fix: use AutoModelForImageTextToText, add spaces dep

app.py

@@ -6,7 +6,7 @@ import os
 import spaces
 import gradio as gr
 import torch
-from transformers import AutoModelForVision2Seq, AutoProcessor
+from transformers import AutoModelForImageTextToText, AutoProcessor
 from qwen_vl_utils import process_vision_info
 from PIL import Image
 
@@ -23,7 +23,7 @@ processor = AutoProcessor.from_pretrained(
     trust_remote_code=True,
 )
 
-model = AutoModelForVision2Seq.from_pretrained(
+model = AutoModelForImageTextToText.from_pretrained(
     MODEL_ID,
     torch_dtype=torch.bfloat16,
     device_map="auto",

requirements.txt

@@ -3,3 +3,4 @@ transformers>=4.55.0
 accelerate
 qwen_vl_utils
 Pillow
+spaces