Update app.py

app.py

@@ -42,23 +42,17 @@ CHAT_TEMPLATE= """{% for message in messages %}
 {% endfor %}""" # Asegúrate de usar tu CHAT_TEMPLATE aquí
 tokenizer.chat_template = CHAT_TEMPLATE
 
-# Función para generar respuestas del modelo
-import gradio as gr
-
-# Asume que todas tus importaciones previas y configuraciones del modelo están aquí
-
-# Aquí deberías tener definida la función `generate_response` tal como la compartiste
+chat_history = []
 # Función para generar respuestas del modelo
 def generate_response(user_input, chat_history):
     # Preparar el input agregando el historial de chat
     chat_history.append({"content": user_input, "role": "user"})
-    print(chat_history)
     user_input  = tokenizer.apply_chat_template(chat_history, tokenize=False)
     
     input_tokens = tokenizer(user_input, return_tensors='pt', padding=True, truncation=True, max_length=1024).to(device)
     
     # Generar la respuesta
-    output_tokens = model_with_adapter.generate(**input_tokens, max_length=1024, pad_token_id=tokenizer.eos_token_id, top_k=50, top_p=0.95, temperature=0.7)
+    output_tokens = model_with_adapter.generate(**input_tokens, max_length=1024, pad_token_id=tokenizer.eos_token_id, top_k=50, do_sample=True, top_p=0.95, temperature=0.7)
     generated_text = tokenizer.decode(output_tokens[0], skip_special_tokens=True)
 
     last_us = generated_text.rfind("</user>") + len("</user>")
@@ -67,52 +61,11 @@ def generate_response(user_input, chat_history):
     chat_history.append({"content": generated_text, "role": "assistant"})
     return generated_text, chat_history
     
-History = List[Tuple[str, str]]
-
-def clear_session() -> History:
-    return '', []
-
-def modify_system_session(system: str) -> str:
-    if system is None or len(system) == 0:
-        system = default_system
-    return system, system, []
-
-def model_chat(query: Optional[str], history: Optional[History], system: str
-) -> Tuple[str, str, History]:
-    if query is None:
-        query = ''
-    if history is None:
-        history = []
-    if not history or history[-1][0] != query:  # Asegurar que no se repita la última pregunta
-        history.append((query, ''))
-    response, history = generate_response(query, history)  # Tu función de modelo
-    return '', response, history
-
-placeholder = "This is a simulated conversation with your model."
-
-with gr.Blocks() as demo:
-    gr.Markdown("<center><font size=6>Chatbot with Your Model</center>")
-
-    with gr.Row():
-        system_input = gr.Textbox(value=default_system, lines=1, label='System')
-        modify_system = gr.Button("🛠️ Set system prompt and clear the history")
-        system_state = gr.Textbox(value=default_system, visible=False)
-    
-    chatbot = gr.Chatbot(label='Chat with Your Model', placeholder=placeholder)
-    textbox = gr.Textbox(lines=2, label='Input')
-
-    with gr.Row():
-        clear_history = gr.Button("🧹 Clear history")
-        submit = gr.Button("🚀 Send")
+def response(user_input, chat_history):
+    response, chat_history = generate_response(user_input, chat_history)
+    print(chat_history)
+    return response
 
-    submit.click(model_chat,
-                 inputs=[textbox, chatbot, system_state],
-                 outputs=[textbox, chatbot, system_input])
-    clear_history.click(fn=clear_session,
-                        inputs=[],
-                        outputs=[textbox, chatbot])
-    modify_system.click(fn=modify_system_session,
-                        inputs=[system_input],
-                        outputs=[system_state, system_input, chatbot])
+iface = gr.ChatInterface(fn=response, inputs="text", outputs="text")
 
-demo.launch()
+iface.launch()