Fix bug in save_mem_new_scan due to broadcasting by set_subtensor

Author: ricardoV94

pytensor/scan/rewriting.py

@@ -1516,13 +1516,14 @@ def save_mem_new_scan(fgraph, node):
                     if (
                         nw_inputs[offset + idx].owner
                         and isinstance(nw_inputs[offset + idx].owner.op, IncSubtensor)
+                        and nw_inputs[offset + idx].owner.op.set_instead_of_inc
                         and isinstance(
                             nw_inputs[offset + idx].owner.op.idx_list[0], slice
                         )
+                        # Don't try to create a smart Alloc, if set_subtensor is broadcasting the fill value
+                        # As it happens in set_subtensor(empty(2)[:], 0)
+                        and not(nw_inputs[offset + idx].ndim > nw_inputs[offset + idx].owner.inputs[1].ndim)
                     ):
-                        assert isinstance(
-                            nw_inputs[offset + idx].owner.op, IncSubtensor
-                        )
                         _nw_input = nw_inputs[offset + idx].owner.inputs[1]
                         cval = at.as_tensor_variable(val)
                         initl = at.as_tensor_variable(init_l[i])

tests/scan/test_rewriting.py

@@ -1487,6 +1487,22 @@ def test_while_scan_taps_and_map(self):
         assert stored_ys_steps == 2
         assert stored_zs_steps == 1
 
+    def test_vector_zeros_init(self):
+        ys, _ = pytensor.scan(
+            fn=lambda ytm2, ytm1: ytm1 + ytm2,
+            outputs_info=[{"initial": at.zeros(2), "taps": range(-2, 0)}],
+            n_steps=100,
+        )
+
+        fn = pytensor.function([], ys[-50:], mode=self.mode)
+        assert tuple(fn().shape) == (50,)
+
+        # Check that rewrite worked
+        [scan_node] = (n for n in fn.maker.fgraph.apply_nodes if isinstance(n.op, Scan))
+        _, ys_trace = scan_node.inputs
+        debug_fn = pytensor.function([], ys_trace.shape[0], accept_inplace=True)
+        assert debug_fn() == 50
+
 
 def test_inner_replace_dot():
     """