update

app.py

@@ -12,8 +12,8 @@ def chat_with_moi(message, history, temperature, top_p, beta):
     # Set the MIXINPUTS_BETA env var *per request*
     os.environ["MIXINPUTS_BETA"] = str(beta)
 
-    setup_mixinputs()
-    launch_vllm_server(beta=beta)
+    # setup_mixinputs()
+    # launch_vllm_server(beta=beta)
 
     payload = {
         "model": "Qwen/QwQ-32B",  # match what your vLLM server expects

server.py

@@ -7,6 +7,7 @@ def setup_mixinputs():
     # Step 1: Run mixinputs setup
     subprocess.run(["mixinputs", "setup"], check=True)
 
+@spaces.GPU(duration=240)
 def launch_vllm_server(beta=1.0):
     # Step 2: Set environment variables
     env = os.environ.copy()
@@ -24,11 +25,11 @@ def launch_vllm_server(beta=1.0):
     ]
     subprocess.run(cmd, env=env)
 
-# # Step 1: Setup
-# setup_mixinputs()
+# Step 1: Setup
+setup_mixinputs()
 
-# # Step 2: Launch vLLM server in background
-# threading.Thread(target=launch_vllm_server, daemon=True).start()
+# Step 2: Launch vLLM server in background
+threading.Thread(target=launch_vllm_server, daemon=True).start()
 
-# # Step 3: Give time for server to initialize
-# time.sleep(20)
+# Step 3: Give time for server to initialize
+time.sleep(20)