diff --git a/llmc/utils.h b/llmc/utils.h
index 6f33c40..4223e6b 100644
--- a/llmc/utils.h
+++ b/llmc/utils.h
@@ -15,6 +15,7 @@
 // implementation of dirent for Windows is in dev/unistd.h
 #ifndef _WIN32
 #include <dirent.h>
+#include <arpa/inet.h>
 #else
 #include <winsock2.h>
 #endif
diff --git a/llmc/zero.cuh b/llmc/zero.cuh
index fb0821a..2b19b1b 100644
--- a/llmc/zero.cuh
+++ b/llmc/zero.cuh
@@ -5,12 +5,6 @@ Utilities for ZeRO sharding
 #ifndef LLMC_ZERO_CUH
 #define LLMC_ZERO_CUH
 
-#ifdef _WIN32
-#include <winsock2.h>
-#else
-#include <arpa/inet.h>
-#endif
-
 #include <cuda_runtime_api.h>
 #include <stdint.h>
 #include <stdlib.h>