Hi, TX2 (Jetpack4.2.3 TensorRT5.1) was used in the previous project, because the network model has a custom layer, and the TensorRT API ‘createConcatPlugin’ conversion is used, which can normally achieve model acceleration and inference.
Now testing on AGX (Jetpack4.4 TensorRT7.1), an error message appears when compiling the project, ‘createConcatPlugin’ is not a member of’nvinfer1::plugin’. Compare the API interfaces of TensorRT5.1 and TensorRT7.1 and find TensorRT7.1 The interface has changed a lot.
Is there any other interface for TensorRT7.1 that can replace ‘createConcatPlugin’?
Or other methods can be solved, thank you very much.
The Concat plugin layer basically performs the concatention for 4D tensors. Unlike the Concatenation layer in early version of TensorRT, it allows the user to specify the axis along which to concatenate. The axis can be 1 (across channel), 2 (across H), or 3 (across W). More particularly, this Concat plugin layer also implements the “ignoring the batch dimension” switch. If turned on, all the input tensors will be treated as if their batch sizes were 1.