From 63ed4bba6e98265ee70a15df67b237897e392c20 Mon Sep 17 00:00:00 2001
From: Marc Blatter <marc.blatter@outlook.com>
Date: Sun, 26 Jul 2026 12:46:47 +0000
Subject: [PATCH] Korrektur durch ChatGPT

---
 composte-hermes.yml | 6 ++++--
 1 file changed, 4 insertions(+), 2 deletions(-)

diff --git a/composte-hermes.yml b/composte-hermes.yml
index c68ca98..931f843 100644
--- a/composte-hermes.yml
+++ b/composte-hermes.yml
@@ -24,13 +24,15 @@ services:
     command:
        - --model=Qwen/Qwen3-14B-AWQ
        - --tensor-parallel-size=2
-       - --gpu-memory-utilization=0.90
+       - --gpu-memory-utilization=0.92
        - --max-model-len=65536
-       - --hf-overrides={"rope_scaling":{"rope_type":"yarn","factor":2.0,"original_max_position_embeddings":32768}}
+       - --hf-overrides={"rope_scaling":{"rope_type":"yarn","factor":2.0,"original_max_position_embeddings":40960}}
        - --enable-prefix-caching
        - --enable-auto-tool-choice
        - --tool-call-parser=hermes
        - --max-num-seqs=4
+       - --generation-config=vllm
+       - --disable-custom-all-reduce
     healthcheck:
       test: ["CMD", "curl", "-f", "http://localhost:8005/health"]
       interval: 30s