multi-task-speech-classification

Multi-Task Speech classification of accent and gender of an english speaker on Mozilla's common voice dataset. Paper can be found here

Run instructions

1. To preprocess the audio data, run

   python main.py preprocess -r <audio_data_path>

2. To train the model using the preprocessed audio data, run

   python main.py train -r <audio_data_path> -m <model_name> 

   Models Implemented: simple_cnn, resnet18, resnet34, resnet50, simple_lstm, bi_lstm, lstm_attention, bi_lstm_attention

3. To test the model on the test data, run

   python main.py test -r <audio_data_path> -m <model_name> -c <saved_model_path> 

4. To perform inference on the audio files directly, run

   python main.py inference -r <audio_files_path> -m <saved_model_path>