From d3e7c655e5b4ce3fb15ecd7e27064f8df727ce14 Mon Sep 17 00:00:00 2001
From: Matthew Jenkins <40323108+Matthew-Jenkins@users.noreply.github.com>
Date: Sun, 20 Apr 2025 22:06:24 -0400
Subject: [PATCH] Add support for llama-cpp builds from
 https://github.com/ggml-org/llama.cpp (#6862)

---
 modules/llama_cpp_server.py | 12 +++++++++++-
 1 file changed, 11 insertions(+), 1 deletion(-)

diff --git a/modules/llama_cpp_server.py b/modules/llama_cpp_server.py
index 02a56b3c..c88f945d 100644
--- a/modules/llama_cpp_server.py
+++ b/modules/llama_cpp_server.py
@@ -1,4 +1,5 @@
 import json
+import os
 import pprint
 import socket
 import subprocess
@@ -281,12 +282,21 @@ class LlamaServer:
         if shared.args.rope_freq_base > 0:
             cmd += ["--rope-freq-base", str(shared.args.rope_freq_base)]
 
+        env = os.environ.copy()
+        if os.name == 'posix':
+            current_path = env.get('LD_LIBRARY_PATH', '')
+            if current_path:
+                env['LD_LIBRARY_PATH'] = f"{current_path}:{os.path.dirname(self.server_path)}"
+            else:
+                env['LD_LIBRARY_PATH'] = os.path.dirname(self.server_path)
+
         # Start the server with pipes for output
         self.process = subprocess.Popen(
             cmd,
             stderr=subprocess.PIPE,
             text=True,
-            bufsize=1
+            bufsize=1,
+            env=env
         )
 
         def filter_stderr(process_stderr):