A neural network that classifies common sounds for individuals who are hard of hearing.
classify.py is used to run inference.
train.py is the file that was used for training the neural network (utilizing GPUs).
The models directory stores the saved TensorFlow models.
The Flask directory is used for the implementation of the neural network with web services.
The TensorFlow deep learning library was used to create the neural networks using Convolutional Neural Networks with Spectrogram Analysis.
The Librosa and PyDub libraries were used for handling audio files with python.
The model was trained on the ECS-50 dataset.