From 605c6d37091e5a16e2a715086128ce65802dad58 Mon Sep 17 00:00:00 2001
From: ginnoir <ginnoir@users.noreply.github.com>
Date: Fri, 26 Jun 2026 16:31:01 -0500
Subject: [PATCH] fix(llm): use --flash-attn on (this llama.cpp build requires
 explicit value)

The server-cuda image parses -fa as --flash-attn [on|off|auto], so a bare -fa
swallowed the following --cache-type-k as its value and crash-looped.

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
---
 stacks/llm/docker-compose.yml | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

diff --git a/stacks/llm/docker-compose.yml b/stacks/llm/docker-compose.yml
index f1123e5..6b17f6d 100644
--- a/stacks/llm/docker-compose.yml
+++ b/stacks/llm/docker-compose.yml
@@ -36,7 +36,8 @@ services:
       - "99"
       - "--ctx-size"
       - "32768"
-      - "-fa"
+      - "--flash-attn"
+      - "on"
       - "--cache-type-k"
       - "q8_0"
       - "--cache-type-v"