Update app.py

app.py

@@ -1,68 +1,15 @@
 #10
-import gradio as gr
-from huggingface_hub import InferenceClient
-import os
+from transformers import AutoTokenizer, AutoModelForSeq2SeqLM, pipeline
 
-# Suponiendo que 'token' es tu token de autenticación de Hugging Face
-token = os.getenv("Mitoken")
-if not token:
-    raise ValueError("Token de Hugging Face no encontrado.")
+# Cargar el tokenizer y el modelo
+tokenizer = AutoTokenizer.from_pretrained("PlanTL-GOB-ES/roberta-base-bne")
+model = AutoModelForSeq2SeqLM.from_pretrained("PlanTL-GOB-ES/roberta-base-bne")
 
-# Crear un único cliente de inferencia con el token de autenticación
-client = InferenceClient(token=token)
+# Inicializar la pipeline de generación de texto
+text_generator = pipeline("text-generation", model=model, tokenizer=tokenizer)
 
-# Lista de modelos
-models = [
-    "google/gemma-7b",
-    "google/gemma-7b-it",
-    "google/gemma-2b",
-    "google/gemma-2b-it"
-]
+# Generar una respuesta a una pregunta en español
+question = "¿Cuál es la capital de España?"  # Ejemplo de pregunta
+response = text_generator(question, max_length=50, do_sample=True)
 
-def chat_inf(prompt, history, memory, client_choice, seed, temp, tokens, top_p, rep_p, chat_mem, cust_p):
-    model = models[int(client_choice)]  # Obtener el ID del modelo basado en la elección del usuario
-    try:
-        parameters = {
-            "temperature": temp,
-            "max_new_tokens": tokens,
-            "top_p": top_p,
-            "repetition_penalty": rep_p,
-            "seed": seed,
-        }
-        formatted_prompt = f"{prompt}"
-        # Realiza la solicitud de inferencia especificando el model_id directamente aquí
-        response = client(inputs=formatted_prompt, parameters=parameters, model_id=model, wait_for_model=True)
-        
-        output = response[0]["generated_text"] if response else "No response."
-        new_history = history + [(prompt, output)]
-        return new_history, memory
-    except Exception as e:
-        error_message = f"An error occurred: {str(e)}"
-        new_history = history + [(prompt, error_message)]
-        return new_history, memory
-
-with gr.Blocks() as app:
-    history = gr.State(default=[])
-    memory = gr.State(default=[])
-    
-    with gr.Row():
-        inp = gr.Textbox(label="Prompt")
-        client_choice = gr.Dropdown(label="Choose Model", choices={name: i for i, name in enumerate(models)}, value=0)
-        temp = gr.Slider(label="Temperature", minimum=0.1, maximum=1.0, value=0.7)
-        tokens = gr.Slider(label="Max new tokens", minimum=1, maximum=512, value=100)
-        top_p = gr.Slider(label="Top-P", minimum=0.1, maximum=1.0, value=0.9)
-        rep_p = gr.Slider(label="Repetition Penalty", minimum=1.0, maximum=2.0, value=1.1)
-        seed = gr.Number(label="Seed", value=42)
-        chat_mem = gr.Slider(label="Chat Memory", minimum=1, maximum=10, value=3)
-        cust_p = gr.Textbox(label="Custom Prompt", value="<start_of_turn>user{prompt}<end_of_turn><start_of_turn>model", visible=False)
-        generate_button = gr.Button("Generate")
-        
-    chat = gr.Chatbot()
-    
-    generate_button.click(
-        fn=chat_inf,
-        inputs=[inp, history, memory, client_choice, seed, temp, tokens, top_p, rep_p, chat_mem, cust_p],
-        outputs=[chat, memory]
-    )
-
-app.launch()
+print(response[0]['generated_text'])