Spaces:
Paused
Paused
File size: 2,317 Bytes
236b637 |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 |
import streamlit as st
from llama_cpp import Llama
import os
from rag_model import RAGEngine
#from rag_model_optimise import RAGEngine
import logging
from huggingface_hub import hf_hub_download
import time
import os
os.environ["NLTK_DATA"] = "/home/appuser/nltk_data"
# Appliquer le patch avant tout import de llama_index
from patches.llama_patch import patch_llamaindex_nltk
patch_llamaindex_nltk()
logger = logging.getLogger("Streamlit")
logger.setLevel(logging.INFO)
handler = logging.StreamHandler()
formatter = logging.Formatter("[%(asctime)s] %(levelname)s - %(message)s")
handler.setFormatter(formatter)
logger.addHandler(handler)
ENV = os.getenv("ENV", "space")
logger.info(f"ENV :{ENV}")
#time.sleep(5)
if ENV == "local":
model_path = "chatbot-models/Nous-Hermes-2-Mistral-7B-DPO.Q4_K_M.gguf"
faiss_index_path="chatbot-models/vectordb_docling/index.faiss"
vectors_path="chatbot-models/vectordb_docling/chunks.pkl"
else:
# Télécharger le modèle GGUF
model_path = hf_hub_download(
repo_id="rkonan/chatbot-models",
filename="chatbot-models/Nous-Hermes-2-Mistral-7B-DPO.Q4_K_M.gguf",
repo_type="dataset"
)
# Télécharger les fichiers FAISS
faiss_index_path = hf_hub_download(
repo_id="rkonan/chatbot-models",
filename="chatbot-models/vectordb_docling/index.faiss",
repo_type="dataset"
)
vectors_path = hf_hub_download(
repo_id="rkonan/chatbot-models",
filename="chatbot-models/vectordb_docling/chunks.pkl",
repo_type="dataset"
)
st.set_page_config(page_title="Chatbot RAG local",page_icon="🤖")
@st.cache_resource
def load_rag_engine():
rag = RAGEngine(
model_path=model_path,
vector_path=vectors_path,
index_path=faiss_index_path,
model_threads=8 # ✅ plus rapide
)
# 🔥 Warmup pour éviter latence au 1er appel
rag.llm("Bonjour", max_tokens=1)
return rag
rag=load_rag_engine()
st.title("🤖 Chatbot LLM Local (CPU)")
user_input=st.text_area("Posez votre question :", height=100)
if st.button("Envoyer") and user_input.strip():
with st.spinner("Génération en cours..."):
response = rag.ask(user_input)
st.markdown("**Réponse :**")
st.success(response) |