Update app.py

app.py

@@ -58,7 +58,6 @@ class StarlingBot:
             gc.collect()
             torch.cuda.empty_cache()
 
-starling_bot = StarlingBot()
 examples = [
     [
         "The following dialogue is a conversation between Emmanuel Macron and Elon Musk:",  # user_message
@@ -69,21 +68,37 @@ examples = [
         1.9,  # repetition_penalty
     ]
 ]
+# Initialize StarlingBot
+starling_bot = StarlingBot()
+
+def gradio_starling(user_message, assistant_message, system_message, do_sample, temperature, max_new_tokens, top_p, repetition_penalty):
+    response = starling_bot.predict(user_message, assistant_message, system_message, do_sample, temperature, max_new_tokens, top_p, repetition_penalty)
+    return response
+
+with gr.Blocks(theme="ParityError/Anime") as demo:
+    gr.Markdown(title)  
+    gr.Markdown(description)  
+    with gr.Row():
+        system_message = gr.Textbox(label="Optional💫🌠Starling System Message", lines=2)
+        assistant_message = gr.Textbox(label="💫🌠Starling Assistant Message", lines=2)
+        user_message = gr.Textbox(label="Your Message", lines=3)
+    with gr.Row():
+        do_sample = gr.Checkbox(label="Advanced", value=True)
+    
+    with gr.Accordion("Advanced Settings", open=lambda do_sample: do_sample):
+        with gr.Row():
+            temperature = gr.Slider(label="Temperature", value=0.7, minimum=0.05, maximum=1.0, step=0.05)
+            max_new_tokens = gr.Slider(label="Max new tokens", value=100, minimum=25, maximum=256, step=1)
+            top_p = gr.Slider(label="Top-p (nucleus sampling)", value=0.90, minimum=0.01, maximum=0.99, step=0.05)
+            repetition_penalty = gr.Slider(label="Repetition penalty", value=1.9, minimum=1.0, maximum=2.0, step=0.05)
+
+    submit_button = gr.Button("Submit")
+    output_text = gr.Textbox(label="💫🌠Starling Response")
+
+    submit_button.click(
+        gradio_starling,
+        inputs=[user_message, assistant_message, system_message, do_sample, temperature, max_new_tokens, top_p, repetition_penalty],
+        outputs=output_text
+    )
 
-iface = gr.Interface(
-    fn=starling_bot.predict,
-    inputs=[
-        gr.Textbox(label="🌟🤩User Message", type="text", lines=5),
-        gr.Textbox(label="💫🌠Starling Assistant Message or Instructions ", lines=2),
-        gr.Textbox(label="💫🌠Starling System Prompt or Instruction", lines=2),
-        gr.Checkbox(label="Advanced", value=False),
-        gr.Slider(label="Temperature", value=0.7, minimum=0.05, maximum=1.0, step=0.05),
-        gr.Slider(label="Max new tokens", value=100, minimum=25, maximum=256, step=1),
-        gr.Slider(label="Top-p (nucleus sampling)", value=0.90, minimum=0.01, maximum=0.99, step=0.05),
-        gr.Slider(label="Repetition penalty", value=1.9, minimum=1.0, maximum=2.0, step=0.05)
-    ],
-    outputs="text",
-    gr.Markdown(title), 
-    gr.Markdown(description),
-    theme="ParityError/Anime"
-)
+demo.launch()