Verbessung gemäss ChatGPT

composte-hermes.yml

@@ -21,19 +21,14 @@ services:
               count: 2
               capabilities: [gpu]
     command:
-      - --model=Qwen/Qwen3-14B-AWQ
-      - --port=8005
-      - --quantization=awq
-      - --tensor-parallel-size=2
-      - --gpu-memory-utilization=0.82
-      - --max-model-len=65536
-      - --hf-overrides={"rope_scaling":{"rope_type":"yarn","factor":2.0,"original_max_position_embeddings":32768}}
-      - --max-num-seqs=16
-      - --kv-cache-dtype=fp8
-      - --served-model-name=qwen3-14b-awq
-      - --enable-auto-tool-choice
-      - --tool-call-parser=hermes
-      - --reasoning-parser=deepseek_r1
+       - --model=Qwen/Qwen3-14B-AWQ
+       - --tensor-parallel-size=2
+       - --gpu-memory-utilization=0.90
+       - --max-model-len=65536
+       - --enable-prefix-caching
+       - --enable-auto-tool-choice
+       - --tool-call-parser=hermes
+       - --max-num-seqs=4
     healthcheck:
       test: ["CMD", "curl", "-f", "http://localhost:8005/health"]
       interval: 30s