Update model_api.py

model_api.py

@@ -2,7 +2,7 @@ import requests
 import os
 import time
 
-API_URL = "https://api-inference.huggingface.co/models/mistralai/Mistral-7B-Instruct-v0.2"
+API_URL = "https://router.huggingface.co/hf-inference/models/mistralai/Mistral-7B-Instruct-v0.2"
 
 HF_TOKEN = os.getenv("HF_TOKEN")
 
@@ -23,9 +23,11 @@ def query_model(prompt):
         response = requests.post(API_URL, headers=headers, json=payload)
         result = response.json()
 
+        # Success response
         if isinstance(result, list):
             return result[0]["generated_text"]
 
+        # Model loading
         if "estimated_time" in result:
             time.sleep(result["estimated_time"])
         else: