Where to specify custom plugin's name and op name

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

I mimic plugin/leakyReluPlugin to write a custom plugin(replace name of leakyRelu as Custom), using nvinfer1::IPluginV2DynamicExt( while leakyReluPlugin* uses BasePlugin,this is different)
it compiled successfully, but failed when coverting a onnx model(contain Custom op), output message as following,
It seems the TensorRT lib not register my Custom plugin, I guess the problem is related to plugin name, name space, where is the problem?
so
1)what’s the exact plugin name? what’s the op name? where to specify op name?
(plugin name, op name, foldername and filename can be different from plugin name? plugin name is class name?)
2)where to specified the the plugin name and its space?

2

Hi,
Please refer to below links related custom plugin implementation and sample:

While IPluginV2 and IPluginV2Ext interfaces are still supported for backward compatibility with TensorRT 5.1 and 6.0.x respectively, however, we recommend that you write new plugins or refactor existing ones to target the IPluginV2DynamicExt or IPluginV2IOExt interfaces instead.

Thanks!

4

I have read the document carefully , get no answer, the sample is code,no more comments

5

Hi,

Please refer the following doc, which can help you.

Thank you.