handle output properly

app.py

@@ -2,15 +2,18 @@ import gradio as gr
 from evaluation_logic import run_evaluation, AVAILABLE_PROMPT_FORMATS
 
 def gradio_run_evaluation(model_name, prompt_format):
-    return run_evaluation(model_name, prompt_format)
+    output = []
+    for result in run_evaluation(model_name, prompt_format):
+        output.append(result)
+        yield "\n".join(output)
 
 with gr.Blocks() as demo:
     gr.Markdown("# DuckDB SQL Evaluation App")
 
     model_name = gr.Textbox(label="Model Name (e.g., qwen/qwen-2.5-72b-instruct)")
     prompt_format = gr.Dropdown(
-        label="Prompt Format", 
-        choices=AVAILABLE_PROMPT_FORMATS, 
+        label="Prompt Format",
+        choices=AVAILABLE_PROMPT_FORMATS,
         value="duckdbinstgraniteshort"
     )
     start_btn = gr.Button("Start Evaluation")