcastable) <= sum(x_broad)
         axis_to_sum = []
         for i in range(gx.ndim):
             if gx.broadcastable[i] is False and x_broad[i] is True:

tests/tensor/test_subtensor.py

@@ -1593,6 +1593,15 @@ class TestIncSubtensor:
                 ),
             )
 
+            # Broadcastable leading dim
+            utt.verify_grad(
+                f_slice(slice(None, None), slice(1, 3)),
+                (
+                    np.asarray([0, 1, 2, 3, 4, 5.0])[None, ...],
+                    np.asarray([9, 9.0]),
+                ),
+            )
+
 
 class TestIncSubtensor1:
     def setup_method(self):