Update app.py

app.py

@@ -32,6 +32,7 @@ def load_model():
         attn_implementation="sdpa",  # SDPA is fine on CPU; avoid flash-attn on CPU
         torch_dtype=torch.float32,   # Safer default for CPU
         device_map="cpu"             # Ensure CPU execution
+        quantization_config=None,
     )
     model.eval()