Update app.py

app.py

@@ -1,5 +1,6 @@
 import gradio as gr
 from transformers import AutoModelForCausalLM, AutoTokenizer, AutoConfig, TextStreamer
+from transformers import Qwen2VLForConditionalGeneration, AutoTokenizer, AutoProcessor
 import torch
 from PIL import Image
 import re 
@@ -27,13 +28,13 @@ import tempfile
 from transformers import AutoTokenizer
 from transformers.generation import GenerationConfig
 
-DEFAULT_CKPT_PATH = 'Qwen/Qwen2-VL-7B-Instruct-GPTQ-Int4'
-device_map = "cpu"
+DEFAULT_CKPT_PATH = 'Qwen/Qwen2-VL-2B-Instruct-GPTQ-Int4'
+device_map = "auto"
 
 tokenizer = AutoTokenizer.from_pretrained(
         DEFAULT_CKPT_PATH, trust_remote_code=True,
     )
-model = AutoModelForCausalLM.from_pretrained(
+model = Qwen2VLForConditionalGeneration.from_pretrained(
         DEFAULT_CKPT_PATH,
         device_map=device_map,
         trust_remote_code=True,