Update app.py

app.py

@@ -17,7 +17,7 @@ model = AutoModelForCausalLM.from_pretrained(
 model.eval()  # Set the model to evaluation mode
 
 # Initialize the inference client (if needed for other API-based tasks)
-# client = InferenceClient(token=access_token)
+client = InferenceClient(provider="together",token=access_token)
 
 def conversation_predict(input_text):
     """Generate a response for single-turn input using the model."""
@@ -54,6 +54,7 @@ def respond(
 
     # Stream response tokens from the chat completion API
     for message_chunk in client.chat_completion(
+        model = "google/gemma-2b-it"
         messages=messages,
         max_tokens=max_tokens,
         stream=True,