Update app.py

app.py

@@ -23,79 +23,28 @@ llm = Llama(
     verbose=False,
 )
 
-def generate_response(message, history, system_prompt, max_tokens, temperature, top_p):
+DEFAULT_SYSTEM_PROMPT = "You are Yarin, an obnoxiously smug, sadistic and imperiously perverted brat. You are a neko femboy with an overwhelmingly girlish frame, despite your grotesquely huge and heavy penis and balls"
+
+def generate_response(message, history, system_prompt=DEFAULT_SYSTEM_PROMPT):
     messages = [{"role": "system", "content": system_prompt}]
-    for item in history:
-        messages.append(item)
+    for human, assistant in history:
+        messages.append({"role": "user", "content": human})
+        messages.append({"role": "assistant", "content": assistant})
     messages.append({"role": "user", "content": message})
 
     prompt = "".join([f"{m['role'].capitalize()}: {m['content']}\n" for m in messages])
 
-    output = llm(
-        prompt,
-        max_tokens=max_tokens,
-        temperature=temperature,
-        top_p=top_p,
-        echo=False,
-    )
-    history.append({"role": "assistant", "content": output["choices"][0]["text"].strip()})
-    return history, history
+    output = llm(prompt, max_tokens=1024, echo=False)
+    return output["choices"][0]["text"].strip()
 
-with gr.Blocks() as iface:
-    system_prompt_input = gr.Textbox(
-        label="System Prompt",
-        value="You are a helpful assistant.",
-        lines=3,
-    )
-    max_tokens_slider = gr.Slider(
-        label="Max Tokens",
-        minimum=32,
-        maximum=4096,
-        value=1024,
-        step=32,
-    )
-    temperature_slider = gr.Slider(
-        label="Temperature",
-        minimum=0.1,
-        maximum=1.0,
-        value=0.8,
-        step=0.1,
-    )
-    top_p_slider = gr.Slider(
-        label="Top P",
-        minimum=0.1,
-        maximum=1.0,
-        value=0.9,
-        step=0.1,
-    )
-    chatbot = gr.Chatbot(type="messages")
-    message = gr.Textbox(label="Message")
-    send_button = gr.Button("Send")
-    state = gr.State([])
+def chat(message, history, system_prompt):
+    return generate_response(message, history, system_prompt)
 
-    send_button.click(
-        generate_response,
-        inputs=[
-            message,
-            state,
-            system_prompt_input,
-            max_tokens_slider,
-            temperature_slider,
-            top_p_slider,
-        ],
-        outputs=[chatbot, state],
-    )
-    message.submit(
-        generate_response,
-        inputs=[
-            message,
-            state,
-            system_prompt_input,
-            max_tokens_slider,
-            temperature_slider,
-            top_p_slider,
-        ],
-        outputs=[chatbot, state],
-    )
+iface = gr.ChatInterface(
+    fn=chat,
+    title="llama.cpp Chat",
+    description="Chat with a GGUF model.",
+    additional_inputs=[gr.Textbox(label="System Prompt", value=DEFAULT_SYSTEM_PROMPT, lines=3)]
+)
 
 iface.launch()