feat(*):增加了一些模型

vllm-qwen25-72b-instruct-awq/.env

@@ -0,0 +1 @@
+# VARIABLE=value #comment

vllm-qwen25-72b-instruct-awq/compose.yaml

@@ -0,0 +1,30 @@
+version: "3.9"
+services:
+  vllm-Qwen2.5-72B-Instruct-AWQ:
+    image: docker.citory.tech/mirror/vllm/vllm-openai:latest
+    container_name: vllm-Qwen2.5-72B-Instruct-AWQ
+    ports:
+      - 10580:8080
+    tty: true
+    deploy:
+      resources:
+        reservations:
+          devices:
+            - driver: nvidia
+              count: 2
+              capabilities:
+                - gpu
+    ipc: host
+    runtime: nvidia
+    volumes:
+      - /home/deepgeek/data/data_local/server/vllm/models:/models
+    restart: always
+    command: --served-model-name Qwen2.5-72B-Instruct-AWQ --model /models/Qwen/Qwen2.5-72B-Instruct-AWQ
+      --trust-remote-code --host 0.0.0.0 --port 8080 --max-model-len 8192
+      --tensor-parallel-size 2 --gpu_memory_utilization 0.9 --enforce-eager
+      --dtype auto --swap-space 8 --enable-auto-tool-choice --tool-call-parser
+      llama3_json
+x-dockge:
+  urls:
+    - http://local.citory.tech:10580
+networks: {}