How to cast the INT64 weights down to INT32

[Neronjust2017]

Hi, I convert my keras model to onnx using keras2onnx. However, when I try to convert the onnx file to TensorRT file, I encountered the following warnning "Your ONNX model has been generated with INT64 weights, while TensorRT does not natively support INT64. Attempting to cast down to INT32". How can I get a onnx model converted from keras with INT32 weights? Thanks!