Update app.py

app.py

@@ -2,7 +2,7 @@ import os
 import torch
 from threading import Thread
 import gradio as gr
-from transformers import Qwen2_5_VLForConditionalGeneration, AutoProcessor,TextIteratorStreamer
+from transformers import Qwen2_5_VLForConditionalGeneration, AutoProcessor,TextIteratorStreamer,AutoTokenizer
 from qwen_vl_utils import process_vision_info
 
 # 3D mesh dependencies
@@ -35,7 +35,7 @@ model = Qwen2_5_VLForConditionalGeneration.from_pretrained(
     trust_remote_code=True
 )
 processor = AutoProcessor.from_pretrained(MODEL_DIR)
-
+terminators = [tokenizer.eos_token_id]
 
 def chat_qwen_vl(message: str, history: list, temperature: float = 0.1, max_new_tokens: int = 1024):
     # —— 原有多模态输入构造 —— #
@@ -62,7 +62,7 @@ def chat_qwen_vl(message: str, history: list, temperature: float = 0.1, max_new_
 
     # 2. 把 streamer 和生成参数一起传给 model.generate
     streamer = TextIteratorStreamer(
-        tokenizer,
+        processor,
         timeout=100.0,
         skip_prompt=True,
         skip_special_tokens=True