Does Torch-TensorRT fall back to Torch supported in C++ also?

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

As desribed in the following link Figure 4:
https://developer.nvidia.com/blog/accelerating-inference-up-to-6x-faster-in-pytorch-with-torch-tensorrt/

Is it supported also by the Torch, TRT and Torch-TensorRT c++ versions?

Thanks,