phi4

main.py

@@ -1,77 +1,71 @@
-import requests
+import streamlit as st
+import yaml
+from transformers import AutoTokenizer, AutoModelForCausalLM
 import torch
-import os
-import io
-from PIL import Image
-import soundfile as sf
-from transformers import AutoModelForCausalLM, AutoProcessor, GenerationConfig
-from urllib.request import urlopen
 
+class ModelManager:
+    def __init__(self, model_name="microsoft/Phi-4-mini-instruct"):
+        # Diccionario de modelos: usa los identificadores de Hugging Face
+        self.models = {
+            "microsoft/Phi-4-mini-instruct": "microsoft/Phi-4-mini-instruct",
+            "meta-llama/Llama-3.3-70B-Instruct": "meta-llama/Llama-3.3-70B-Instruct"
+        }
+        self.current_model_name = model_name
+        self.tokenizer = None
+        self.model = None
+        self.load_model(model_name)
+    
+    def load_model(self, model_name):
+        self.current_model_name = model_name
+        model_path = self.models[model_name]
+        st.info(f"Cargando modelo: {model_name} ...")
+        self.tokenizer = AutoTokenizer.from_pretrained(model_path)
+        self.model = AutoModelForCausalLM.from_pretrained(model_path)
+    
+    def generate(self, prompt, max_length=50, temperature=0.7):
+        inputs = self.tokenizer(prompt, return_tensors="pt")
+        outputs = self.model.generate(inputs["input_ids"], max_length=max_length, temperature=temperature)
+        return self.tokenizer.decode(outputs[0], skip_special_tokens=True)
+    
+    def switch_model(self, model_name):
+        if model_name in self.models:
+            self.load_model(model_name)
+        else:
+            raise ValueError(f"El modelo {model_name} no está disponible.")
 
-# Define model path
-model_path = "microsoft/Phi-4-multimodal-instruct"
+@st.cache_data
+def load_prompts():
+    with open("prompt.yml", "r", encoding="utf-8") as f:
+        prompts = yaml.safe_load(f)
+    return prompts
 
-# Load model and processor
-processor = AutoProcessor.from_pretrained(model_path, trust_remote_code=True)
-model = AutoModelForCausalLM.from_pretrained(
-    model_path, 
-    device_map="cuda", 
-    torch_dtype="auto", 
-    trust_remote_code=True,
-    # if you do not use Ampere or later GPUs, change attention to "eager"
-    _attn_implementation='flash_attention_2',
-).cuda()
+def main():
+    st.title("Switcher de Modelos de Transformers")
+    
+    # Cargar configuración de prompts
+    prompts_config = load_prompts()
+    
+    # Selección de modelo desde la barra lateral
+    st.sidebar.title("Selección de Modelo")
+    model_choice = st.sidebar.selectbox("Selecciona un modelo", list(prompts_config.keys()))
+    
+    # Instanciar el manejador de modelos
+    model_manager = ModelManager(model_name=model_choice)
+    
+    # Obtener el prompt de estilo para el modelo seleccionado
+    style_prompt = prompts_config.get(model_choice, prompts_config.get("default_prompt", ""))
+    
+    st.write(f"**Modelo en uso:** {model_choice}")
+    
+    # Área de texto para ingresar el prompt, iniciando con el estilo predefinido
+    user_prompt = st.text_area("Ingresa tu prompt:", value=style_prompt)
+    
+    max_length = st.slider("Longitud máxima", min_value=10, max_value=200, value=50)
+    temperature = st.slider("Temperatura", min_value=0.1, max_value=1.0, value=0.7)
+    
+    if st.button("Generar respuesta"):
+        result = model_manager.generate(user_prompt, max_length=max_length, temperature=temperature)
+        st.text_area("Salida", value=result, height=200)
 
-# Load generation config
-generation_config = GenerationConfig.from_pretrained(model_path)
-
-# Define prompt structure
-user_prompt = '<|user|>'
-assistant_prompt = '<|assistant|>'
-prompt_suffix = '<|end|>'
-
-# Part 1: Image Processing
-print("\n--- IMAGE PROCESSING ---")
-image_url = 'https://www.ilankelman.org/stopsigns/australia.jpg'
-prompt = f'{user_prompt}<|image_1|>What is shown in this image?{prompt_suffix}{assistant_prompt}'
-print(f'>>> Prompt\n{prompt}')
-
-# Download and open image
-image = Image.open(requests.get(image_url, stream=True).raw)
-inputs = processor(text=prompt, images=image, return_tensors='pt').to('cuda:0')
-
-# Generate response
-generate_ids = model.generate(
-    **inputs,
-    max_new_tokens=1000,
-    generation_config=generation_config,
-)
-generate_ids = generate_ids[:, inputs['input_ids'].shape[1]:]
-response = processor.batch_decode(
-    generate_ids, skip_special_tokens=True, clean_up_tokenization_spaces=False
-)[0]
-print(f'>>> Response\n{response}')
-
-# Part 2: Audio Processing
-print("\n--- AUDIO PROCESSING ---")
-audio_url = "https://upload.wikimedia.org/wikipedia/commons/b/b0/Barbara_Sahakian_BBC_Radio4_The_Life_Scientific_29_May_2012_b01j5j24.flac"
-speech_prompt = "Transcribe the audio to text, and then translate the audio to French. Use <sep> as a separator between the original transcript and the translation."
-prompt = f'{user_prompt}<|audio_1|>{speech_prompt}{prompt_suffix}{assistant_prompt}'
-print(f'>>> Prompt\n{prompt}')
-
-# Downlowd and open audio file
-audio, samplerate = sf.read(io.BytesIO(urlopen(audio_url).read()))
-
-# Process with the model
-inputs = processor(text=prompt, audios=[(audio, samplerate)], return_tensors='pt').to('cuda:0')
-
-generate_ids = model.generate(
-    **inputs,
-    max_new_tokens=1000,
-    generation_config=generation_config,
-)
-generate_ids = generate_ids[:, inputs['input_ids'].shape[1]:]
-response = processor.batch_decode(
-    generate_ids, skip_special_tokens=True, clean_up_tokenization_spaces=False
-)[0]
-print(f'>>> Response\n{response}')
+if __name__ == "__main__":
+    main()

prompts.yml

@@ -0,0 +1,3 @@
+default_prompt: "Eres un asistente de inteligencia artificial. Responde de manera clara y concisa."
+phi4-mini: "Estás usando el modelo phi4-mini. Proporciona respuestas técnicas y resumidas."
+llama: "Estás utilizando el modelo LLaMA. Ofrece explicaciones detalladas y en profundidad."

requirements.txt

@@ -1,10 +1,4 @@
-flash_attn==2.7.4.post1
-torch==2.6.0
-transformers==4.48.2
-accelerate==1.3.0
-soundfile==0.13.1
-pillow==11.1.0
-scipy==1.15.2
-torchvision==0.21.0
-backoff==2.2.1
-peft==0.13.2
+streamlit
+PyYAML
+transformers
+torch