Update app.py

app.py

@@ -12,7 +12,10 @@ from huggingface_hub import snapshot_download
 MODEL_BASE_SINGLE = snapshot_download(repo_id="nvidia/audio-flamingo-3")
 MODEL_BASE_THINK = os.path.join(MODEL_BASE_SINGLE, 'stage35')
 
-model_single = llava.load(MODEL_BASE_SINGLE, model_base=None, devices=[0])
+# model_single = llava.load(MODEL_BASE_SINGLE, model_base=None, devices=[0])
+model_single = llava.load(MODEL_BASE_SINGLE, model_base=None)
+model_single = model_single.to("cuda")
+
 generation_config_single = model_single.default_generation_config
 
 model_think = PeftModel.from_pretrained(