MLBEDSW-4219: Add tensor allocation info to summary

Added the theoretically minimum max memory usage and
the allocator overhead to the Vela summary.

Signed-off-by: erik.andersson@arm.com <erik.andersson@arm.com>
Change-Id: If373dfeaac50d6f8b56554d435bf22af2c3acda3
diff --git a/ethosu/vela/nn_graph.py b/ethosu/vela/nn_graph.py
index c45d0e3..1973487 100644
--- a/ethosu/vela/nn_graph.py
+++ b/ethosu/vela/nn_graph.py
@@ -154,6 +154,7 @@
 
         self.memory_used = {}
         self.memory_used_per_type = {}
+        self.min_mem_usage = 0
 
     def __str__(self):
         return "<nng.Subgraph '%s',  n_passes=%d, n_cascaded_passes=%d>" % (