Running the script

[Manik-Taneja]

Hey, I am trying to run this to identify stress detection in a corpus, I am unable to understand how to run your code over a corpus, if you could explain that would be helpful.

[vivekkr12]

We don't have the pre trained model here in this repo, however you can train your own model to make inferences. You'll need to follow the steps mentioned in the README file - follow the section - To retrain the model follow the steps below

The phone alignment script takes a CSV as input parameter. If you're using a corpus same as this project, you can easily use one of our scripts to generate that CSV.

• script for LibriSpeech
• script for Oscaar

Once you have the output of phoneme alignment, use the output and follow the next steps in README for training the model. The hyperparameters that we used for training are in the report.

With a trained model, making inference will follow the same steps - alignment > feature generation > prediction