Update app.py

app.py

@@ -47,7 +47,7 @@ vision_model = AutoModelForCausalLM.from_pretrained(
 vision_processor = AutoProcessor.from_pretrained(VISION_MODEL_ID, trust_remote_code=True)
 
 # Helper functions
-@spaces.GPU  # Add this decorator
+@spaces.GPU
 def stream_text_chat(message, history, system_prompt, temperature=0.8, max_new_tokens=1024, top_p=1.0, top_k=20):
     conversation = [{"role": "system", "content": system_prompt}]
     for prompt, answer in history:
@@ -78,7 +78,7 @@ def stream_text_chat(message, history, system_prompt, temperature=0.8, max_new_t
     buffer = ""
     for new_text in streamer:
         buffer += new_text
-        yield buffer
+        yield history + [[message, buffer]]
 
 @spaces.GPU  # Add this decorator
 def process_vision_query(image, text_input):
@@ -97,6 +97,7 @@ def process_vision_query(image, text_input):
     response = vision_processor.batch_decode(generate_ids, skip_special_tokens=True, clean_up_tokenization_spaces=False)[0]
     return response
 
+# Update the Gradio interface part
 with gr.Blocks() as demo:
     gr.HTML(TITLE)
     gr.Markdown(DESCRIPTION)