Add Qwen3-VL vision runtime dependency

app.py

@@ -3,13 +3,13 @@ import threading
 import gradio as gr
 import spaces
 import torch
-from transformers import AutoModelForMultimodalLM, AutoProcessor
+from transformers import AutoProcessor, Qwen3VLForConditionalGeneration
 
 
 MODEL_ID = "ForeverBlue/Qwen3-VL-2B-GRACE-BF16"
 
 processor = AutoProcessor.from_pretrained(MODEL_ID)
-model = AutoModelForMultimodalLM.from_pretrained(
+model = Qwen3VLForConditionalGeneration.from_pretrained(
     MODEL_ID,
     dtype=torch.bfloat16,
     low_cpu_mem_usage=True,
@@ -99,4 +99,3 @@ with gr.Blocks(title="GRACE-VLM Demo") as demo:
 
 if __name__ == "__main__":
     demo.queue(default_concurrency_limit=1).launch()
-

requirements.txt

@@ -1,4 +1,5 @@
 accelerate==1.13.0
 pillow>=11,<12
 torch==2.8.0
+torchvision==0.23.0
 transformers==5.9.0