chore: test gpt-oss-120b, nemotron and qwen3

Author: ZengSpace

config/config.yaml

@@ -46,15 +46,15 @@ prompt_guard:
 # NOT supported: domain names (example.com), protocol prefixes (http://), paths (/api), ports in address (use 'port' field)
 vllm_endpoints:
   - name: "gpt-oss-120b-endpoint"
-    address: "172.28.0.1" # Docker gateway IP to access host machine
+    address: "192.168.48.1" # Host machine IP address
     port: 8001
     weight: 1
   - name: "qwen3-32b-endpoint"
-    address: "172.28.0.1" # Docker gateway IP to access host machine
+    address: "192.168.48.1" # Host machine IP address
     port: 8002
     weight: 1
   - name: "nemotron-super-endpoint"
-    address: "172.28.0.1" # Docker gateway IP to access host machine
+    address: "192.168.48.1" # Host machine IP address
     port: 8003
     weight: 1
 

deploy/docker-compose/docker-compose.yml

@@ -20,7 +20,7 @@ services:
       - HUGGINGFACE_HUB_CACHE=/root/.cache/huggingface
       - HF_HUB_ENABLE_HF_TRANSFER=1
     extra_hosts:
-      - "host.docker.internal:172.28.0.1"  # Allow container to reach host LLM endpoints
+      - "host.docker.internal:172.32.0.1"  # Allow container to reach host LLM endpoints
     networks:
       - semantic-network
     healthcheck:
@@ -43,7 +43,7 @@ services:
       - ./addons/envoy.yaml:/etc/envoy/envoy.yaml:ro,z
     command: ["/usr/local/bin/envoy", "-c", "/etc/envoy/envoy.yaml", "--component-log-level", "ext_proc:trace,router:trace,http:trace"]
     extra_hosts:
-      - "host.docker.internal:172.28.0.1"  # Allow container to reach host LLM endpoints
+      - "host.docker.internal:172.32.0.1"  # Allow container to reach host LLM endpoints
     depends_on:
       semantic-router:
         condition: service_healthy
@@ -67,7 +67,7 @@ services:
       - "8000:8000"
     networks:
       semantic-network:
-        ipv4_address: 172.28.0.10
+        ipv4_address: 172.32.0.10
     healthcheck:
       test: ["CMD", "curl", "-fsS", "http://localhost:8000/health"]
       interval: 10s
@@ -219,7 +219,7 @@ services:
       - hf-cache:/home/llmkatan/.cache/huggingface
     networks:
       semantic-network:
-        ipv4_address: 172.28.0.20
+        ipv4_address: 172.32.0.20
     command: ["llm-katan", "--model", "/app/models/Qwen/Qwen3-0.6B", "--served-model-name", "qwen3", "--host", "0.0.0.0", "--port", "8002"]
     healthcheck:
       test: ["CMD", "curl", "-fsS", "http://localhost:8002/health"]
@@ -278,7 +278,7 @@ networks:
     driver: bridge
     ipam:
       config:
-        - subnet: 172.28.0.0/16
+        - subnet: 172.32.0.0/16
 
 volumes:
   models-cache: