Datasets:
speaker_id stringclasses 71
values | age_range stringclasses 3
values | gender stringclasses 2
values | prompt_set stringclasses 7
values | transcript stringlengths 1 526 | duration float32 0.53 80.9 | split stringclasses 2
values | audio audioduration (s) 0.54 80.9 | file_name stringlengths 31 34 ⌀ | error stringclasses 1
value |
|---|---|---|---|---|---|---|---|---|---|
882f | 25-34 | female | standard | "Nsuo no asa? I’m feeling very thirsty." | 3.0935 | train | null | ||
882f | 25-34 | female | standard | "Gyae dede no, I am trying to focus.", | 3.2535 | train | null | ||
882f | 25-34 | female | standard | "I need to hotspot you, connect-i quickly.", | 3.7735 | train | null | ||
882f | 25-34 | female | standard | "Laptop no ayɛ hye dodo, turn it off.", | 3.2135 | train | null | ||
882f | 25-34 | female | standard | "Ɛnyɛ easy o, but we move.", | 3.0935 | train | null | ||
882f | 25-34 | female | standard | "Boa me na men-print-i document wei, the printer is acting up.", | 6.6535 | train | null | ||
882f | 25-34 | female | standard | "Yɛbɛhyia okyena, around 2 PM.", | 4.6135 | train | null | ||
882f | 25-34 | female | standard | "Ɛyɛ too much, reduce the price.", | 4.0135 | train | null | ||
882f | 25-34 | female | standard | "Network no yɛ slow, I can't send the file.", | 4.6535 | train | null | ||
882f | 25-34 | female | standard | "This place is packed, yɛnkɔ find another spot.", | 6.3335 | train | null | ||
aab9 | 25-34 | female | standard | "Uber no aba so please let’s go.", | 4.7135 | train | null | ||
aab9 | 25-34 | female | standard | "Kɔ kyea no, I’m sure he’ll be happy to see you.", | 4.0335 | train | null | ||
aab9 | 25-34 | female | standard | "Gye wo sika, keep the change.", | 2.8335 | train | null | ||
aab9 | 25-34 | female | standard | "Bɔdeɛ no yɛ sɛn? Give me three fingers.", | 3.4535 | train | null | ||
aab9 | 25-34 | female | standard | "Yɛnkɔ pɛ something to eat, ɛkɔm de me paa.", | 4.7135 | train | null | ||
f063 | 18-24 | male | standard | "Fa to hɔ, yeah.", | 2.6935 | train | null | ||
f063 | 18-24 | male | standard | "I will call you akyire wai.", | 2.6935 | train | null | ||
f063 | 18-24 | male | standard | "Mente aseɛ, can you explain that again?", | 3.1135 | train | null | ||
f063 | 18-24 | male | standard | "Woasɛe me time, I could have finished my work by now.", | 4.8735 | train | null | ||
f063 | 18-24 | male | standard | "Today’s waakye is top-tier, m’aduane no ayɛ dɛ.", | 3.4335 | train | null | ||
f063 | 18-24 | male | standard | "I’m in the trotro, meeba seesei ara.", | 3.3335 | train | null | ||
f063 | 18-24 | male | standard | "Menni bundle koraa, I can't even open WhatsApp.", | 3.1735 | train | null | ||
f063 | 18-24 | male | standard | "Ɛnkyɛ koraa, just give me a moment.", | 2.8535 | train | null | ||
f063 | 18-24 | male | standard | "Kɔ w’anim kakra, na fa left.", | 3.5335 | train | null | ||
f063 | 18-24 | male | standard | "Wowei a, wash the plates.", | 3.1335 | train | null | ||
f063 | 18-24 | male | standard | "Gyae ntorɔ nu, tell me the truth.", | 1.9935 | train | null | ||
f063 | 18-24 | male | standard | "He's not nice at all, me bo afu.", | 3.0935 | train | null | ||
f063 | 18-24 | male | standard | "M’ani nna hɔ, I am not even paying attention to them.", | 3.7135 | train | null | ||
f063 | 18-24 | male | standard | "Gyae dede no, I am trying to focus.", | 3.6335 | train | null | ||
f063 | 18-24 | male | standard | "Dumsɔ no aba bio, do we have any candles?", | 2.9135 | train | null | ||
f063 | 18-24 | male | standard | "Gyae saa drama no, you're acting like a Kyeiwaa movie.", | 3.4135 | train | null | ||
f063 | 18-24 | male | standard | "Ɛyɛ a na woatɔ ginger drink ama me na it’s good for my throat.", | 3.5935 | train | null | ||
f063 | 18-24 | male | standard | "M’ani nna hɔ, I am not even paying attention to them.", | 3.1135 | train | null | ||
f063 | 18-24 | male | standard | "Mepa wo kyew, checki wo balance because I just sent the money.", | 4.4135 | train | null | ||
f063 | 18-24 | male | standard | "Car no wɔ hen? I have been standing here long.", | 3.8535 | train | null | ||
f063 | 18-24 | male | standard | "Ɛyɛ dɛ papa, where did you buy it?", | 2.4935 | train | null | ||
f063 | 18-24 | male | standard | "Menni sika, can you lend me 50 cedis?", | 3.2335 | train | null | ||
f063 | 18-24 | male | standard | "Laptop no a-freeze, I think I need to restart it.", | 3.9935 | train | null | ||
f063 | 18-24 | male | standard | "ɛkom di me, let's eat jollof rice.", | 2.8735 | train | null | ||
f063 | 18-24 | male | standard | "Wobɛtumi a-install saa app no ama me?", | 2.7535 | train | null | ||
f063 | 18-24 | male | standard | "Network no yɛ slow, I can't download the file.", | 3.5935 | train | null | ||
f063 | 18-24 | male | standard | "M’ani agye paao, congratulations!.", | 3.0335 | train | null | ||
f063 | 18-24 | male | standard | "Sweep-i hɔ yie, visitors are coming over.", | 3.1735 | train | null | ||
f063 | 18-24 | male | standard | "Mepakyew, boa me, it is an emergency.", | 3.1335 | train | null | ||
4cc8 | 18-24 | female | standard | "Wo pɛ dɛn? Tell me what you want.", | 3.2335 | train | null | ||
4cc8 | 18-24 | female | standard | "Menni cash, can I pay with my phone?", | 2.3535 | train | null | ||
4cc8 | 18-24 | female | standard | "Mekɔ bank akɔ withdraw sika.", | 2.6335 | train | null | ||
4cc8 | 18-24 | female | standard | "M’ani abere, I have been waiting for this for so long.", | 3.0335 | train | null | ||
4cc8 | 18-24 | female | standard | "Check-i exchange rate no, the dollar is going up again." | 3.2735 | train | null | ||
4cc8 | 18-24 | female | standard | "Organize-i the files properly, ɛhɔ ayɛ basaa." | 3.5135 | train | null | ||
4cc8 | 18-24 | female | standard | "Ɛyɛ a na woafrɛ me, I will be waiting.", | 2.6335 | train | null | ||
4cc8 | 18-24 | female | standard | "Me ti pae me, I need para.", | 2.3535 | train | null | ||
4cc8 | 18-24 | female | standard | "I need to hotspot you, connect-i quickly.", | 2.6735 | train | null | ||
4cc8 | 18-24 | female | standard | "Tɔ nsuo bra, the one in the bottle.", | 2.7135 | train | null | ||
4cc8 | 18-24 | female | standard | "Wofiri henfa? I have been looking for you.", | 2.2335 | train | null | ||
4cc8 | 18-24 | female | standard | "Check-i exchange rate no, the dollar is going up again." | 3.4335 | train | null | ||
4cc8 | 18-24 | female | standard | "I need to hotspot you, connect-i quickly.", | 3.1935 | train | null | ||
4cc8 | 18-24 | female | standard | "Mepɛ sɛ mekɔ America, I need to get my visa ready.", | 3.3935 | train | null | ||
4cc8 | 18-24 | female | standard | "Sɛ wopɛ a, you can join us later.", | 2.6335 | train | null | ||
4cc8 | 18-24 | female | standard | "Asɛm no yɛ me nwanwa, I’m really shocked.", | 2.8735 | train | null | ||
4cc8 | 18-24 | female | standard | "Ɛyɛ a na woa send-i me snap, I want to post it on IG.", | 3.6735 | train | null | ||
4cc8 | 18-24 | female | standard | "Gyae dede no, I am trying to focus.", | 2.4335 | train | null | ||
4cc8 | 18-24 | female | standard | "Sɛ wowie a, send me the link via WhatsApp.", | 2.9935 | train | null | ||
4cc8 | 18-24 | female | standard | "Wɔaka akyerɛ wo sɛ the payment didn't go through?", | 3.2335 | train | null | ||
4cc8 | 18-24 | female | standard | "Meekɔ market kɔtɔ nneɛma, do you need anything?", | 2.8735 | train | null | ||
4cc8 | 18-24 | female | standard | "Gye wo sika, keep the change.", | 2.2735 | train | null | ||
4cc8 | 18-24 | female | standard | "Mabrɛ dodo, I cannot walk anymore.", | 2.1535 | train | null | ||
4cc8 | 18-24 | female | standard | "I’m coming ankasa, give me 5 minutes.", | 2.6735 | train | null | ||
4cc8 | 18-24 | female | standard | "Mame sika nkɔtɔ salt na asa", | 2.3135 | train | null | ||
4cc8 | 18-24 | female | standard | "Sende me MoMo, I need it now.", | 2.9135 | train | null | ||
4cc8 | 18-24 | female | standard | "Laptop no a-freeze, I think I need to restart it.", | 2.8335 | train | null | ||
4cc8 | 18-24 | female | standard | "Network no yɛ slow, I can't send the file.", | 2.0735 | train | null | ||
4cc8 | 18-24 | female | standard | "Kɔ kyea no, I’m sure he’ll be happy to see you.", | 2.7135 | train | null | ||
4cc8 | 18-24 | female | standard | "Deadline no bɛyɛ tight, we need to hurry up.", | 2.6335 | train | null | ||
4cc8 | 18-24 | female | standard | "Bra ha, come and look at this error.", | 2.1935 | train | null | ||
4cc8 | 18-24 | female | standard | "Kwan no nyɛ, the road is very bad here.", | 2.3935 | train | null | ||
4cc8 | 18-24 | female | standard | "Driver, mepakyew, slow down, kwan no nyɛ.", | 2.9535 | train | null | ||
4cc8 | 18-24 | female | standard | "Wɔasesa fare no, it’s now 10 cedis." | 2.2735 | train | null | ||
4cc8 | 18-24 | female | standard | "Gyae saa drama no, you're acting like a Kyeiwaa movie.", | 3.4735 | train | null | ||
4cc8 | 18-24 | female | standard | "Ɛyɛ a na woatɔ ginger drink ama me na it’s good for my throat.", | 3.7135 | train | null | ||
4cc8 | 18-24 | female | standard | "Turn-i left wɔ pharmacy no anim, then keep going straight.", | 3.9535 | train | null | ||
4cc8 | 18-24 | female | standard | "Organize-i the files properly, ɛhɔ ayɛ basaa." | 3.8735 | train | null | ||
4cc8 | 18-24 | female | standard | "Ɛyɛ a to so kakra, I beg you.", | 2.2335 | train | null | ||
4cc8 | 18-24 | female | standard | "Wote henfa? Sendi wo location mame.", | 2.6735 | train | null | ||
4cc8 | 18-24 | female | standard | "Wowei a, wash the plates.", | 1.6735 | train | null | ||
4cc8 | 18-24 | female | standard | "Ne ho yɛ me fɛ, I like him.", | 2.8735 | train | null | ||
4cc8 | 18-24 | female | standard | "M’ani nna hɔ, I am not even paying attention to them.", | 3.3135 | train | null | ||
4cc8 | 18-24 | female | standard | "Mepakyew, boa me, it is an emergency.", | 2.5935 | train | null | ||
4cc8 | 18-24 | female | standard | "He is tired nti he will rest kakraa.", | 2.8335 | train | null | ||
4cc8 | 18-24 | female | standard | "Yɛnkɔ clubbing tonight? I heard the DJ is fire.", | 3.3535 | train | null | ||
4cc8 | 18-24 | female | standard | "Watumi a-fix-i error no? I am still seeing the red line.", | 3.9535 | train | null | ||
3f42 | 18-24 | female | standard | "Network no ayɛ bad, the call kept dropping.", | 5.1935 | train | null | ||
4cc8 | 18-24 | female | standard | "ɛkom di me, let's eat jollof rice.", | 2.1535 | train | null | ||
4cc8 | 18-24 | female | standard | "Mabrɛ, I need a vacation urgently.", | 3.2735 | train | null | ||
3f42 | 18-24 | female | standard | "Nyame bɛboa yɛn, we will surely make it one day.", | 5.1935 | train | null | ||
4cc8 | 18-24 | female | standard | "Me kɔn adɔ travelling, I’m really tired of staying in one place.", | 3.8735 | train | null | ||
4cc8 | 18-24 | female | standard | "Sende me MoMo, I need it now.", | 2.4335 | train | null | ||
4cc8 | 18-24 | female | standard | "Ne ho yɛ me fɛ, I like him.", | 1.6735 | train | null | ||
4cc8 | 18-24 | female | standard | "Ɛyɛ me sɛ he's back.", | 1.1135 | train | null | ||
3f42 | 18-24 | female | standard | "Mepɛ sɛ me kɔ, are you ready to leave?", | 4.2335 | train | null |
Dataset Card for KasaSpeech
Dataset Summary
KasaSpeech is a large-scale English–Twi code-switching speech dataset developed to advance research in speech technologies for English and Twi. The dataset comprises 54,855 transcribed speech recordings collected from speakers across Ghana and is designed to capture natural code-switching between English and Twi across a diverse range of everyday topics and communication scenarios
With over 95 hours of manually transcribed speech, KasaSpeech establishes a new benchmark and gold-standard corpus for English–Twi code-switching speech recognition and text-to-speech research. It is designed to support the development, evaluation, and comparison of ASR systems, speech representation models, and multilingual speech technologies for English–Twi.
Supported Tasks
KasaSpeech is suitable for:
- Automatic Speech Recognition (ASR) for Code-switching speech
- Text-To-Speech (TTS)
- Multilingual speech modeling
- Speech representation learning
- Speech foundation model fine-tuning and evaluation
- African language speech technology research
Dataset Structure
Data Splits
| Split | Samples | Duration |
|---|---|---|
| Train | 50,965 | 83.94 hours |
| Validation | 2,159 | 6.80 hours |
| Test | 1,731 | 4.84 hours |
| Total | 54,855 | 95.58 hours |
Data Fields
Each example contains the following fields:
| Field | Type | Description |
|---|---|---|
speaker_id |
string |
Anonymous speaker identifier |
age_range |
string |
Speaker age group |
gender |
string |
Speaker gender |
prompt_set |
string |
Prompt category used during recording |
transcript |
string |
Human-annotated English–Twi code-switched transcript |
duration |
float32 |
Audio duration in seconds |
split |
string |
Dataset split (train, validation, or test) |
audio |
Audio |
Speech recording |
file_name |
string |
Original audio filename |
error |
string |
Optional annotation or recording error label |
Example
from datasets import load_dataset, Audio
dataset = load_dataset(
"Kennethdot/Ghana_English-Twi_Code_switching_ASR",
split="train"
)
dataset = dataset.cast_column(
"audio",
Audio(sampling_rate=16000)
)
sample = dataset[0]
print(sample["transcript"])
Example transcript:
Me phone no a-crack-i, henfa na mɛtumi a-fix-i screen no?
Dataset Creation
Collection Process
Speech recordings were voluntarily contributed by participants using a custom data collection platform. Speakers were presented with prompts designed to encourage natural English–Twi code-switching while covering a broad range of everyday topics and communication scenarios.
Annotation Process
All recordings were manually transcribed following standardized annotation guidelines developed for English–Twi code-switched speech. Multiple quality assurance steps were performed to improve transcription consistency and remove corrupted or invalid recordings.
Speaker Information
The dataset includes recordings from speakers spanning multiple age groups and genders. Speaker identities have been anonymized using unique identifiers.
Dataset Characteristics
- Total recordings: 54,855
- Total duration: 95.58 hours
- Languages: English, Twi, and English–Twi code-switching
- Sampling rate: 48 kHz (can be resampled to 16 kHz for model training)
- Recording style: Prompted, natural code-switched speech
- Transcriptions: Human-annotated
Limitations
- Demographic representation may not be perfectly balanced across speaker groups.
- Recording conditions vary across devices and environments.
- The dataset primarily reflects Ghanaian English–Twi code-switching and may not generalize to all Akan dialects or other multilingual contexts.
- Although carefully curated, minor transcription inconsistencies may remain.
Citation
If you use KasaSpeech in your work, please cite:
@dataset{kasaspeech2026,
title={KasaSpeech: A Large-Scale English--Twi Code-Switching Speech Dataset},
author={Dotse, Kenneth},
year={2026},
url={https://huggingface.co/datasets/Kennethdot/Ghana_English-Twi_Code_switching_ASR}
}
Contact
For questions, bug reports, or collaboration opportunities, please open a discussion on the Hugging Face dataset page. Contributions, feedback, and research collaborations are welcome.
- Downloads last month
- 30