changed logic of source recovery + tune prompt to increase load for answer

Update app.py

change init model to use cpu

Update README.md

add info on the application startup to catch errors (#7)

avoid installing librairies for gpu (#6)

Update requirements.txt

Update requirements.txt

cahnage machine to cuda

Update requirements.txt

Update requirements.txt

Avoid PyTorch GPU build with full CUDA stack (#5)

Update requirements.txt

fixed embeddings loading

Nee to make app public to be used by third party c.a. github io (#4)

must include references when using retreiver (#3)

fixed bug retreiver name

removed magic methods

bmabla

fixed imports

skip import restrict usage

added reqs

load faiss from azure blb storage

load vector store from azure blob storage

load vector store from azure blob storage

Update requirements.txt

Update requirements.txt

Resolve merge: remove ENSCM_2013_CAZELLES.pdf

added usage restriction

github action on push on prod_ready branch

Revert "find a better way to include safeguard"

encryption of data added

added reqs and tests for cd/ci

dev-retreiver --> include git project s readme

find a better way to include safeguard

creation of vector_store

working app with LLM

langchain.chat_models.init_chat_model used -> need tracking (#2)

Update app.py (#1)

Update README.md

initial commit

Duplicate from gradio-templates/chatbot