Classifying bird species from their birdsong.
Install the required dependencies using pip:
pip install -r requirements.txt
-
Prediction.ipynb: Demonstrates how to use the classifier to predict the bird class from both an audio file or a spectrogram and map the bird ID to a bird name. -
ModelTraining.ipynb: Shows how to train the model on the dataset, visualises the loss over training, and tests the model's performance on the test set. -
FilterTesting.ipynb: Tests the model performance on the filtered audio recordings that contain wind for more realistic real-world performance estimation. -
DataVisualization.ipynb: Inspects the dataset and visualizes the data to get a better understanding of the data. -
DataPreprocessing.ipynb: Used to split each audio file in the dataset into 5-second clips.
The dataset used for this project can be found here.