I am practicing machine learning and would like to explore more precise functionality with emotion recognition in speech.

I am looking for a ranked datatset or a trained model. Current libraries are narrow, something over 250 segments with emotion scaled in separate data file from soundbytes should be a good start.


There are many datasets available out there, with different set-ups (no noise/ noisy) targeting various emotion classes (positive/negative, anger/happy/sad etc.) and covering various languages (Arabic, French, German, English etc.). Here are a few:

For more information, please refer to this list https://superkogito.github.io/SER-datasets/, that I maintain on GitHub: SER-datasets. This list includes a small description of each dataset, its languages and the emotions classes included.

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.