WarpForge MiniInfer
A CUDA/C++ kernel library and small Transformer inference runtime, validated against CPU and framework references and profiled with CUDA events and Nsight.
C++17CUDAcuBLASTensorRTPyTorchNsight
A CUDA/C++ kernel library and small Transformer inference runtime, validated against CPU and framework references and profiled with CUDA events and Nsight.