This is My TensorFlow Benchmark.
You can reproduce this result using my source code.
All used source code and environment are included in this repo.
Question/PR/Issues are always welcome
- Python
- C/C++
- Nvidia RTX2080 (8Gb)
- GeForce GTX 1050 Mobile (4Gb)
- AMD Ryzen 2700
- Intel(R) Core(TM) i7-7700HQ CPU @ 2.80GHz
- 128-core NVIDIA Maxwell™ architecture-based GPU
- Quad-core ARM® A57
- ARM Cortex-A53 1.4GHz
- TensorFlow
- TensorFlow-Lite
- x86_64 + TensorFlow + GPU + Python API + Float32
- x86_64 + TensorFlow + CPU + Python API + Float32
- x86_64 + TensorFlow + GPU + C API + Float32
- x86_64 + TensorFlow + CPU + C API + Float32
- x86_64 + TensorFlow-Lite + CPU + Python API + Float32
- x86_64 + TensorFlow-Lite + CPU + Python API + Int8
- aarch64 + TensorFlow + GPU + Python API + Float32
- aarch64 + TensorFlow-Lite + CPU + Python API + Float32
- aarch64 + TensorFlow-Lite + CPU + Python API + Int8
- aarch64 + TensorFlow-Lite + CPU + C++ API + Float32
- aarch64 + TensorFlow-Lite + CPU + C++ API + Int8
- armv7l + TensorFlow + CPU + Python API + Float32
- armv7l + TensorFlow-Lite + CPU + C++ API + Float32
- armv7l + TensorFlow-Lite + CPU + C++ API + Int8
You can build all benchmark tool using the source code in this repo.