Update app.py

app.py

@@ -50,7 +50,8 @@ for file in files:
 model = Qwen2VLForConditionalGeneration.from_pretrained(
     destination_folder,
     torch_dtype=torch.bfloat16,
-    device_map="cpu",
+    # device_map="cpu",
+    attn_implementation = "flash_attention_2",
 )
 
 # Load the processor