fix: use quantization_config

app.py

@@ -1,5 +1,10 @@
 import torch
-from transformers import AutoModelForCausalLM, AutoTokenizer, TextIteratorStreamer
+from transformers import (
+    AutoModelForCausalLM,
+    AutoTokenizer,
+    TextIteratorStreamer,
+    BitsAndBytesConfig,
+)
 from threading import Thread
 
 import gradio as gr
@@ -16,8 +21,9 @@ except:
 
 MODEL_NAME = "hatakeyama-llm-team/Tanuki-8B-Instruct"
 
+quantization_config = BitsAndBytesConfig(load_in_8bit=True)
 model = AutoModelForCausalLM.from_pretrained(
-    MODEL_NAME, load_in_8bit=True, device_map="auto"
+    MODEL_NAME, quantization_config=quantization_config, device_map="auto"
 )
 tokenizer = AutoTokenizer.from_pretrained(MODEL_NAME)