id string | speaker_id string | transcription string | language string | gender string | audio dict |
|---|---|---|---|---|---|
lin_24754 | 4rz2JZIZrUfKaxu5d9zdgrT2B1v1 | Ndaku oyo ezali na bosoto mpo bazali kosala ngo likolo ezali na mwinda misatu ekangami esika moko na se ezali na mabaya oyo ba matelaka. | lin | Female | {
"bytes": [
73,
68,
51,
4,
0,
0,
0,
0,
0,
34,
84,
83,
83,
69,
0,
0,
0,
14,
0,
0,
3,
76,
97,
118,
102,
54,
49,
46,
55,
46,
49,
48,
48,
0,
0,
0,
0,
0,
0,
0,
... |
lin_24796 | 4rz2JZIZrUfKaxu5d9zdgrT2B1v1 | Mobali afandi alati kanzaku,ya langi ya motane na singa ya kingo ya mayaka alati ekoti ya ba silamu. | lin | Female | {
"bytes": [
73,
68,
51,
4,
0,
0,
0,
0,
0,
34,
84,
83,
83,
69,
0,
0,
0,
14,
0,
0,
3,
76,
97,
118,
102,
54,
49,
46,
55,
46,
49,
48,
48,
0,
0,
0,
0,
0,
0,
0,
... |
lin_24732 | OWDZrFDhl8g2R3lZ7encCuWVrFz1 | "Opena nde nazomona ezali on dirait ezali pondu ya kotuta pondu yango kasi eza ya kotekisama kasi ya(...TRUNCATED) | lin | Female | {"bytes":"SUQzBAAAAAAAIlRTU0UAAAAOAAADTGF2ZjYxLjcuMTAwAAAAAAAAAAAAAAD/+5TEAAPAAAGkAAAAIAAANIAAAARMQU(...TRUNCATED) |
lin_24781 | OWDZrFDhl8g2R3lZ7encCuWVrFz1 | "oyo eza bamateriele oyo basalisaka mpona koyoka ndule babuffle babenga yango baffle yango kasi ezal(...TRUNCATED) | lin | Female | {"bytes":"SUQzBAAAAAAAIlRTU0UAAAAOAAADTGF2ZjYxLjcuMTAwAAAAAAAAAAAAAAD/+5TEAAPAAAGkAAAAIAAANIAAAARMQU(...TRUNCATED) |
lin_24790 | FzdvGMcP1mY2v86Jh5RDAqygN1p1 | "awa nazomona matiti pe embeni pe na katikate ya mai pe liboso na mosika nazomona ngomba moko eza na(...TRUNCATED) | lin | Female | {"bytes":"SUQzBAAAAAAAIlRTU0UAAAAOAAADTGF2ZjYxLjcuMTAwAAAAAAAAAAAAAAD/+5TEAAPAAAGkAAAAIAAANIAAAAQBRW(...TRUNCATED) |
lin_24700 | OWDZrFDhl8g2R3lZ7encCuWVrFz1 | "kasi boyo eza neti zando ozomona mituka kilikili esali mikakatano ya molongo kasi yango pembene ya (...TRUNCATED) | lin | Female | {"bytes":"SUQzBAAAAAAAIlRTU0UAAAAOAAADTGF2ZjYxLjcuMTAwAAAAAAAAAAAAAAD/+5TEAAPAAAGkAAAAIAAANIAAAARMQU(...TRUNCATED) |
lin_24718 | aBjQ2BKMCoUIeOIwENeRTCdC5vB2 | "elenge mibali oyo, bazo sala entrainement na stade, nde tina tomoni bango awa, aza gardien, mosusu (...TRUNCATED) | lin | Male | {"bytes":"SUQzBAAAAAAAIlRTU0UAAAAOAAADTGF2ZjYxLjcuMTAwAAAAAAAAAAAAAAD/+5TEAAAAAAGkFAAAJIvEXMM5sAD9jQ(...TRUNCATED) |
lin_24709 | ywYeqEpSB0fAQFMBPhaa0FAUXJi2 | "na liboso nabiso awa bazali bongo kolakisa biso eloko oyo na munoko ya lopoto ba bengaka poncarte (...TRUNCATED) | lin | Male | {"bytes":"SUQzBAAAAAAAIlRTU0UAAAAOAAADTGF2ZjYxLjcuMTAwAAAAAAAAAAAAAAD/+5TEAAPAAAGkAAAAIAAANIAAAAT6b6(...TRUNCATED) |
lin_24789 | aBjQ2BKMCoUIeOIwENeRTCdC5vB2 | "ndaku ya kitoko bazo lakisa biso awa baza na finissage bazo pakola ba langi na kamwa bassin ya vert(...TRUNCATED) | lin | Male | {"bytes":"SUQzBAAAAAAAIlRTU0UAAAAOAAADTGF2ZjYxLjcuMTAwAAAAAAAAAAAAAAD/+5TEAAASZXKMNPQAAuNEX6s60AAE1A(...TRUNCATED) |
lin_24762 | 9J0Xo69bgkhk1s5SeLdLH9UQQC22 | "Na elili oyo nazo mona bazol akisa biso lipa.\nLipa ba tié na kati ba pasoli na kati ba tié na ka(...TRUNCATED) | lin | Male | {"bytes":"SUQzBAAAAAAAIlRTU0UAAAAOAAADTGF2ZjYxLjcuMTAwAAAAAAAAAAAAAAD/+5TEAAPAAAGkAAAAIAAANIAAAARMQU(...TRUNCATED) |
WaxalNLP ASR — Cleaned Subset (Lingala, Shona, Luganda)
This dataset is a cleaned, corrected subset of google/WaxalNLP, covering the train and validation splits for three languages:
- lin_asr — Lingala
- sna_asr — Shona
- lug_asr — Luganda
The test split from the original dataset is intentionally excluded.
What was changed
The original transcriptions for these three languages contained a number of errors. A corrected transcription file was applied on top of the original audio, matched by example id. Corrections were merged in as follows:
| Language | Train | Validation |
|---|---|---|
| lin | 14,399 | 1,844 |
| sna | 14,109 | 1,727 |
| lug | 5,455 | 664 |
Coverage was effectively complete: every row in every split received a corrected transcription, with the exception of a single lin train example whose id had no corresponding entry in the correction file — that row was dropped rather than kept with a potentially unverified transcription.
Audio is unchanged from the original WaxalNLP release and is embedded directly in this dataset (not referenced by external path), so no separate audio download is required.
Fields
Each example contains:
id— unique example identifierspeaker_id— anonymized speaker identifiertranscription— corrected transcription textlanguage— language code (lin,sna, orlug)gender— speaker genderaudio— embedded audio (bytes+ originalpath)
Usage
from datasets import load_dataset
# Load a specific language
lin = load_dataset("Harcuracy/google_waxal_asr_challenge", "lin_asr")
sna = load_dataset("Harcuracy/google_waxal_asr_challenge", "sna_asr")
lug = load_dataset("Harcuracy/google_waxal_asr_challenge", "lug_asr")
train = lin["train"]
val = lin["validation"]
example = train[0]
print(example["transcription"])
print(example["audio"]) # {'bytes': ..., 'path': ...}
Source and license
This dataset is derived from google/WaxalNLP, released by Google under the CC-BY-4.0 license. The original Waxal project collected ASR and TTS data for African languages in partnership with Makerere University, the University of Ghana, Digital Umuganda, Media Trust, Loud and Clear, and AIMS Senegal, with funding from Google and the Gates Foundation.
This derivative work is released under the same CC-BY-4.0 license. Please cite the original WaxalNLP paper and dataset when using this data.
Limitations
- Only train and validation splits are included; no test split is provided here.
- Transcription corrections were applied based on a single external correction pass and have not been independently re-verified beyond automated id/split matching.
- Only three of WaxalNLP's languages are covered (Lingala, Shona, Luganda).
- Downloads last month
- 116