Make cudnn_helper.h compile independently.

Previously, it would depend on <memory> having already been #included, in order to access std::shared_ptr.
galv · Dec 19, 2017 · 7091721 · 7091721
1 parent 67420bf
commit 7091721
Showing 1 changed file with 1 addition and 0 deletions.
diff --git a/code/nvidia/cudnn_helper.h b/code/nvidia/cudnn_helper.h
@@ -1,5 +1,6 @@
 #pragma once
 
+#include <memory>
 #include <iostream>
 #include <sstream>
 #include <vector>