datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
SpokenWOZ-Test-AudioSEA-Spoof
SEA-Spoof
Access And License
Access requires author approval. Please email the authors before requesting or using the dataset:
wu_jinyang@a-star.edu.sg
imcc.sg@gmail.com
This dataset is released for non-commercial academic research only.
Use is restricted to academic institutions and approved research users. Commercial use is not permitted, and this dataset may not be used by commercial companies or for commercial products, services, model training, evaluation… See the full description on the dataset page: https://huggingface.co/datasets/Jack-ppkdczgx/SEA-Spoof.spoken-squad-t2aspoken-multiturn-sft
Spoken Multi-turn SFT Japanese
Japanese spoken multi-turn SFT dataset generated from kanhatakeyama/AutoMultiTurnByCalm3-22B using CosyVoice2 TTS.
Dataset Description
This dataset contains Japanese multi-turn SFT (Supervised Fine-Tuning) data with spoken questions.
q1: First question (text + audio)
a1: First answer (text only)
q2: Follow-up question (text + audio)
a2: Second answer (text only)
Samples
ID
Q1
Q1 Audio
A1
Q2
Q2 Audio
A2
0
鉄は強磁性体ですか？… See the full description on the dataset page: https://huggingface.co/datasets/Atotti/spoken-multiturn-sft.spoken-alpaca-gpt4ANC-Spoof
ANC-Spoof: Audio Neural Codec Spoof Dataset
Overview
ANC-Spoof is a large-scale dataset for studying the
robustness of audio deepfake detection (ADD) systems against distortions introduced by
neural audio codecs. It pairs original (uncompressed) audio from three established ADD
benchmarks with codec-resynthesized versions of the same utterances, produced by eight
different neural codecs (including the uncompressed Original version).
The dataset is built from:… See the full description on the dataset page: https://huggingface.co/datasets/abdulahh35/ANC-Spoof.SpoofCelebindic-voices-hinglish-nospeakeroverlap-spon3.3-acronyms-fixed2SpokenNativQA
SpokenNativQA: Multilingual Everyday Spoken Queries for LLMs
The SpokenNativQA dataset consists of question-answer (QA) pairs, where queries are sourced from real users and answers are manually reviewed and edited. The dataset covers a diverse range of 18 topics that reflect culturally and regionally specific knowledge, as well as everyday queries. These topics include animals, business, clothing, education, events, food and drinks, general knowledge, geography, immigration… See the full description on the dataset page: https://huggingface.co/datasets/QCRI/SpokenNativQA.multi-stream-spontaneous-conversation-training-datasets_chinese
Multi-stream Spontaneous Conversation Training Datasets_Chinese
Every data point counts.
Dataset Basic Info
Dataset Type: ASR Corpus
Language: Chinese
Audio Parameters: 16 kHz, 16 bits
File Format: WAV (PCM)
Recording Equipment: Mobile device
Dataset Description
The Multi-stream conversation dataset developed by MagicData captures each speaker's audio track and labels each speaker separately, thereby preserving the natural occurrences of… See the full description on the dataset page: https://huggingface.co/datasets/MagicHub/multi-stream-spontaneous-conversation-training-datasets_chinese.spoken_squad_testThis dataset is licensed under the terms of the CC-BY-SA-4.0 license.
https://github.com/Chia-Hsuan-Lee/Spoken-SQuAD/blob/master/LICENSE.md
Author: @michaellee886
@article{li2018spoken,
title={Spoken SQuAD: A study of mitigating the impact of speech recognition errors on listening comprehension},
author={Li, Chia-Hsuan and Wu, Szu-Lin and Liu, Chi-Liang and Lee, Hung-yi},
journal={arXiv preprint arXiv:1804.00320},
year={2018}
}
@article{wang2024audiobench,
title={AudioBench: A… See the full description on the dataset page: https://huggingface.co/datasets/AudioLLMs/spoken_squad_test.SpokenSwag
SpokenSwag
We present here SpokenSwag as described in the paper "Slamming: Training a Speech Language Model on One GPU in a Day".
This dataset is based on allenai/swag and synthetised with 4 speakers from hexgrad/Kokoro-82M.
We show that perfoming DPO over the dataset can really improve performance of Speech Language Models.
We encourage you to also see the following resources, for further information:
Project Page: https://pages.cs.huji.ac.il/adiyoss-lab/slamming/ Paper:… See the full description on the dataset page: https://huggingface.co/datasets/slprl/SpokenSwag.SpotSound-Bench
SpotSound-Bench: A 'Needle-in-a-Haystack' Evaluation for Audio Temporal Grounding
Benchmark Summary
SpotSound-Bench is a challenging temporal grounding benchmark designed to evaluate Large Audio-Language Models (ALMs).
Existing benchmarks for audio temporal grounding often feature high ratios of target-window duration to full audio clip duration, which fail to simulate real-world scenarios where short events are obscured by dense background sounds. To bridge… See the full description on the dataset page: https://huggingface.co/datasets/Loie/SpotSound-Bench.indic-voices-hinglish-nospeakeroverlap-sponNeapolitan-Spoken-Corpus
Neapolitan Spoken Corpus (NSC)
A corpus of read Neapolitan speech for ASR evaluation, with a validated
Neapolitan–Italian lexicon, LOSO fine-tuning splits, trained LoRA adapters,
metric implementations, per-clip results, and error annotations.
This release supersedes the earlier 141-clip single-speaker version of this
repository. The earlier release corresponds to Speaker S1 of the present
corpus; the old audioData/ and transcripts.csv are replaced by
data/audio/ and… See the full description on the dataset page: https://huggingface.co/datasets/michaelcacioli/Neapolitan-Spoken-Corpus.indic-voices-hinglish-nospeakeroverlap-spon3podcasts_spotifyindic-voices-hinglish-nospeakeroverlap-spon3.1multi-stream-spontaneous-conversation-training-datasets_chinese
Multi-stream Spontaneous Conversation Training Datasets_Chinese
Every data point counts.
Dataset Basic Info
Dataset Type: ASR Corpus
Language: Chinese
Audio Parameters: 16 kHz, 16 bits
File Format: WAV (PCM)
Recording Equipment: Mobile device
Dataset Description
The Multi-stream conversation dataset developed by MagicData captures each speaker's audio track and labels each speaker separately, thereby preserving the natural occurrences of… See the full description on the dataset page: https://huggingface.co/datasets/MagicDataTech/multi-stream-spontaneous-conversation-training-datasets_chinese.SpokenWOZ-Train-Audiospoken-magpie-ja
Spoken-magpie
LLMの日本語Instruction Tuning用データllm-jp/magpie-sft-v1.0をCosyVoice2 TTSを使用して音声化した商用利用可能な日本語の音声言語モデルのSFT用データセットです。
ある程度の話者多様性を持つように生成されています。
Respone Audioは500文字以下の場合にのみ生成されています。
NVIDIA H200を10枚を使用しvllmで推論しました。
Samples
最初の50サンプルを掲載します。
ID
Instruction
Instruction Audio
Response
Response Audio
0
カボチャを使ったスイーツのレシピをいくつか教えてください。
もちろんです、カボチャを使ったスイーツは秋にぴったりですね。以下にいくつかのレシピをご紹介します。1. カボチャのスフレパウンドケーキ- 材料：カボチャ 200g、生クリーム 50ml、牛乳 50ml、卵 3個、砂糖 100g、薄力粉 70g、バニラエッセンス 少々-… See the full description on the dataset page: https://huggingface.co/datasets/Atotti/spoken-magpie-ja.SpokenTermDetection_Tedlium2Train
Dataset Card for "SpokenTermDetection_Tedlium2Train"
More Information needed
SpokenVisITSpokenVisIT
SpokenVisIT is a real-world visual-speech interaction benchmark built upon VisIT-Bench, designed to evaluate the visual-grounded speech interaction capabilities of omni large multimodal models (LMMs).
Our deepest acknowledgment goes to VisIT-Bench — A Benchmark for Vision-Language Instruction Following Inspired by Real-World Use — which collects a diverse set of real-world visual instructions. SpokenVisIT builds on this foundation by converting the textual instructions into spoken… See the full description on the dataset page: https://huggingface.co/datasets/ICTNLP/SpokenVisIT.spokenwoz_dst
Dataset: spokenwoz_whisper_dst
Short description
Prepared SpokenWoZ training dataset adapted for Whisper-style DST (dialog state tracking) tasks. Each example contains an utterance, normalized text, and aligned audio (16 kHz).
Key metadata
Number of examples: 73,950
Total size on disk: ~9.08 GB
Splits: train, validation
validation examples: 7,284
Features:
text (string): original utterance text
normalized_text (string): normalized form of the… See the full description on the dataset page: https://huggingface.co/datasets/vendrkat/spokenwoz_dst.indic-voices-hinglish-nospeakeroverlap-spon3.2partial-spoof-cross-domain-audit-data
Partial-Spoof Cross-Domain Audit: detector score outputs
Per-utterance and per-frame score outputs of three detectors (MRM, BAM, CFPRF) on
PartialSpoof, LlamaPartialSpoof, PartialEdit, and HQ-MPSD. Together with the analysis code they reproduce the reported
results of the cross-domain operational audit.
An earlier version of this study was submitted to IJCB 2026; that submission was
withdrawn and was never published. These arrays support one manuscript, currently in
preparation… See the full description on the dataset page: https://huggingface.co/datasets/sukhdeveyash/partial-spoof-cross-domain-audit-data.SPoRC
SPoRC: the Structured Podcast Open Research Corpus (V 1.1)
SPoRC is a large multimodal dataset for studying the podcast ecosystem. It contains metadata, full transcripts, speaker-turn-level diarization, speaker-role labels, and acoustic features for over 1.1 million podcast episodes across 228,000 podcasts.
Paper: Mapping the Podcast Ecosystem with the Structured Podcast Research Corpus (ACL 2025)
Upgrading from version 1.0? Read What changed since version 1.0 first. The file… See the full description on the dataset page: https://huggingface.co/datasets/blitt/SPoRC.indic-voices-hinglish-nospeakeroverlap-spon3.3LIEPA-3_read-spon_16kHz
LIEPA-3 Bendrinis Garsynas
Didysis lietuvių kalbos garsynas (LIEPA-3) – atviras kalbos duomenų rinkinys, skirtas šnekos atpažinimo tikslams ir moksliniams tyrimams.
Aprašymas
Šis duomenų rinkinys apima LIEPA-3 garsyno read ir spon dalis (fonetinė phon ir dialektų dial dalys neįtrauktos). Duomenys paruošti Whisper mokymui.
Pakeitimų istorija
1 versija — 2026-08-25 (20260825)
Pradinis pilno garsyno įkėlimas: 6,729,976 įrašai (be P)
Kalbėtojų… See the full description on the dataset page: https://huggingface.co/datasets/liepa-project/LIEPA-3_read-spon_16kHz.multi-stream-spontaneous-conversation-training-datasets_english
Multi-stream Spontaneous Conversation Training Datasets_English
Every data point counts.
Dataset Basic Info
Dataset Type: ASR Corpus
Language: English
Audio Parameters: 16 kHz, 16 bits
File Format: WAV (PCM)
Recording Equipment: Mobile device
Dataset Description
The Multi-stream conversation dataset developed by MagicData captures each speaker's audio track and labels each speaker separately, thereby preserving the natural occurrences of… See the full description on the dataset page: https://huggingface.co/datasets/MagicHub/multi-stream-spontaneous-conversation-training-datasets_english.