Yapdo-Mini is a sample of the Yapdo dataset, a conversational speech corpus drawn from 109,804 hours of approved recordings from 17,008 speakers across 67 languages.

Yapdo Data Highlights


Total approved audio	109,804 hours
Unique speakers	17,008
Languages	67 (human-verified labels)
Format	48 kHz, 16-bit PCM WAV per speaker
Channel separation	Each speaker on a dedicated, time-aligned track
Speech type	Spontaneous, unscripted, multi-party conversations
Code-switching	Yoruba-English, Hindi-English, Swahili-English ("Sheng"), Tagalog-Cebuano, and more
Mean SNR	~33 dB
Median RMS	-26 dBFS

Top 10 Languages

Language	Hours	Language	Hours
English	31,660	Tagalog	2,014
Hindi	8,412	Spanish	1,651
Arabic	2,427	Nigerian Pidgin	1,382
Swahili	2,075	Tamil	1,288
Hausa	2,074	Cebuano	848

Hindi alone exceeds FLEURS (~~12h) and Common Voice (~~18h) by over 100x.

Combined vs. Separated Audio

Each sample in this mini dataset is a combined mix of all speakers. The parent Yapdo corpus stores each speaker on a separate, time-aligned track. Here's what that difference sounds like — a Sheng (Swahili-English) conversation with 3 speakers:

Combined (all speakers mixed)

"Juu mbona iko iko ama ni pengine nmetoa hizi earphones ndo imeacha, imepunguza kurekodi. Unaniskia clear?"

Speaker 1 (isolated track)

Speaker 2 (isolated track)

Speaker 3 (isolated track)

All 17 Samples

#	Language	Speakers	Speech %	Notes
1	sw	3	62%	Sheng
2	hi	3	69%
3	tl	3	59%	Tagalog-English
4	te	3	56%	Telugu
5	sw	3	63%	Sheng
6	te	3	60%	Telugu
7	ar	4	61%	Egyptian Arabic
8	ar	4	68%	Egyptian Arabic
9	ta	3	52%	Tamil
10	pcm	3	58%	Nigerian Pidgin
11	en	4	64%	Egyptian accent
12	pcm	3	64%	Nigerian Pidgin
13	ta	3	64%	Tamil
14	tl	4	60%	Tagalog
15	hi	3	66%	Hindi-English
16	en	4	66%	Indian accent
17	en	3	63%	Nigerian accent

Schema

Column	Type	Description
`audio`	`Audio(16kHz)`	Combined multi-speaker audio, 16 kHz mono
`text`	`string`	Combined transcript from all speakers (AI-generated)
`language`	`string`	Primary ISO 639-1 language code
`num_speakers`	`int`	Number of speakers in the clip
`accents_self_reported`	`string`	Self-reported accent/dialect from user profiles
`recording_id`	`string`	Session ID linking to the source corpus
`duration_s`	`float`	Clip duration in seconds
`rms_dbfs`	`float`	RMS loudness in dBFS
`peak_amplitude`	`float`	Peak sample amplitude (0.0–1.0)
`speech_ratio`	`float`	Fraction of frames containing speech
`full_recording_duration_s`	`float`	Total duration of the original recording session in seconds
`notes`	`string`	Additional context (accent, language variety)

Usage

from datasets import load_dataset

ds = load_dataset("liva-ai/yapdo-mini", split="train")

for example in ds:
    print(f"{example['language']:>3s} | {example['num_speakers']} speakers | {example['notes']}")
    print(f"   Transcript: {example['text'][:100]}...")
    print()

Downloads last month: 79