Datasets:

speedykom-group
/

karamojong-speech-dataset

audio audioduration (s) 2 12.1	transcription stringlengths 11 122
	samuson kaŋikue ŋul
	tolimok yesu ŋun akemut
	ŋeenan erubakinete anakirionot kaŋina emam nyedaun erai ketwaka ŋakicauate da nai yesu kristo erai ekapolon akica
	irai iyoŋ eberu ŋina ejokan
	ani ebala aburaam ecum isaako
	naiteanen abu akuj
	nai nyebuno yesu tasec
	ani ikiwapit ioni yesu
	apei koloŋit toloto saul
	iteyaruni eyokotau
	ŋol lo abutosec anenikeŋ
	ŋul daadaŋ iwapito erot ayesu
	abu erwosit iinakin eoda
	abu akuj tolimoŋ isirael ebe ŋilemut idiobore
	kipeyiok naman elisa alakaikeŋ
	alosit iurea akitojor eŋimai a isirael
	kiŋolik yesu iŋes alodiete abwan
	isudok iyoŋ mere neni
	mam pa pedori atwanare
	abu stepanokitoyeniŋ yudayan ŋul ebe
	togiela apolou ŋina karaaku ekiyai alokal kotere
	totiyarosikop sodoma
	nai apakeŋ arai akuj apei ŋina akire
	erai ŋul ijaunete akirot
	emaikin aiwon ominasi akuj
	keyai koloŋ etuŋanan ŋolo abaran
	kijau erwosit yes ta
	abu eyenet inak boas
	anakisicenoka twanare ŋina aŋikar kaŋikar
	taarai keke yesu ŋola
	ape koloŋit elos daudi alokitiama kaikeŋ
	toriam iŋes ŋican nooi ka tonyam akoro
	epedorio akuj ainakin
	apedor neni kaaberu
	pakiti epei kanupan ŋunakeŋ
	einakinio akuj akiar ŋina emam
	etorube ŋolo aŋitomon aŋiarei
	kilelebun paulo
	pa arai ijael ŋina isirael
	yai ŋes akiboi keeli
	aninataparas ŋin ace
	apaginiti eloete kes
	epurosi iŋin yudayan akuj
	ainakia yoŋ etaunu
	katorotokini ameli
	abwa akuju tema nabo neni
	ekemaimai aeoŋ yes ŋikonei
	amurunoto ŋina kaarecit
	akipup italakarit akuj
	apotu ŋkonei kiŋolikis yesu bon
	inyenyami nabo akoro ŋes jik
	acamit ŋes aker daudi
	eyakaunor isimael ŋudwe ŋulu alalak
	abu kitaŋan ewekile ŋoloadiakaon
	aŋun aeporo ŋes emam ŋin ebuci
	erai lo lokokukaŋ
	abu ŋes tosub kidiama
	aŋuna itia iyoŋ nu
	arai nemia eyudayat
	anen toasik aguj saul
	ekiŋarin alakara saul aŋuna daudi
	tarau ŋina aŋkar kaŋkar
	abu aku iyaun kabilai aŋitiaŋ da daoŋ nen anaa
	ekile ece ŋolo
	potu nen ilala tonupa
	apotu ŋisirael kiiras samuel
	abu akuju kiricaŋakiciketekeŋ
	ayaŋ imucurus toma
	koriara amuduko akisicanet nenika akuj
	ayayek ekelipan
	togolok akuj ikes toma
	abu etau ka akuj kitogogoŋ
	abu kitui aceumokinet
	alorokara eoŋ noi eturi yesu
	ŋituŋa ŋul luaponito anakwapin dadaŋ
	kimaima aburaam akuj
	ebunit yesu anenikaguj
	atwanare alokitoi aŋuna aŋituŋa
	etiy eoŋ apaarasia dadaŋ
	kiŋwolik etorube ŋol ce iirari iyoŋ eruye lo
	ŋisigaria ŋes lomusalaba oŋolo
	tobuluŋ kaswamakka tamasi
	ka iwani da da ŋul epani tasecai nen ka akuj
	apotu kiiras etoil
	ayakatar eyenet ece apedori ŋina kiitaret
	abu yakob inaki yosep ekaosoŋ loabusan nooi
	ka tocam akuj amunyar kes
	elelebun akwap kan aronis
	abu akuj totub
	kelunyareŋ atwanaria yesu alom salaba
	katacitere ŋasacisio ayoŋ
	erai ekapolon yesu kristo
	kiyakiya musaŋ kar ebok
	tama is ŋ yudaan akeleleŋicak
	nai ani itami iyoŋ
	yesu epiori ŋican ka edeke ka arieboka kemea satani euruwana ayok kiyaral akwap
	tokianatar ŋin ice kiliokoi ŋes
	ŋikuei ŋamiyai ŋauni
	einakini yesu apedor ka kiar
	emaiki n akiŋarak ŋulce daŋ

End of preview. Expand in Data Studio

Ng'akarimojong Speech Dataset

Speech dataset for Ng'akarimojong (kdj) — Eastern Nilotic language, ~370,000 speakers, Karamoja, NE Uganda.

Property	Value
Format	WAV, 16 kHz, mono / UTF-8 transcripts
Source	GRN recordings, segmented via silence detection
Transcription	Auto-generated via `facebook/mms-1b-all` (Teso adapter)

About

This dataset was created by Speedykom as part of an effort to advance speech technology for underserved African languages. Ng'akarimojong is spoken by approximately 370,000 people in the Karamoja sub-region of northeastern Uganda and belongs to the Ateker (Teso-Turkana) language cluster within the Eastern Nilotic family.

Usage

from datasets import load_dataset

ds = load_dataset("speedykom-group/karamojong-speech-dataset")
print(ds["train"][0])

Notes

Transcriptions were generated using the Teso (teo) ASR adapter from facebook/mms-1b-all — the closest available language to Karamojong. Manual review by a native speaker is recommended.
Audio sourced from publicly available GRN recordings. Please respect the original license terms.

Citation

If you use this dataset, please credit:

Speedykom - karamojong-speech-dataset
https://huggingface.co/datasets/speedykom-group/karamojong-speech-dataset
Created by Speedykom (https://speedykom.de)

Created by Speedykom

Downloads last month: 94

Models trained or fine-tuned on speedykom-group/karamojong-speech-dataset

speedykom-group/karamojong-tts-vits-beta

Text-to-Speech • Updated 2 days ago • 70