From 79b88f8ddcc22b098cfefb7448ee189fe29b6c86 Mon Sep 17 00:00:00 2001
From: Justine Tunney <jtunney@mozilla.com>
Date: Tue, 23 Jan 2024 01:43:15 -0800
Subject: [PATCH] Sanitize -ngl flag on Apple Metal

Passing `-ngl 35` (as shown by examples) to Apple Metal causes weird
output. This should be `-ngl 1` since it enables full GPU offloading
---
 llamafile/gpu.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/llamafile/gpu.c b/llamafile/gpu.c
index 2a1bfb5e05..440a5b8e85 100644
--- a/llamafile/gpu.c
+++ b/llamafile/gpu.c
@@ -70,7 +70,7 @@ int llamafile_gpu_layers(int n_gpu_layers) {
     }
 
     // Apple Metal is safe enough to enable by default.
-    if (n_gpu_layers == -1 && ggml_metal_supported()) {
+    if ((n_gpu_layers < 0 || n_gpu_layers > 1) && ggml_metal_supported()) {
         n_gpu_layers = 1;
     }