Update app.py

app.py

@@ -23,17 +23,17 @@ regular_servers = [
     "https://rajhuggingface4253-compressor3pro.hf.space"
 ]
 
-# OPTIMIZED: Simpler prompts for faster responses
+# REVERTED: Use the EXACT format that worked for Qwen 2
 models_to_warm = [
     {
         "name": "Qwen 1",
         "url": "https://rajhuggingface4253-qwen.hf.space",
         "endpoint": "/chat",
         "type": "streaming_chat",
-        "timeout": 45.0,  # Increased timeout
+        "timeout": 45.0,
         "payload": {
-            "prompt": "Hello, respond with OK",
-            "max_new_tokens": 10,  # Shorter response
+            "prompt": "--- HISTORY START ---\nUser: Say 'ready' if you're working\n--- HISTORY END ---\n\nUser's latest message: \"Say 'ready' if you're working\"",
+            "max_new_tokens": 50,
             "temperature": 0.1,
             "enable_code_execution": False,
             "enable_web_search": False
@@ -46,8 +46,8 @@ models_to_warm = [
         "type": "streaming_chat",
         "timeout": 45.0,
         "payload": {
-            "prompt": "Hello, respond with OK",
-            "max_new_tokens": 10,
+            "prompt": "--- HISTORY START ---\nUser: Say 'ready' if you're working\n--- HISTORY END ---\n\nUser's latest message: \"Say 'ready' if you're working\"",
+            "max_new_tokens": 50,
             "temperature": 0.1,
             "enable_code_execution": False,
             "enable_web_search": False
@@ -60,8 +60,8 @@ models_to_warm = [
         "type": "streaming_chat",
         "timeout": 45.0,
         "payload": {
-            "prompt": "Hello, respond with OK",
-            "max_new_tokens": 10,
+            "prompt": "--- HISTORY START ---\nUser: Say 'ready' if you're working\n--- HISTORY END ---\n\nUser's latest message: \"Say 'ready' if you're working\"",
+            "max_new_tokens": 50,
             "temperature": 0.1,
             "enable_code_execution": False,
             "enable_web_search": False
@@ -105,7 +105,7 @@ async def ping_server(url: str) -> Dict:
         }
 
 async def warmup_chat_model(model_config: Dict) -> Dict:
-    """Warm up chat models with optimized approach"""
+    """Warm up chat models with PROPER error handling"""
     try:
         start_time = time.time()
         async with httpx.AsyncClient(timeout=model_config.get('timeout', 45.0)) as client:
@@ -153,11 +153,11 @@ async def warmup_chat_model(model_config: Dict) -> Dict:
                         'timestamp': datetime.now().isoformat()
                     }
             else:
-                # Try to get error details
+                # Try to get error details for debugging
                 error_detail = ""
                 try:
                     error_text = await response.text()
-                    error_detail = f" - {error_text[:100]}"
+                    error_detail = f" - {error_text[:200]}"
                 except:
                     pass