In original paper authors used Tensorflow v1. Tensorflow v1 is depricated and hard to implement. So I implement the same architecture using Pytorch.
Main file is Brain_Typing_Reproducing_with_Pytorch.ipynb. It is fully compatible with Google Colab.
Codes, data for brain typing paper.
Converting Your Thoughts to Texts: Enabling Brain Typing via Deep Feature Learning of EEG Signals (https://arxiv.org/pdf/1709.08820.pdf)
brain_typing.py: the python code
emotiv_7sub_5class.mat: the local dataset used in the paper
data_readme: the decription of local dataset
S1_nolabel.mat: an example of 1 subject, which is a subset of Physionet EEG MI database.
More information can be find in my homepage: http://www.cse.unsw.edu.au/~z5119405/
Please feel free to contact me if you have any questions or suggestions. (1015xzhang@gmail.com)