Update app.py

app.py

@@ -16,7 +16,6 @@ from numpy import ndarray
 from outlines import models
 from llama_cpp import Llama
 import hydralit_components as hc
-from transformers import AutoTokenizer
 from qdrant_client import QdrantClient
 from optimum_encoder import OptimumEncoder
 from huggingface_hub import snapshot_download
@@ -25,6 +24,7 @@ from fastembed import SparseEmbedding, SparseTextEmbedding
 from unstructured.nlp.tokenize import download_nltk_packages
 from scipy.sparse import csr_matrix, save_npz, load_npz, vstack
 from langchain_experimental.text_splitter import SemanticChunker
+from mistral_common.tokens.tokenizers.mistral import MistralTokenizer
 from langchain_community.document_loaders import WikipediaLoader, WebBaseLoader
 from qdrant_client.models import (
     NamedSparseVector,
@@ -147,6 +147,8 @@ def main(query: str, client: QdrantClient, collection_name: str, llm: Llama, den
 def load_models_and_documents():
     with st.spinner('Load models...'):
         model_path = snapshot_download(repo_id="Ichigo2899/mistralai-Mistral-Nemo-Instruct-2407-AWQ")
+
+        tokenizer = MistralTokenizer.from_file(f"{model_path}/tekken.json")
         
         llm = vllm.LLM(
             model_path,