Datasets:

sulabhkatiyar
/

ne-asr-njm-aug

text stringlengths 22 155	language stringclasses 1 value	augmentation stringclasses 3 values
keba mhierhi ha nu themia puo bagi kezha puo mhodzü nu ketha ngu ba	Angami	original
keba mhierhi ha nu themia puo bagi kezha puo mhodzü nu ketha ngu ba	Angami	speed_0.9
keba mhierhi ha nu themia puo bagi kezha puo mhodzü nu ketha ngu ba	Angami	speed_1.1
hanu cycle ngu ba gari ngu ba themia kekra line chüpfüthacü ngu ba süshie puo mho nu thu keba sbi yono cash idi thu khaba	Angami	original
hanu cycle ngu ba gari ngu ba themia kekra line chüpfüthacü ngu ba süshie puo mho nu thu keba sbi yono cash idi thu khaba	Angami	speed_0.9
hanu cycle ngu ba gari ngu ba themia kekra line chüpfüthacü ngu ba süshie puo mho nu thu keba sbi yono cash idi thu khaba	Angami	speed_1.1
siro nipukoi puo mevi mu ketho ko pese kertso kezivi se pie ba	Angami	original
siro nipukoi puo mevi mu ketho ko pese kertso kezivi se pie ba	Angami	speed_0.9
siro nipukoi puo mevi mu ketho ko pese kertso kezivi se pie ba	Angami	speed_1.1
sibo puo phinu ro nyiepfhü bou puo kepezho di ba	Angami	original
sibo puo phinu ro nyiepfhü bou puo kepezho di ba	Angami	speed_0.9
sibo puo phinu ro nyiepfhü bou puo kepezho di ba	Angami	speed_1.1
mhizhüu ro puo tsü gei ro khrüphi pie chü pie ba mu puo phiko ro si pie chü pie ba	Angami	original
mhizhüu ro puo tsü gei ro khrüphi pie chü pie ba mu puo phiko ro si pie chü pie ba	Angami	speed_0.9
mhizhüu ro puo tsü gei ro khrüphi pie chü pie ba mu puo phiko ro si pie chü pie ba	Angami	speed_1.1
mia ho mia ro ba di keloo le ba mu miaho mia ro pfhüdi le ba miaho mia ro thadi gali kerhü tha	Angami	original
mia ho mia ro ba di keloo le ba mu miaho mia ro pfhüdi le ba miaho mia ro thadi gali kerhü tha	Angami	speed_0.9
mia ho mia ro ba di keloo le ba mu miaho mia ro pfhüdi le ba miaho mia ro thadi gali kerhü tha	Angami	speed_1.1
mhanya sü chü pie su pie ba mu kese pie ba	Angami	original
mhanya sü chü pie su pie ba mu kese pie ba	Angami	speed_0.9
mhanya sü chü pie su pie ba mu kese pie ba	Angami	speed_1.1
kiu mhodzü nü ro bagi kekra kenie mu keti puo ngu ba	Angami	original
kiu mhodzü nü ro bagi kekra kenie mu keti puo ngu ba	Angami	speed_0.9
kiu mhodzü nü ro bagi kekra kenie mu keti puo ngu ba	Angami	speed_1.1
boulie pfhenyi keze ki khekaa ro boulie keterhü themia puo ba	Angami	original
boulie pfhenyi keze ki khekaa ro boulie keterhü themia puo ba	Angami	speed_0.9
boulie pfhenyi keze ki khekaa ro boulie keterhü themia puo ba	Angami	speed_1.1
mhirhi hanu ki ko rhupie khashüba ki nu sei kropuo nguba	Angami	original
mhirhi hanu ki ko rhupie khashüba ki nu sei kropuo nguba	Angami	speed_0.9
mhirhi hanu ki ko rhupie khashüba ki nu sei kropuo nguba	Angami	speed_1.1
mhierhie hanu mia lhizeba puo nguba	Angami	original
mhierhie hanu mia lhizeba puo nguba	Angami	speed_0.9
mhierhie hanu mia lhizeba puo nguba	Angami	speed_1.1
kehouki u gei henbenji baptist church idi thupie ba	Angami	original
kehouki u gei henbenji baptist church idi thupie ba	Angami	speed_0.9
kehouki u gei henbenji baptist church idi thupie ba	Angami	speed_1.1
mhierhie hanu chüpahimia kropo vo rüzhüya kezha thechü ki puo nu ba kehou keba ngu ba	Angami	original
mhierhie hanu chüpahimia kropo vo rüzhüya kezha thechü ki puo nu ba kehou keba ngu ba	Angami	speed_0.9
mhierhie hanu chüpahimia kropo vo rüzhüya kezha thechü ki puo nu ba kehou keba ngu ba	Angami	speed_1.1
kieu ro phipra chü sie di sie sei geinu chü ba	Angami	original
kieu ro phipra chü sie di sie sei geinu chü ba	Angami	speed_0.9
kieu ro phipra chü sie di sie sei geinu chü ba	Angami	speed_1.1
kenie bagi puo se vor di dupie ba	Angami	original
kenie bagi puo se vor di dupie ba	Angami	speed_0.9
kenie bagi puo se vor di dupie ba	Angami	speed_1.1
sie kitie u ro kepejo kekra mu kemerie pie thepie ba	Angami	original
sie kitie u ro kepejo kekra mu kemerie pie thepie ba	Angami	speed_0.9
sie kitie u ro kepejo kekra mu kemerie pie thepie ba	Angami	speed_1.1
nyiepu po suru pie ketie gei dupie zha	Angami	original
nyiepu po suru pie ketie gei dupie zha	Angami	speed_0.9
nyiepu po suru pie ketie gei dupie zha	Angami	speed_1.1
mirhi hau liro garinya chüya kezha ki puo	Angami	original
mirhi hau liro garinya chüya kezha ki puo	Angami	speed_0.9
mirhi hau liro garinya chüya kezha ki puo	Angami	speed_1.1
tekhu sewa d telhako cü par kezhü susü nguba	Angami	original
tekhu sewa d telhako cü par kezhü susü nguba	Angami	speed_0.9
tekhu sewa d telhako cü par kezhü susü nguba	Angami	speed_1.1
kibu hau leshüda khaba ya kezha kibu puo	Angami	original
kibu hau leshüda khaba ya kezha kibu puo	Angami	speed_0.9
kibu hau leshüda khaba ya kezha kibu puo	Angami	speed_1.1
mhanyako liro si pie he chü pie keba ngu ba	Angami	original
mhanyako liro si pie he chü pie keba ngu ba	Angami	speed_0.9
mhanyako liro si pie he chü pie keba ngu ba	Angami	speed_1.1
hau nu nko khriesamie mie se ngu ba de	Angami	original
hau nu nko khriesamie mie se ngu ba de	Angami	speed_0.9
hau nu nko khriesamie mie se ngu ba de	Angami	speed_1.1
rüso kiu mh kh khenu mo te motha nipu nhasi ko ze di rüso kinu pie le lie keto puo ze bate	Angami	original
rüso kiu mh kh khenu mo te motha nipu nhasi ko ze di rüso kinu pie le lie keto puo ze bate	Angami	speed_0.9
rüso kiu mh kh khenu mo te motha nipu nhasi ko ze di rüso kinu pie le lie keto puo ze bate	Angami	speed_1.1
pesotsa colour kemerie pie ketso pie ba mu pekhro tsa kekra pie zhü pie ba	Angami	original
pesotsa colour kemerie pie ketso pie ba mu pekhro tsa kekra pie zhü pie ba	Angami	speed_0.9
pesotsa colour kemerie pie ketso pie ba mu pekhro tsa kekra pie zhü pie ba	Angami	speed_1.1
nhya kepejo kro puo rei puo khe puo mhu ki ba shimonyü puo sheiu nu liro kecie kro puo rei ba	Angami	original
nhya kepejo kro puo rei puo khe puo mhu ki ba shimonyü puo sheiu nu liro kecie kro puo rei ba	Angami	speed_0.9
nhya kepejo kro puo rei puo khe puo mhu ki ba shimonyü puo sheiu nu liro kecie kro puo rei ba	Angami	speed_1.1
puo mhodzü tsa rüso chü pie keba ngu	Angami	original
puo mhodzü tsa rüso chü pie keba ngu	Angami	speed_0.9
puo mhodzü tsa rüso chü pie keba ngu	Angami	speed_1.1
uko temvü color li ro brown kropo mou kekra kropo keze tha	Angami	original
uko temvü color li ro brown kropo mou kekra kropo keze tha	Angami	speed_0.9
uko temvü color li ro brown kropo mou kekra kropo keze tha	Angami	speed_1.1
pou ko ruo nyepu se vor thevie kehuki va kezevi tha	Angami	original
pou ko ruo nyepu se vor thevie kehuki va kezevi tha	Angami	speed_0.9
pou ko ruo nyepu se vor thevie kehuki va kezevi tha	Angami	speed_1.1
gari kropu thadi nguwa mu gari kropu tadi nguwa mu taxi kori nguwa mu alto gari silver color shie gari kekri kekri di ngwa mu pushenu ki kazha puo ri nguwa	Angami	original
gari kropu thadi nguwa mu gari kropu tadi nguwa mu taxi kori nguwa mu alto gari silver color shie gari kekri kekri di ngwa mu pushenu ki kazha puo ri nguwa	Angami	speed_0.9
gari kropu thadi nguwa mu gari kropu tadi nguwa mu taxi kori nguwa mu alto gari silver color shie gari kekri kekri di ngwa mu pushenu ki kazha puo ri nguwa	Angami	speed_1.1
hanu tent thenyie ki puo ngu wa nko mu tent colour liro grey	Angami	original
hanu tent thenyie ki puo ngu wa nko mu tent colour liro grey	Angami	speed_0.9
hanu tent thenyie ki puo ngu wa nko mu tent colour liro grey	Angami	speed_1.1
ha ki ha liro signboard kepezho kezha puo ngu bate	Angami	original
ha ki ha liro signboard kepezho kezha puo ngu bate	Angami	speed_0.9
ha ki ha liro signboard kepezho kezha puo ngu bate	Angami	speed_1.1
hanu mie kropo badi ngwu mi kropo shi tha di ngwa	Angami	original
hanu mie kropo badi ngwu mi kropo shi tha di ngwa	Angami	speed_0.9
hanu mie kropo badi ngwu mi kropo shi tha di ngwa	Angami	speed_1.1
hau shie nu ro ki kezha puo ngu wa hau mu gari kro puo ngu wa	Angami	original
hau shie nu ro ki kezha puo ngu wa hau mu gari kro puo ngu wa	Angami	speed_0.9
hau shie nu ro ki kezha puo ngu wa hau mu gari kro puo ngu wa	Angami	speed_1.1
hau nu liro hall nu te	Angami	original

End of preview. Expand in Data Studio

NE ASR Augmented Dataset -- Angami (njm)

Augmented automatic speech recognition dataset for Angami (njm), a Tibeto-Burman language spoken in Nagaland, India.

Source

Augmented from sulabhkatiyar/ne-asr-njm (original transcribed speech data from the ARTPARK-IISc Vaani project).

Language Information

Property	Value
Language	Angami
ISO 639-3	`njm`
Family	Tibeto-Burman
Region	Nagaland, India
Tonal	Yes
Tier	A (0.17h original data)

Dataset Statistics

Original training samples: 88
Augmented training samples: 264 (3x augmentation)
Train shards: 1
Estimated original duration: ~0.2 hours
Estimated augmented duration: ~0.5 hours

Split	Samples
train	264
validation	16
test	7

Transformations Applied

Each original training sample produces 3 samples (1 original + 2 speed + 0 pitch):

Speed perturbation: 0.9x, 1.1x (2 variants per sample)
Pitch shift: Disabled (tonal language -- pitch shift would alter lexical meaning)
Noise augmentation: Not applied

SpecAugment Parameters (for training, NOT in this dataset)

These parameters are consumed by the training script and are not baked into the audio files:

mask_time_prob: 0.03
mask_time_length: 10
mask_feature_prob: 0.0
mask_feature_length: 10
layerdrop: 0.0

Full augmentation config: configs/augmentation_config.yaml

Dataset Format

Audio: 16kHz mono WAV (stored as Parquet with audio bytes)
Text: Transcriptions
Features: audio, text, language, augmentation
Augmentation labels: original, speed_0.9, speed_1.1

How to Use

from datasets import load_dataset

# Load the full dataset
ds = load_dataset("sulabhkatiyar/ne-asr-njm-aug")

# Load only the training split
train = load_dataset("sulabhkatiyar/ne-asr-njm-aug", split="train")

# Filter to only original (non-augmented) samples
original_only = train.filter(lambda x: x["augmentation"] == "original")

# Filter to a specific augmentation type
speed_09 = train.filter(lambda x: x["augmentation"] == "speed_0.9")

Original Data

Source dataset: sulabhkatiyar/ne-asr-njm
Project: ARTPARK-IISc Vaani
License: CC-BY-4.0

Citation

If you use this dataset, please cite the Vaani project and acknowledge the augmentation pipeline.

Downloads last month: 25

Models trained or fine-tuned on sulabhkatiyar/ne-asr-njm-aug

sulabhkatiyar/ne-asr-njm

Automatic Speech Recognition • Updated May 29