Stage 2: Accelerating the Pre/Post Processing Steps with CUDA - [ ] Utilize the CuPy library to accelerate the NumPy operations and compare the results - [ ] Experiment with some alternatives for CUDA pre/post processing and compare the results - [ ] ex. PyTorch NMS module - [ ] ex. OpenCV CUDA Module - [ ] Try executing operations against different object types - [ ] ex. OpenCV gpu.mat - [ ] ex. CuPy Matrix - [ ] ex. CUDA Stream - [ ] Find out how to profile GPU utilisation using Nsight Systems Helpful Links for CUDA runtime management: - [ ] [Invalid Resource Handle](https://forums.developer.nvidia.com/t/genericreformat-cuh-1487-error-code-1-cuda-runtime-invalid-resource-handle/303815) - [ ] [Running Multiple Engines](https://forums.developer.nvidia.com/t/running-multiple-tensorrt-engines-on-jetson-agx/181654) - [ ] [Multithreaded Engine Management](https://forums.developer.nvidia.com/t/how-to-use-tensorrt-by-the-multi-threading-package-of-python/123085) - [ ] [TRT Classes](https://forums.developer.nvidia.com/t/tensorrt-error-1-reformat-cu-1038-error-code-1-cuda-runtime-invalid-resource-handle/193394) - [ ] [Nvidia Guide for Multithreaded Engines](https://forums.developer.nvidia.com/t/how-to-use-tensorrt-by-the-multi-threading-package-of-python/123085/8)
Stage 2: Accelerating the Pre/Post Processing Steps with CUDA
Helpful Links for CUDA runtime management: