hi i am trying to work on the chatbot and i have the answer generated something like this
i have tried to pickle the embedding and the code is as follows:
file_path = "vectorStore.pkl"
pdf_file_path = "legal_women.pdf"
embedding_file_path = "embeddings.pkl"
# Load embeddings
if os.path.exists(embedding_file_path):
with open(embedding_file_path, "rb") as f:
embeddings = pickle.load(f)
else:
embeddings = HuggingFaceEmbeddings()
# Pickle embeddings
with open(embedding_file_path, "wb") as f:
pickle.dump(embeddings, f)
# Text splitter
text_splitter = CharacterTextSplitter(
separator="\n",
chunk_size=1000,
chunk_overlap=200
)
# Load PDF and split into documents
loader = PDFPlumberLoader(pdf_file_path)
pages = loader.load()
docs = text_splitter.split_documents(pages)
db = Chroma.from_documents(docs, embeddings)
# Pickle the extracted documents
with open(file_path, "wb") as f:
pickle.dump(docs, f)
prompt = hub.pull("rlm/rag-prompt", api_url="https://api.hub.langchain.com")
def model(user_query, max_length, temp):
if os.path.exists(file_path):
with open(file_path, "rb") as f:
vectorstore = pickle.load(f)
repo_id = 'mistralai/Mistral-7B-Instruct-v0.2'
llm = HuggingFaceHub(
repo_id=repo_id, model_kwargs={"max_length": max_length, "temperature": temp})
qa = RetrievalQA.from_chain_type(llm=llm,
chain_type="stuff",
retriever=db.as_retriever(k=2),
return_source_documents=True,
verbose=True,
chain_type_kwargs={"prompt": prompt})
return qa(user_query)["result"]
can u help me solve the correct way of generation of answer.
