From: Slaren <redacted>
Date: Wed, 29 Mar 2023 06:53:14 +0000 (+0200)
Subject: Always initialize mm_addr and mm_length in llama_model
X-Git-Tag: gguf-v0.4.0~1044
X-Git-Url: https://git.djapps.eu/?a=commitdiff_plain;h=ac184d514723902f9b05b688703b1be6e8dc65de;p=pkg%2Fggml%2Fsources%2Fllama.cpp

Always initialize mm_addr and mm_length in llama_model
---

diff --git a/llama.cpp b/llama.cpp
index 0c220e4a..aaf5f0ad 100644
--- a/llama.cpp
+++ b/llama.cpp
@@ -150,8 +150,8 @@ struct llama_model {
     std::vector<uint8_t> buf;
 
     // model memory mapped file
-    void * mm_addr;
-    size_t mm_length;
+    void * mm_addr = NULL;
+    size_t mm_length = 0;
 
     // tensors
     int n_loaded;