Switch to qwen3:1.7b for faster inference

docker-compose.full.yml

@@ -42,7 +42,7 @@ services:
       DATABASE_URL: postgresql://brain:${POSTGRES_PASSWORD:-braindb2026}@postgres:5432/second_brain
       OLLAMA_URL: ${OLLAMA_URL:-http://192.168.1.217:11434}
       EMBEDDING_MODEL: ${EMBEDDING_MODEL:-nomic-embed-text}
-      CHAT_MODEL: ${CHAT_MODEL:-qwen3:8b}
+      CHAT_MODEL: ${CHAT_MODEL:-qwen3:1.7b}
       LOG_LEVEL: ${LOG_LEVEL:-INFO}
       SEARCH_TOP_K: ${SEARCH_TOP_K:-10}
       SEARCH_THRESHOLD: ${SEARCH_THRESHOLD:-0.65}
@@ -90,7 +90,7 @@ services:
       DATABASE_URL: postgresql://brain:${POSTGRES_PASSWORD:-braindb2026}@postgres:5432/second_brain
       OLLAMA_URL: ${OLLAMA_URL:-http://192.168.1.217:11434}
       VAULT_PATH: /vault
-      CHAT_MODEL: ${CHAT_MODEL:-qwen3:8b}
+      CHAT_MODEL: ${CHAT_MODEL:-qwen3:1.7b}
     volumes:
       - /opt/second-brain-vault:/vault:ro
     depends_on: