Skip to content

Latest commit

 

History

History
32 lines (16 loc) · 1.25 KB

README.md

File metadata and controls

32 lines (16 loc) · 1.25 KB

How to use

To use this model converter you should have a .h5 keras model (trained) and a C++ compiler active (g++ > 8, make > 4.2, pip > 20.0 & python > 3.6). If these prerequisite are fulfill run these commands :

git clone https://gitlab.com/eg-julien/keras-model-compiler-cpp.git

cd keras-model-compiler-cpp

pip install requirements.txt

Change the content of the variable model_name to fulfill your requirements and then run :

python3 converter_keras.py

converter_keras.py is used to create a .tflite model from a .h5 keras model. It's also create a .h cpp library with model coefficient.

Inference the tflite model

First of all you'll need to clone and compile tensorflow lite. For that you just have to run these commands :

git clone --depth 1 https://github.com/tensorflow/tensorflow.git

cd tensorflow

make -f tensorflow/lite/micro/tools/make/Makefile generate_projects

Then you'll have to copy your models in your mbed project (maybe create a folder named models ?) after that you'll have to create a folder named tensorflow_lite and then copy tensforflow & third_party folders from tensorflow/lite/micro/tools/make/gen/$youros/prj/hello_world/mbed/ in it.

Add those folders to the path compiler and implements the inference.