GitHub - jay-z20/TensorRT-RollPugin

swin-transformer roll插件实现

C++ Tensorrt 实现 pytroch 中 torch.roll()

优化代码实现矩阵 3.5 倍加速

// 保存 engine

roll.exe -s -shift 1 1 -dims 1 2

// 推理

roll.exe -shift 1 1 -dims 1 2

输入矩阵 3x4x5

Name		Name	Last commit message	Last commit date
Latest commit History 7 Commits
include		include
CMakeLists.txt		CMakeLists.txt
README.md		README.md
logging.h		logging.h
main.cpp		main.cpp
roll.cu		roll.cu
roll.h		roll.h
utilsn.h		utilsn.h