Spaces:
Paused
Paused
File size: 2,414 Bytes
5e75742 228558e 5e75742 65e0560 5e75742 |
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 |
import streamlit as st
import csv
import chromadb
from chromadb.utils import embedding_functions
from transformers import AutoTokenizer, AutoModelForSeq2SeqLM
from transformers import pipeline
from langchain.llms import HuggingFacePipeline
# Instantiate chromadb instance. Data is stored on disk (a folder named 'my_vectordb' will be created in the same folder as this file).
chroma_client = chromadb.PersistentClient(path="vector_db")
# Select the embedding model to use.
# List of model names can be found here https://www.sbert.net/docs/pretrained_models.html
sentence_transformer_ef = embedding_functions.SentenceTransformerEmbeddingFunction(model_name="all-mpnet-base-v2")
# Use this to delete the database
# chroma_client.delete_collection(name="my_collection")
# Create the collection, aka vector database. Or, if database already exist, then use it. Specify the model that we want to use to do the embedding.
collection = chroma_client.get_or_create_collection(name="my_collection", embedding_function=sentence_transformer_ef)
# Add the data to the collection
# collection.add(
# documents=documents,
# metadatas=metadatas,
# ids=ids
# )
# Streamlit app layout
st.title("ChromaDB and HuggingFace Pipeline Integration")
query = st.text_input("Enter your query:", value="director")
if st.button("Search"):
results = collection.query(
query_texts=[query],
n_results=1,
include=['documents', 'distances', 'metadatas']
)
st.write("Query Results:")
st.write(results['metadatas'])
if results['documents']:
context = results['documents']
st.write("Context:")
st.write(context)
tokenizer = AutoTokenizer.from_pretrained("MBZUAI/LaMini-T5-738M")
model = AutoModelForSeq2SeqLM.from_pretrained("MBZUAI/LaMini-T5-738M")
pipe = pipeline(
"text2text-generation",
model=model,
tokenizer=tokenizer,
max_length=512
)
local_llm = HuggingFacePipeline(pipeline=pipe)
l = f"""
Use the following pieces of context to answer the question at the end. If you don't know the answer, just say that you don't know, don't try to make up an answer.
{context}
Question: {query}
Helpful Answer:
"""
answer = local_llm(l)
st.write("Answer:")
st.write(answer)
|