The implementation is made in Tensorflow for python and has been tested on Tensorflow 1.12 and python3.6. Other required packages are sklearn and numpy. No data is included to keep the repo small but can be generated using generated_data.m
- Install tensorflow and dependecies
- Run the following to generate data in matlab
generate_data("data_train.mat",10000)generate_data("data_test.mat",1000)
- Start training
pyhton main.py