{"slug": "show-hn-kuralhub-1-0-speech-emotion-datasets-for-70-languages-in-one-place", "title": "Show HN: KuralHub 1.0, Speech emotion datasets for 70 languages in one place", "summary": "KuralHub 1.0, a catalog of speech emotion datasets covering the 70 most widely spoken languages, launched with benchmark results for 11 pretrained speech models across 33 datasets, according to the project's Show HN post. The release packages 90 datasets with emotion labels, size, licence, access link and citation per language, plus all 363 benchmark runs as downloadable CSV and JSON with run logs, an Interspeech 2026 paper, and code updated for current PyTorch and Transformers. The benchmarks found Whisper had the best average but HuBERT led for nine languages, larger checkpoints did not help, Tamil, Telugu and Kannada were hard for every model, and 25 of the 70 languages — including Sinhala, Nepali, Hausa and Yoruba — still have no known emotional speech dataset.", "body_md": "Update to an earlier Show HN. KuralHub is a catalog of speech emotion datasets for the 70 most widely spoken languages, with benchmark results for 11 pretrained speech models on 33 of those datasets. Version 1.0 adds:\n\n- A searchable catalog with one page per language. Each of the 90 datasets lists its emotion labels, size, licence, access link and citation. - All 363 benchmark runs as downloadable CSV and JSON, with the logs for each run. - The published paper (Interspeech 2026), and code that runs on current PyTorch and Transformers. - A short form for suggesting a missing dataset or a model to benchmark.\n\nThe most practical finding: the right model depends on the language. Whisper had the best average, but HuBERT came out on top for nine languages, bigger checkpoints did not help, and Tamil, Telugu and Kannada were hard for every model.\n\n25 of the 70 languages still have no known emotional speech dataset, including Sinhala, Nepali, Hausa and Yoruba.\n\nCatalog: [https://aaivu.github.io/KuralHub/datasets.html](https://aaivu.github.io/KuralHub/datasets.html)\nResults: [https://aaivu.github.io/KuralHub/benchmark.html](https://aaivu.github.io/KuralHub/benchmark.html)\nWhat changed: [https://github.com/aaivu/KuralHub/releases/tag/v1.0.0](https://github.com/aaivu/KuralHub/releases/tag/v1.0.0)\nPaper: [https://www.isca-archive.org/interspeech_2026/thavarasa26_in...](https://www.isca-archive.org/interspeech_2026/thavarasa26_interspeech.pdf)\n\nSuggestions for missing datasets are welcome, especially for those 25 languages, and so are new languages beyond the current 70.\n\nComments URL: [https://news.ycombinator.com/item?id=50314172](https://news.ycombinator.com/item?id=50314172)\n\nPoints: 2\n\n# Comments: 0", "url": "https://wpnews.pro/news/show-hn-kuralhub-1-0-speech-emotion-datasets-for-70-languages-in-one-place", "canonical_source": "https://aaivu.github.io/KuralHub/datasets.html", "published_at": "2026-10-31 08:25:27+00:00", "updated_at": "2026-10-31 08:48:31.839440+00:00", "lang": "en", "topics": ["machine-learning", "natural-language-processing", "ai-research", "ai-tools"], "entities": ["KuralHub", "Whisper", "HuBERT", "Interspeech 2026", "PyTorch", "Transformers", "aaivu"], "also_reported_by": [], "alternates": {"html": "https://wpnews.pro/news/show-hn-kuralhub-1-0-speech-emotion-datasets-for-70-languages-in-one-place", "markdown": "https://wpnews.pro/news/show-hn-kuralhub-1-0-speech-emotion-datasets-for-70-languages-in-one-place.md", "text": "https://wpnews.pro/news/show-hn-kuralhub-1-0-speech-emotion-datasets-for-70-languages-in-one-place.txt", "jsonld": "https://wpnews.pro/news/show-hn-kuralhub-1-0-speech-emotion-datasets-for-70-languages-in-one-place.jsonld"}}