datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
yodas-granary
Dataset Card for YODAS-Granary
Repository: NeMo-speech-data-processor: Granary
Paper: Granary: Speech Recognition and Translation Dataset in 25 European Languages
Shared by: ESPnet
Dataset Description
YODAS-Granary is a curated subset of the larger nvidia/Granary dataset, focusing on high-quality pseudo-labeled speech data for Automatic Speech Recognition (ASR) and Automatic Speech Translation (AST) across 23 European languages.
Overview… See the full description on the dataset page: https://huggingface.co/datasets/espnet/yodas-granary.yodasUpdates
2024/07/09: we also uploaded a new version of YODAS as YODAS2, it provides unsegmented audios and higher sampling rate (24k)
README
This is the YODAS manual/automatic subset from our YODAS dataset, it has 369,510 hours of speech.
This dataset contains audio utterances and corresponding captions (manual or automatic) from YouTube. Note that manual caption only indicates that it is uploaded by users, but not necessarily transcribed by a human
For more details about YODAS… See the full description on the dataset page: https://huggingface.co/datasets/espnet/yodas.yodas2YODAS2 is the long-form dataset from YODAS dataset.
It provides the same dataset as espnet/yodas but YODAS2 has the following new features:
formatted in the long-form (video-level) where audios are not segmented.
audios are encoded using higher sampling rates (i.e. 24k)
For detailed information about YODAS dataset, please refer to our paper and the espnet/yodas repo.
Usage:
Each data point corresponds to an entire video on YouTube, it contains the following fields:
video_id:… See the full description on the dataset page: https://huggingface.co/datasets/espnet/yodas2.Bagpiper_SFT_Data
Bagpiper SFT Data
Release status: the validated Parquet release is being uploaded. The
homepage and metadata may appear before every large shard is committed.
Bagpiper SFT Data is the supervised fine-tuning corpus for
Bagpiper, an open-ended audio language model
that understands and generates speech, music, environmental sound, and their
mixtures through rich textual captions and planning.
The public release has exactly two configurations:
Configuration
Direction… See the full description on the dataset page: https://huggingface.co/datasets/espnet/Bagpiper_SFT_Data.floras
FLORAS
FLORAS is a 50-language benchmark For LOng-form Recognition And Summarization of spoken language.
The goal of FLORAS is to create a more realistic benchmarking environment for speech recognition, translation, and summarization models.
Unlike typical academic benchmarks like LibriSpeech and FLEURS that uses pre-segmented single-speaker read-speech, FLORAS tests the capabilities of models on raw long-form conversational audio, which can have one or many speakers.
To… See the full description on the dataset page: https://huggingface.co/datasets/espnet/floras.yodas_owsmv4🏆 News: Our OWSM v4 paper won the Best Student Paper Award at INTERSPEECH 2025!
Dataset Card for YODAS_OWSMv4
Paper: OWSM v4: Improving Open Whisper-Style Speech Models via Data Scaling and Cleaning (Best Student Paper at INTERSPEECH 2025)
Authors: Yifan Peng, Muhammad Shakeel, Yui Sudo, William Chen, Jinchuan Tian, Chyi-Jiunn Lin, Shinji Watanabe
Data Cleaning Scripts: ESPnet
Model Demo: Gradio
Dataset Description
Open Whisper-style Speech Model (OWSM)is the first… See the full description on the dataset page: https://huggingface.co/datasets/espnet/yodas_owsmv4.Bagpiper_PreTrain_Data
Bagpiper Pretraining Data
Bagpiper Pretraining Data is the public rich-captioned audio snapshot associated
with Bagpiper, an open-ended audio language
model that learns bidirectional mappings between audio and comprehensive text
descriptions across speech, music, environmental sound, and mixtures.
The en metadata describes the primary rich-caption language. Source audio can
contain speech or singing in other languages; it is not an English-only audio
guarantee.
The repository… See the full description on the dataset page: https://huggingface.co/datasets/espnet/Bagpiper_PreTrain_Data.Bagpiper_TTS_SFT_Data
Bagpiper-TTS SFT Data
Release status: the validated Parquet release is being uploaded. The
homepage and metadata may appear before every large shard is committed.
Bagpiper-TTS SFT Data supports
Bagpiper-TTS, a universal
speech-synthesis model that interprets free-form natural-language requests,
plans the requested delivery, produces a rich textual caption, and synthesizes
the target audio.
The release is organized into the six applications used by the paper:… See the full description on the dataset page: https://huggingface.co/datasets/espnet/Bagpiper_TTS_SFT_Data.raw_tts_emilia_ESPnet_espnet_mls-multi_soundstream_16kace-opencpop-segments
Citation Information
@misc{shi2024singingvoicedatascalingup,
title={Singing Voice Data Scaling-up: An Introduction to ACE-Opencpop and ACE-KiSing},
author={Jiatong Shi and Yueqian Lin and Xinyi Bai and Keyi Zhang and Yuning Wu and Yuxun Tang and Yifeng Yu and Qin Jin and Shinji Watanabe},
year={2024},
eprint={2401.17619},
archivePrefix={arXiv},
primaryClass={cs.SD},
url={https://arxiv.org/abs/2401.17619},
}
raw_tts_esc_ESPnet_espnet_mls-audioset_soundstream_16kmms_ulab_v2MMS ulab v2 is a a massively multilingual speech dataset that contains 8900 hours of unlabeled speech across 4023 languages. In total, it contains 189 language families.
It can be used for language identification, spoken language modelling, or speech representation learning.
MMS ulab v2 is a reproduced and extended version of the MMS ulab dataset originally proposed in Scaling Speech Technology to 1000+ Languages, covering more languages and containing more data.
This dataset includes the raw… See the full description on the dataset page: https://huggingface.co/datasets/espnet/mms_ulab_v2.OE-LoL-Esports-Datasetspeechlm_bpe_tts_mls_mix_train_valle_espnet_mls-multi_soundstream_16kLichessGamesraw_tts_esc_ESPnet_espnet_mls-multi_soundstream_16klol-esports-matches
GPTilt: League of Legends Esports Matches
This dataset is part of the GPTilt open-source initiative, aimed at democratizing access to high-quality LoL data for research and analysis, fostering public exploration, and advancing the community's understanding of League of Legends through data science and AI. It provides a clean, canonical record of the competitive matches and games of professional League of Legends.
By using this dataset, users accept full responsibility for any… See the full description on the dataset page: https://huggingface.co/datasets/gptilt/lol-esports-matches.code_search_net_python_10000_examplesESpeech-webinars2
Webinar Audio Dataset
Dataset Description
This dataset contains 850 hours processed webinar audio segments with corresponding metadata. Each audio file represents a segment extracted from webinar recordings, processed at 44.1kHz sample rate.
Dataset Summary
Language: Russian
Task: TTS, ASR, Quality Asessment
Audio format: MP3, 44.1kHz sample rate
Structure: Segmented audio files with JSON metadata
Dataset Structure
Data Fields… See the full description on the dataset page: https://huggingface.co/datasets/ESpeech/ESpeech-webinars2.raw_bpe_tts_mls_ESPnet_espnet_mls-english_soundstream_16kace-kising-segments
Citation Information
@misc{shi2024singingvoicedatascalingup,
title={Singing Voice Data Scaling-up: An Introduction to ACE-Opencpop and ACE-KiSing},
author={Jiatong Shi and Yueqian Lin and Xinyi Bai and Keyi Zhang and Yuning Wu and Yuxun Tang and Yifeng Yu and Qin Jin and Shinji Watanabe},
year={2024},
eprint={2401.17619},
archivePrefix={arXiv},
primaryClass={cs.SD},
url={https://arxiv.org/abs/2401.17619},
}
speechlm_bpe_tts_mls_mix_train_valle_espnet_mls-audioset_soundstream_16kspeechlm_bpe_tts_mls_mix_train_valle_espnet_mls-english_soundstream_16kEsportsBench
EsportsBench: A Collection of Datasets for Benchmarking Rating Systems in Esports
EsportsBench is a collection of 20 esports competition datasets. Each row of each dataset represents a match played between either two players or two teams in a professional video game tournament.
The goal of the datasets is to provide a resource for comparison and development of rating systems used to predict the results of esports matches based on past results. Date is complete up to 2026-03-31.… See the full description on the dataset page: https://huggingface.co/datasets/EsportsBench/EsportsBench.raw_tts_esc_ESPnet_espnet_mls-english_soundstream_16kDSUChallenge2024
The Interspeech 2024 Challenge on Speech Processing Using Discrete Units
Paper: https://www.isca-archive.org/interspeech_2024/chang24b_interspeech.html
Arxiv: https://arxiv.org/abs/2406.07725
Challenge details: https://www.wavlab.org/activities/2024/Interspeech2024-Discrete-Speech-Unit-Challenge/
To cite:
@inproceedings{chang24b_interspeech,
title = {The Interspeech 2024 Challenge on Speech Processing Using Discrete Units},
author = {Xuankai Chang and Jiatong Shi and… See the full description on the dataset page: https://huggingface.co/datasets/espnet/DSUChallenge2024.lol-esports-entities
GPTilt: League of Legends Esports Directory
This dataset is part of the GPTilt open-source initiative, aimed at democratizing access to high-quality LoL data for research and analysis, fostering public exploration, and advancing the community's understanding of League of Legends through data science and AI. It provides a clean, canonical reference for the people and organizations of competitive League of Legends.
By using this dataset, users accept full responsibility for any… See the full description on the dataset page: https://huggingface.co/datasets/gptilt/lol-esports-entities.espeech_podcasts_chunked_tts_train
espeech_podcasts_chunked_tts_train
This is a gated Russian TTS training dataset from instinct-org.
This repository contains tokenized or prepared speech data for text-to-speech training workflows.
Language
Primary language: ru (Russian)
Intended Use
text-to-speech training
Internal dataset curation, quality checks, and model evaluation
Research or commercial use only after access approval and license review
Data Notes
Prepared for TTS training… See the full description on the dataset page: https://huggingface.co/datasets/instinct-org/espeech_podcasts_chunked_tts_train.wikitonguesThe WikiTongues speech corpus is a collection of conversational audio across 700+ languages.
It can be used for spoken language modelling or speech representation learning.
This dataset includes the raw unsegmented audio in a 16kHz single channel format.
Each clip is usually 2-10 minutes long, and contains one or more speakers conversing in their language(s).
Sometimes, a speaker may switch languages within a single clip.
The total dataset size is around 70 hours.
The current version of the… See the full description on the dataset page: https://huggingface.co/datasets/espnet/wikitongues.ml_superb_hf