From 51355234290ac3adb0ee0df597aa6a3bb9189cb4 Mon Sep 17 00:00:00 2001
From: oobabooga <112222186+oobabooga@users.noreply.github.com>
Date: Fri, 18 Apr 2025 17:10:26 -0700
Subject: [PATCH] Fix the new llama.cpp loader failing to unload models

---
 modules/llama_cpp_server.py | 16 ++++++++++------
 1 file changed, 10 insertions(+), 6 deletions(-)

diff --git a/modules/llama_cpp_server.py b/modules/llama_cpp_server.py
index a6eb9b20..8180f974 100644
--- a/modules/llama_cpp_server.py
+++ b/modules/llama_cpp_server.py
@@ -280,13 +280,17 @@ class LlamaServer:
             bufsize=1
         )
 
-        def filter_stderr():
-            for line in iter(self.process.stderr.readline, ''):
-                if not line.startswith(('srv ', 'slot ')) and 'log_server_r: request: GET /health' not in line:
-                    sys.stderr.write(line)
-                    sys.stderr.flush()
+        def filter_stderr(process_stderr):
+            try:
+                for line in iter(process_stderr.readline, ''):
+                    if not line.startswith(('srv ', 'slot ')) and 'log_server_r: request: GET /health' not in line:
+                        sys.stderr.write(line)
+                        sys.stderr.flush()
+            except (ValueError, IOError):
+                # Handle pipe closed exceptions
+                pass
 
-        threading.Thread(target=filter_stderr, daemon=True).start()
+        threading.Thread(target=filter_stderr, args=(self.process.stderr,), daemon=True).start()
 
         # Wait for server to be healthy
         health_url = f"http://localhost:{self.port}/health"