.cu plugin files about tensorrtx HOT 4 CLOSED

Any explanation for this wang-xinyu?

from tensorrtx.

For how to use plugin in tensorrt4, you can refer to https://docs.nvidia.com/deeplearning/tensorrt/archives/tensorrt_401/tensorrt-developer-guide/index.html#extending

Basically, a plugin need to implement a IPlugin class, and the .cu file implements the IPlugin methods. Some plugins need CUDA to accelerate the process, so it is using .cu as extension.

And for tensorrt7, refer to https://docs.nvidia.com/deeplearning/tensorrt/developer-guide/index.html#extending.

from tensorrtx.

What is your idea of BN algorithm in yolov3 in
IScaleLayer* addBatchNorm2d(INetworkDefinition *network, std::map<std::string, Weights>& weightMap, ITensor& input, std::string lname, float eps)
? i have a little feel that the BN algorithm in yolov3 is inconsistent with
Conventional formula。thank you

from tensorrtx.

@persuelx please open another thread for this question.

from tensorrtx.