zhangwq740 / SER-ESC-50

pytorch - DSP - audio_classification

SER-ESC-50

pytorch - DSP - audio classification

Sonic Environment Recognition

A simple audio classifier that recieves a wav file and returns a classification from 50 categories

How to Use

git clone https://github.com/FinchMF/SER-ESC-50.git

pip install -r requirements.txt

bash run.sh

python app.py

Once all requirements are installed, executing run.sh will:

• fetch all necessary data (ESC-50 audio dataset)
• train both a CNN and a modified resNet 34

Once the models are trained, executing app.py will:

• run a local server with a simplistic front end to interact with the classifier
• upload wav file and the classifier will respond with a classificaiton

Coming Soon

Producionized application using AWS