How to convert a multi-input tensorflow model to tensorRT?

When I use uff.register_input() to register multiple inputs.
Something like:

parser = uffparser.create_uff_parser()
parser.register_input("image", (3,64,192),0)
parser.register_input("next_image", (3,64,192),1)

Then error occur:

python3: Network.h:103: virtual nvinfer1::DimsHW nvinfer1::NetworkDefaultConvolutionFormula::compute(nvinfer1::DimsHW, nvinfer1::DimsHW, nvinfer1::DimsHW, nvinfer1::DimsHW, nvinfer1::DimsHW, const char*): Assertion `(input.h() + padding.h() * 2) >= dkh && "Image height with padding must always be at least the height of the dilated filter."' failed.
Aborted (core dumped)

There is no error when using single input.


Any update with this question? I have the same doubt.


No update yet… May be solved in new version of TensorRT.
I just concat multiple input to simulate single input in my application