Update app.py

app.py

@@ -13,7 +13,7 @@ footer {
 ACCESS_TOKEN = os.getenv("HF_TOKEN")
 
 client = OpenAI(
-    base_url="https://api-inference.huggingface.co/v1/",
+    base_url="https://api.deepinfra.com/v1/openai/chat/completions",
     api_key=ACCESS_TOKEN,
 )
 
@@ -53,7 +53,7 @@ def respond(
     response = ""
     
     for message in client.chat.completions.create(
-        model="meta-llama/Meta-Llama-3.1-8B-Instruct",
+        model="mistralai/Mistral-Small-24B-Instruct-2501",
         max_tokens=max_tokens,
         stream=True,
         temperature=temperature,