A repository for the assignment of the course Deep Learning shared with @yoshago and @Yishay1.
This repository implements a softmax using TensorFlow.
Our goal is to classify records by the accent of the recorder into five classes.
The five classes are the next one: UK, USA, French, Hebrew, USSR.
- The folder Data-set include all the data, adding the csv table.
- The folder Multinomial_logistic_regression include the code in python.
- The folder Paper include a paper describing the work, and the data.
- The folder Screens include screen-shots of the result of the softmax.
- The folder Presentation include an explanative explanation of the project.
To run the code, please type the next command from the terminal (terminal linux or gitbash for windows):
git clone https://github.com/SamuelBismuth/Voice_Recognition
Then you should be able to run the python script from your favorite IDE.
Please run only the main file. There is no need to change anything from the code, only download the packages.
Notice that the running time is long since our project is not finish yet (once the project finish, the data will no need to be converter from wav to mfcc, then the running time will be shorter).
- Add Hebrew data. @Yishay1
- Classing mfcc into folders ready to be read. @yoshago
- Adding hidden layers. @yoshago @SamuelBismuth
- Adding regularization. @yoshago @SamuelBismuth
- Paper. @SamuelBismuth
- Cheking the paper and submitting. @Yishay1 @SamuelBismuth @yoshago