Update app.py

app.py

@@ -4,13 +4,12 @@ import spaces
 import gradio as gr
 from peft import PeftModel
 from transformers import AutoModelForCausalLM, AutoTokenizer, TextIteratorStreamer
-import torch
 import threading
 
-# Load the base model
+# Load the base model without quantization to avoid bitsandbytes issues
 base_model = AutoModelForCausalLM.from_pretrained(
     "unsloth/qwen2.5-3b-instruct-unsloth-bnb-4bit",
-    device_map="auto",
+    device_map="cpu",  # Force CPU to avoid bitsandbytes dependency
     torch_dtype="auto"
 )
 
@@ -57,4 +56,4 @@ demo = gr.Interface(
     allow_flagging="never"
 )
 
-demo.launch(share=True)
+demo.launch(share=True)