sarthak268/audio_classification_using_lstm

Classification of Urban Sound Audio Dataset using LSTM-based model.

Audio Classification using LSTM

Classification of Urban Sound Audio Dataset using LSTM-based model.

Requirements

- pytorch==1.0.1
- scipy==1.2.0
- torchvision==0.2.1
- pandas==0.24.1
- numpy==1.14.3
- torchaudio==0.2
- librosa==0.6.3
- pydub==0.23.1

Steps to follow for testing on your Test Data

• Create a folder named data/test in the current directory which will contain all the '.wav' files that are to be tested.

• Download 'bestModel.pt' from this Link and place in the current directory.

• Run the following commands:

python preprocess.py
python eval.py

• A csv file named 'test_predictions.csv' will be generated in the current directory containing all the test files along with their corresponding predicted labels.

Citation

In case you find any of this useful, consider citing:

@misc{audio-classification-using-LSTM,
  author = {Shagun Uppal, Anish Madan, Sarthak Bhagat},
  title = {sarthak268/Audio_Classification_using_LSTM},
  url = {https://github.com/sarthak268/Audio_Classification_using_LSTM},
  year = {2019}
}

Team

• Anish Madan
• Sarthak Bhagat
• Shagun Uppal

audio_classification_using_lstm's Issues

Dataset

Thanks for your hard work. Have you found that there are several audios that lasts more than 4sec? And another issue is that there are some format inconsistencies that may cause some audios cannot be read by pydub.

数据集

您好，感谢您伟大的工程分享。请问音频数据集是需要自己放进去么