Build failed in TensorRT C++ API in addElementwise() layer

Description

I build the Restormer model using TensorRT C++ API and in the last decoder level need to calculate X-mean formula so I used the addElementwise layer to calculate subtraction input X size {1, 2073600, 48} and mean size {1, 2073600, 1} the error is show as

2: [elementwiseBuilder.cpp::splitVectorizedHub::186] Error Code 2: Internal Error (Assertion !never(a.z >= 0) failed.)
2: [builder.cpp::buildSerializedNetwork::417] Error Code 2: Internal Error (Assertion enginePtr != nullptr failed.)

so I can’t fixed it out why this layer can’t be calculate while before layer use the same func to calculate and it can produced until this last decoder layer. I think it about the size of this input layer or memory?
plz help, thank you

Environment

TensorRT Version: 8.2.3
GPU Type: NVIDIA GeForce RTX 3060
CUDA Version: 11.5
CUDNN Version:
Operating System + Version: Ubuntu 18.04

Hi,

Could you please share with us complete error logs and issue repro script/model for better debugging.

Thank you.

Thank you I fixed it out it cuz the problem with my GPU memory size

This topic was automatically closed 14 days after the last reply. New replies are no longer allowed.