Error converting ONNX model to TensorRT engine on Jetson with CUDA 11.4 and TensorRT 8.4.6 (polygraphy)

Issue Description:

When attempting to convert an ONNX model to a TensorRT engine in a Jetson environment, the following error occurs:

[W] onnx2trt_utils.cpp:363: Your ONNX model has been generated with INT64 weights, while TensorRT does not natively support INT64. Attempting to cast down to INT32.
[W] onnx2trt_utils.cpp:391: One or more weights outside the range of INT32 was clamped
[W] IElementWiseLayer with inputs onnx::Add_554 and (Unnamed Layer* 652) [Shuffle]_output: first input has type Int32 but second input has type Float.
[W] IElementWiseLayer with inputs (Unnamed Layer* 654) [ElementWise]_output and (Unnamed Layer* 653) [Shuffle]_output: first input has type Int32 but second input has type Float.
[E] [layers.cpp::validate::2419] Error Code 4: Internal Error ((Unnamed Layer* 654) [ElementWise]: operation MAX has incompatible input types Int32 and Float)
[E] ModelImporter.cpp:748: While parsing node number 444 [Clip -> "onnx::Cast_565"]:
[E] ModelImporter.cpp:749: --- Begin node ---
[E] ModelImporter.cpp:750: input: "onnx::Add_554"
    input: "onnx::Clip_2582"
    input: "onnx::Clip_564"
    output: "onnx::Cast_565"
    name: "Clip_444"
    op_type: "Clip"
[E] ModelImporter.cpp:751: --- End node ---
[E] ModelImporter.cpp:753: ERROR: ModelImporter.cpp:179 In function parseGraph:
    [6] Invalid Node - Clip_444
    [layers.cpp::validate::2419] Error Code 4: Internal Error ((Unnamed Layer* 654) [ElementWise]: operation MAX has incompatible input types Int32 and Float)
[E] In node 444 (parseGraph): INVALID_NODE: Invalid Node - Clip_444
    [layers.cpp::validate::2419] Error Code 4: Internal Error ((Unnamed Layer* 654) [ElementWise]: operation MAX has incompatible input types Int32 and Float)
[!] Could not parse ONNX correctly

Environment Information:

• Jetson
• CUDA Version: 11.4.239
• cuDNN Version: 8.3.2.49
• TensorRT Version: 8.4.6
• Jetpack Version: 5.0.1

Steps to Reproduce:

1. Use the provided environment information and attempt to convert the model to a TensorRT engine.
2. Encounter the above error message.

Expected Behavior:

Expect successful conversion of the ONNX model to a TensorRT engine without encountering type mismatch errors.

Additional Information:

Provide any additional information or attempted solutions to help better understand and resolve the issue.

Start by reproducing the conversion in the stated Jetson, CUDA 11.4, TensorRT 8.4.6, and JetPack 5.0.1 environment. Inspect ONNX node 444, Clip_444, and the reported INT32/Float inputs to identify the model-side or parser-side mismatch. Done means the provided ONNX model converts successfully to a TensorRT engine without the type error.