# [Jeff Liao 20200329] The performance of models with OpenVINO Tool

**Hardware Device**
* OS : Windows 10 Professional (x64)
* IDE : Visual Studio 2019 Community
* CPU : Intel® Core™ i5-7500-CPU @ 3.4GHz 3.41GHz
* GPU : None
* RAM : 16.0GB
* Environment : Anaconda Python 3.7.6
* Library :
* tensorflow : 1.15.2 (CPU)
* OpenCV : 4.2.0-openvino
* OpenVINO : 2020 R1

## Benchmark app test
**Configuration Setup**
* Topology : Function_0
* Target device : CPU
* API : async
* Precision : MIXED
* Batch size : 1
* number of iterations : 0
* number of parrallel infer request : 4
* duration (ms) : 60000
* number of CPU streams : 4

**Result**

Note:
  Benchmark app is OpenVINO tool to test performance if model. It can test what you use on differnt device, including datatype, GPU, TPU, batch size, and API,etc.

## Implementation test
**Using Python to test models**