KuralHub lists 2 speech emotion recognition datasets for Russian, 2 of them openly available. Each entry below gives the emotion labels, size, recording setup, licence, where to get it and how to cite it. Russian is also part of the benchmark. The best of 11 models reached 51.0% test accuracy (hubert-base-ls960).
Language family
Indo-European (Slavic)
ISO code
ru
Spoken mainly in
Russia, Belarus, Ukraine, Kazakhstan, and Kyrgyzstan
Russian is one of the 29 languages in the KuralHub benchmark. Each of the 11 frozen speech models was trained with a small classification head on the shared emotions and scored on a held-out test split. Over all 11 runs the mean test accuracy was 34.5% (median 40.0%). The best run reached 51.0% with hubert-base-ls960. Compare with other languages.
RESD (run ID RESD). Scores on the test split, best run highlighted.
Dusha is a bi-modal corpus suitable for speech emotion recognition (SER) tasks. The dataset consists of about 300 000 audio recordings with Russian speech, their transcripts and emotional labels. The corpus contains approximately 350 hours of data. Four basic emotions that usually appear in a dialog with a virtual assistant were selected: Happiness (Positive), Sadness, Anger and Neutral emotion.
The dataset focuses on emotional speech dialogues in Russian, capturing live speech performed by actors who express specific emotions in pre-scripted dialogues. The dataset contains approximately 3.5 hours of speech, with each dialogue lasting around 3 minutes. It includes audio samples recorded at either 16,000 Hz or 44,100 Hz, corresponding to one of seven labeled emotional states: anger, disgust, fear, enthusiasm, happiness, neutral, and sadness. In addition to the emotional labels, each sample also includes speech-to-text transcriptions of the spoken dialogue. The dataset is useful for tasks related to Speech Emotion Recognition (SER), emotion detection in speech, and speech-to-text modeling.
Used in the KuralHub benchmark (run ID RESD). See the results above.
Citation
@misc{Aniemore,
author = {Артем Аментес, Илья Лубенец, Никита Давидчук},
title = {Открытая библиотека искусственного интеллекта для анализа и выявления эмоциональных оттенков речи человека},
year = {2022},
publisher = {Hugging Face},
journal = {Hugging Face Hub},
howpublished = {\url{https://huggingface.com/aniemore/Aniemore}},
email = {hello@socialcode.ru}
}
Know another Russian speech emotion dataset, or a newer version of one listed here? Suggest it or open a pull request.
This page is generated from datasets/Russian/ in the KuralHub repository.
If you use a dataset, cite its authors. If KuralHub helped you find it, please also cite the KuralHub paper.