Is there any information available on the relative performance of the Python and C++ interfaces to TensorRT?
Has anyone benchmarked the performance of these two inference implementations, for the same trained model?
Thanks
Is there any information available on the relative performance of the Python and C++ interfaces to TensorRT?
Has anyone benchmarked the performance of these two inference implementations, for the same trained model?
Thanks
Hello,
Informal TRT5 tests suggest negligible overhead with Python.