Albert / app.py
Albert-lab's picture
Create app.py
5f35428 verified
Raw History Blame Contribute Delete
831 Bytes
from transformers import AutoTokenizer, AutoModelForCausalLM
import torch
import gradio as gr
# Chargement du modèle plus léger
model_name = "tiiuae/falcon-rw-1b"
tokenizer = AutoTokenizer.from_pretrained(model_name)
model = AutoModelForCausalLM.from_pretrained(model_name)
def albert_respond(prompt):
inputs = tokenizer(prompt, return_tensors="pt")
outputs = model.generate(**inputs, max_length=100, do_sample=True, temperature=0.7)
response = tokenizer.decode(outputs[0], skip_special_tokens=True)
return response[len(prompt):].strip()
# Interface Gradio
iface = gr.Interface(fn=albert_respond,
inputs="text",
outputs="text",
title="Albert v0.1",
description="Assistant personnel hébergé sur Hugging Face")
iface.launch()