Update app.py

app.py

@@ -24,7 +24,8 @@ tokenizer, llava_model, image_processor, context_len = load_pretrained_model(
 )
 
 # Define the prediction function
-def predict(image, text, history):
+@spaces.GPU
+def bot_streaming(image, text, history):
     output = chat_llava(
         args=None,
         image_file=image,
@@ -44,8 +45,8 @@ with gr.Blocks() as demo:
     text_input = gr.Textbox(label="Enter your message")
     submit_btn = gr.Button("Submit")
 
-    # Define interaction: when submit is clicked, call predict and update the chatbot
-    submit_btn.click(fn=predict, inputs=[image_input, text_input, chatbot], outputs=[chatbot, text_input])
+    # Define interaction: when submit is clicked, call bot_streaming and update the chatbot
+    submit_btn.click(fn=bot_streaming, inputs=[image_input, text_input, chatbot], outputs=[chatbot, text_input])
 
     # Add example inputs
     gr.Examples(