fix[hardware_predict]: Force x array to float64

Author: Djokzer

hls4ml/backends/vitis_accelerator/vitis_accelerator_backend.py

@@ -138,7 +138,7 @@ def hardware_predict(self, model, x, target="hw", debug=False, profilingRepeat=-
         elif method == "lib":
             """Run the hardware prediction using a shared library."""
             # Set array to contiguous memory layout
-            X_test = np.ascontiguousarray(x)
+            X_test = np.ascontiguousarray(x, dtype=np.float64)
 
             # Create prediction array
             config = VitisAcceleratorConfig(model.config)