update app

app.py

@@ -226,11 +226,11 @@ class RadioAnimated(gr.HTML):
 def apply_gpu_duration(val: str):
     return int(val)
 
-MODEL_ID_V = "prithivMLmods/Qwen2.5-VL-7B-Instruct-Unredacted-MAX-FP8"
+MODEL_ID_V = "nanonets/Nanonets-OCR2-3B"
 processor_v = AutoProcessor.from_pretrained(MODEL_ID_V, trust_remote_code=True)
 model_v = Qwen2_5_VLForConditionalGeneration.from_pretrained(
     MODEL_ID_V,
-   # attn_implementation="kernels-community/flash-attn2",
+    attn_implementation="kernels-community/flash-attn2",
     trust_remote_code=True,
     torch_dtype=torch.float16
 ).to(device).eval()