Pocket AI

A Portable Toolkit for deploying Edge AI and HPC

pai::engine

• cl: A small computing framework based on opencl. This framework is designed to help you quickly call Opencl API to do the calculations you need.

• vk: A small computing framework based on vulkan. This framework is designed to help you quickly call vulkan's computing API to do the calculations you need.

• graph: A small multitasking scheduler that can quickly build efficient pipelines for your multiple tasks.

pai::memory

• align_alloc
• allocator
• blocking_queue
• frame_shift_cache
• huffman_encoder
• ring_buffer

pai::thread

• internal_thread
• thread_pool

tests

Unit tests

cd tests && ./build_win_x86.bat # windows
cd tests && ./build.sh          # linux
# run
.\bin\unit_tests

NOTE

test -> 单元测试

pai::engine -> graph / vk / cl / infer

pai::type -> Vector (相同使用方式，添加约束进行加速) / Mat

pai::thread -> pthread（gemmlowp-pthread_everywhere） / InternalThread / ThreadPool

pai::memory -> Align / RingBuffer / FrameShift / BlockingQueue / MemReuser(内存复用 / 内存分区，用于管理高速缓存的使用) / MemPool(内存池) / Compressor(数组/权重压缩)

pai::prof -> AsmPeakPerf(宏定义代码块) / Timer（gflops） / MemRecorder (直接替换malloc/calloc)

pai::util -> basic_marco / Logger / Type / PcmReader / BmpReader

scrip -> obfuscator (.a / .hpp)

---

Usage

直接链接头文件，即可使用。

如:

#include "pai/engine/graph/graph.hpp"
#include "pai/engine/cl/engine.hpp"
#include "pai/engine/vk/engine.hpp"