Datasets:

sulabhkatiyar
/

ne-asr-dataset-nnp

audio audioduration (s) 1.12 15	text stringlengths 14 252	language stringclasses 1 value
	pha ama gaga aga gih pan tega mai tai	Wancho
	pan chu wai ngan ang kapley ema	Wancho
	lah ley eya alu ah balti khama tailey	Wancho
	eni tokphai ma e pan hagong boh jaw ngaila taiga	Wancho
	forest ma maikhee chu mo jau taiaa	Wancho
	church ku kunu ke kumei jaw tega	Wancho
	iyaa ah lainot ham chang ahh	Wancho
	hantoh lah ley ema ladka ani ngan ley tai ley	Wancho
	taka kam wall mame jari danla le photo man le cham taga	Wancho
	ehe bu siri sah mem dum la tai	Wancho
	busta busta ma sholey ley taga hancha busta	Wancho
	toma apak hagai ngantaiga	Wancho
	mala maipua lo jikko lela	Wancho
	kom gaga tingthak tam cham le taiga	Wancho
	kabu putla to hun zikku yang toh ah taega	Wancho
	kam kha cha chi ngan taiga	Wancho
	photo ma mihpa ani ngan taiga	Wancho
	age gang tan me nen ngan le taiga ache kawni gaga shua tok ma taiga	Wancho
	aga tokchem sanpe ngan taiga	Wancho
	ema jantam chu nganley tailey	Wancho
	pan nanlong ka sha chu te tailey	Wancho
	lato aa hago pa hatam cha la chola ngan taiga	Wancho
	ag photo ma mihnu adam ngan taiga	Wancho
	khirki tamta chu zangngai nganley tsezang ley ngan ang kapley	Wancho
	mahu phang ma kantong chu ta danley taiya	Wancho
	mihhuak ani chu chak ahet ley azong ngo taiya chutoh	Wancho
	khanak hu lah ley tale tale	Wancho
	ley chawak waklagley tai aa	Wancho
	khoma pasak danley ngan taiga	Wancho
	aga photo ma kui sui cha lai ngoipa ngan chong taga	Wancho
	tega khonyak ke mihnyu mihpa e bapley chong ngan taiga	Wancho
	drum hekai pat a ngan ley tai ley	Wancho
	ham kha bon tam ali taiga	Wancho
	to ma long long pu luk cha le ngan taiga	Wancho
	ama pan bu chu nam ra pan bu chu rae tai huley	Wancho
	nokphai ma a sintex ngan taiga	Wancho
	ma gaga sho ma ga nao am gaga khanyak ji ja kam lele ngan taga	Wancho
	ta te eko tang kau ringchi ngan taiga	Wancho
	aga khirki ma parda tuk lah a se ngan taiga	Wancho
	ekao side ma chu bus am danle ley tai tiya	Wancho
	kumaan ma bag ke ngan taiga	Wancho
	e photo ma tinu ngan tai a	Wancho
	aga photo ma ngan pa ham nyu kha sang a	Wancho
	aga photo ma ngan pe school kha sang ga	Wancho
	ama gaga white colour pa leley gaga jan jiji nu sa ngan taga	Wancho
	akao chan ma gaga tham chu pan chu moa ngan tega	Wancho
	tantohley tatama seko hame chu ngan ang kapley	Wancho
	gama kam chen kage tham che pan chego jaw moh a nyemley ngan tega	Wancho
	ham tuakphaima pan hingnu am	Wancho
	e nokfai machu seko du azam taiya	Wancho
	mihuak tareni khipa nipong sohley	Wancho
	photo ma ham ani athan ley ngan ley	Wancho
	aga naw uam e zoli khoma chongla dam agong taiga	Wancho
	minu am ee sari sohla ee ley taiga	Wancho
	ekhama teche ma ham chu ngan taiya	Wancho
	ega gasam mic chola jongja ngan taiga	Wancho
	kuni ma ema martun kapley taiga	Wancho
	ho nyak e kam hochong ho ngan taiga	Wancho
	hochen pan chak sa nanlong taiga	Wancho
	enu khama chu nao am anyi danle le tai hua	Wancho
	goal hat a pe sene cha taiga	Wancho
	nipong he khe khe kya mon mon kya ley apong lah le	Wancho
	hatoh lampho chen ma potsa am taiga	Wancho
	karpet tamcha ley nyem chem taiga	Wancho
	iya ozong hahwan long kho ee sii man wan	Wancho
	gym ding pa kam mong ngai la taiga	Wancho
	building chen ma cha pan ngai tai ga	Wancho
	antoley side ma parda bu chu nganley tailey	Wancho
	haisoi ma van post ta ayong taiya	Wancho
	ate vangu sa a che ngan tega	Wancho
	e tanki azam taiga tangma tuta pannu chu ajong ley tai a	Wancho
	untohley ema mistri mistri mamai ngan taiya	Wancho
	antoley jan ma danla taiya	Wancho
	khama kuye nausam sa bu ngan le laino ko lailai le	Wancho
	ema ta ham bahar ngan ley tai ley	Wancho
	panka tok to light jao laila ngan taga	Wancho
	akam floor ma marble lagai le tai a	Wancho
	hom ham tokto ye phoon hagam boh tula ngantaiga	Wancho
	chat ley pansa pansa yanu chu ngan ley tai ley	Wancho
	mihhuak cha e jong la hula taiga	Wancho
	chair chenma nyakmai pa chong taiga	Wancho
	ema chong a washolia taiga	Wancho
	gaman ma buty phunlo tham taiga	Wancho
	kom bama ye partition kori lah taiga	Wancho
	ate gama first ma bike kaunyak e ani ngan tega ete ate gaga lam ma	Wancho
	pha ag pan nyem loi ngan teyu	Wancho
	azam tailey azam jaw nakpa colour taila	Wancho
	aga ma gai chon nyai kam chu ngan taiga	Wancho
	shisha tok toh agin dan taiga	Wancho
	atey kam bo kanyu che tega atey ganong ka building noi ngan tega	Wancho
	tebul khoma sisa tingma phul anyi danla taiga	Wancho
	aa photo ma ham ngantaiga	Wancho
	loh ham chu nganley tailey	Wancho
	ikha ma machine ani ngan ley tai ley	Wancho
	jaga gaga kam ka boh chegai taiga	Wancho
	aga kamnu thao taiga khanyak ajok ngan taiga	Wancho
	ag photo ma kobuh ngai taiyo long kam chi ngantaiga	Wancho
	ema tamta lakka ngan ley ley tai ley	Wancho
	gaari ley ley te kaw e ya	Wancho
	ama mihnu cha mihpa cha le taga	Wancho

End of preview. Expand in Data Studio

Wancho (nnp) — ASR dataset

A small Wancho (nnp) speech-to-text dataset for automatic speech recognition (ASR) of a low-resource North-East India language. Each example pairs a short audio clip with its Romanized (Latin-script) transcript.

Source

Derived from the ARTPARK-IISc Vaani project (https://vaani.iisc.ac.in/)

Splits

Split	Samples
train	8,860
validation	1,163
test	1,099

Data fields

Each example has:

audio — the audio recording (16 kHz mono).
text — the Romanized (Latin-script) transcript.
language — the ISO 639-3 language code (nnp).

Used by

This dataset trains and evaluates the paired ASR model sulabhkatiyar/ne-asr-nnp — a character-level CTC adapter over a frozen facebook/mms-1b-all encoder. An augmented training variant is available at sulabhkatiyar/ne-asr-dataset-nnp-aug.

Part of the NE Speech AI V1 collection.

Limitations

Wancho is a low-resource language; this dataset is small and may not cover the full range of speakers, dialects, domains, and recording conditions.
Transcriptions are in Latin / Roman script (romanized); romanization conventions can introduce label noise.

License

cc-by-4.0, following the ARTPARK-IISc Vaani source dataset (https://vaani.iisc.ac.in/), from which this data is derived. Please retain that attribution and comply with the license terms when using this dataset.

Downloads last month: 78

Models trained or fine-tuned on sulabhkatiyar/ne-asr-dataset-nnp

sulabhkatiyar/ne-asr-nnp

Automatic Speech Recognition • Updated 5 days ago • 29

Collection including sulabhkatiyar/ne-asr-dataset-nnp

NE Speech AI V1

Collection

V1: character-level CTC ASR adapters over a frozen facebook/mms-1b-all for 14 low-resource North-East India languages. Honest per-sample WER. • 30 items • Updated 4 days ago