Update vlm.py

vlm.py

@@ -5,6 +5,7 @@ Author: Didier Guillevic
 Date: 2025-03-16
 """
 
+import spaces
 from transformers import AutoProcessor, Gemma3ForConditionalGeneration
 from transformers import TextIteratorStreamer
 from threading import Thread
@@ -67,6 +68,7 @@ def build_messages(message: dict, history: list[tuple]):
 #
 # Streaming response
 #
+@spaces.GPU
 @torch.inference_mode()
 def stream_response(messages: list[dict]):
     """Stream the model's response to the chat interface.
@@ -101,6 +103,7 @@ def stream_response(messages: list[dict]):
 #
 # Response (non-streaming)
 #
+@spaces.GPU
 @torch.inference_mode()
 def get_response(messages: list[dict]):
     """Get the model's response.