Update app.py

app.py

@@ -10,13 +10,12 @@ import numpy as np
 import streamlit as st
 from typing import List
 from numpy import ndarray
-from qdrant_client import QdrantClient, models
-from llama_cpp import Llama, GGML_TYPE_I8
+from llama_cpp import Llama
 from optimum_encoder import OptimumEncoder
+from qdrant_client import QdrantClient, models
 from unstructured.partition.auto import partition
 from fastembed import SparseEmbedding, SparseTextEmbedding
 from unstructured.nlp.tokenize import download_nltk_packages
-from fastembed.sparse.splade_pp import supported_splade_models
 from scipy.sparse import csr_matrix, save_npz, load_npz, vstack
 from langchain_experimental.text_splitter import SemanticChunker
 from langchain_community.document_loaders import WikipediaLoader, WebBaseLoader
@@ -190,7 +189,8 @@ def load_models_and_documents():
         )
 
         dense_model = OptimumEncoder(
-            device="cuda"
+            device="cuda",
+            cache_dir=os.getenv('HF_HOME')
         )
         
         sparse_model = SparseTextEmbedding(