Using custom layer IPluginV2Ext with UFF parser

AI & Data Science Deep Learning (Training & Inference) TensorRT
1

Hello.

I’m currently trying to load a UFF model into TensorRT 5.1 in C++. I have some custom layers inside the model that would like to implement using the IPluginV2Ext interface but it seems as though the nvuffparser::IPluginFactoryExt only supports IPluginExt.

I have looked at the AastaNV/Face-Recognition example where an IPluginFactory is used with an IPlugin custom layer but the documentation suggests to use the newer IPluginV2 API.

I am unclear on what is the intended way to go about loading a pretrained UFF model from disk with custom layers. Could someone give me some advice on how to go about doing this?

Thanks.