Automatic Speech Recognition

2023 Spring, Introduction to Artificial Intelligence

National Yang Ming Chiao Tung University
Final Project, Group 26
Professor: Yi-Ting, Chen
For more information of this project, please check our report and slides.

Dataset source : https://www.kaggle.com/competitions/tensorflow-speech-recognition-challenge/data

code description

• preprocess.py : preprocess data with MFCC and fBank method
• lstm_keras.py : 1st approach for LSTM model
• lstm_torch.py : 2nd approach for LSTM model, with control group (convolutional network)
• model.py : for lstm_torch.py to use, main model structure
• plot.py : plot the accuracy and loss results and save .png
• elevator.py : main UI interface, with audio recording function

not used code

• preprocess_test.py : for debugging :)

credit:

陳昱喬 - code: preprocessing & UI design; presentation
林怡秀 - code: LSTM model; presentation
傅莉妮 - code: LSTM model; report (RNN LSTM); presentation
許維也 - report (preprocess method); presentation