Akan - UNICEF WCARO NLP

🌍 General Information

ISO 639-3aka

ISO 639-1ak

French Nameakan

Glottocodeakan1250

Language FamilyNiger–Congo

Writing SystemLatin

L1 Speakers8.9 million

L2 Speakers1 million

Population11000000 (10⁶+)

Endangermentsafe

CountriesBenin, Ghana

WikipediaAkan on Wikipedia

🗺️ Language Use

Data from CLEAR Global's Language Use Data Platform.

👥 Actors (4)

AdwumaTech AI (Startup)

AdwumaTech AI provides enterprise-grade AI training data infrastructure powered by African engineering talent. Core offerings include RLHF for code ge...

Farmerline / Darli AI (Startup)

Ghanaian agritech company building technology solutions for smallholder farmers. Core platform Mergdata is a cloud-based farm management system used b...

HCI and Persuasive Systems Lab, University of Ghana (University)

Research lab within the Department of Computer Science at University of Ghana, led by Prof. Isaac Wiafe. Focused on Human-Computer Interaction, persua...

UBC Deep Learning & NLP Lab (University)

Research lab at the University of British Columbia led by Muhammad Abdul-Mageed, focused on deep representation learning and natural language socio-pr...

🤖 NLP & Tech Resources

✓ NLLB-200

✓ MADLAD-400 Docs: 4800 MAD-Docs

✓ MADLAD-400 Sentences: 210200 MAD-Sents

✓ CMU Wilderness

✓ MAFAND

✓ MMS-ASR

✓ MMS-TTS

✓ MMS-LID

✓ Omnilingual ASR

✓ SimbaBench

✓ WaxalNLP

Benchmarks

ASR (Speech-to-Text)

SimbaBench (FinancialSpeech, Akuapem Twi)

Model	WER	CER	Source
facebook/mms-1b-all	85.89	38.51	reported
openai/whisper-large-v3	633.04	623.03	reported
openai/whisper-large-v3-turbo	249.56	243.12	reported
facebook/seamless-m4t-v2-large	230.19	198.78	reported
UBC-NLP/Simba-S	12.01	7.87	reported
UBC-NLP/Simba-M	16.94	8.01	reported
UBC-NLP/omniASR_LLM_7B_v2	89.16	47.42	reported
UBC-NLP/omniASR_LLM_1B_v2	92.84	61.5	reported
UBC-NLP/omniASR_LLM_300M_v2	106.37	71.87	reported
UBC-NLP/omniASR_CTC_7B_v2	85.29	41.92	reported
UBC-NLP/omniASR_CTC_3B_v2	89.32	41.35	reported
UBC-NLP/omniASR_CTC_1B_v2	87.21	43.37	reported

TTS (Text-to-Speech)

SimbaBench (bibleTTS, Akuapem Twi)

Model	WER	UTMOS	PESQ	Source
UBC-NLP/Simba-TTS	59.45	2.79	1.05	reported

MT (Machine Translation)

FLORES-200 devtest

Model	chrF++	direction	Source
facebook/nllb-200-3.3B	36.2	eng→aka	reported
facebook/nllb-200-3.3B	46.7	aka→eng	reported

FLORES+ 10-sample

Model	direction	SpBLEU	ChrF	Source
google/gemini-2.5-pro	aka→eng	30.2	54.8	reported
google/gemini-2.5-pro	eng→aka	17.1	47.2	reported
google/gemini-3-pro-preview	aka→eng	32.9	55.1	reported
google/gemini-3-pro-preview	eng→aka	17.0	48.5	reported
google/translate-v2	aka→eng	28.1	51.3	reported
google/translate-v2	eng→aka	17.5	46.4	reported

LLM (Large Language Models)

FLORES-200

Model	Log-perplexity	Source
goldfish-models/aka_latn_full	132.48	reported
MaLA-LM/mala-500-10b-v2	128.37	reported

ARC Synthetic 10-sample

Model	ARC Acc.	Source
anthropic/claude-3.7-sonnet	100.0	reported
anthropic/claude-opus-4.5	100.0	reported
anthropic/claude-opus-4.6	100.0	reported

MGSM 10-sample

Model	MGSM Acc.	Source
anthropic/claude-opus-4.6	88.9	reported
openai/gpt-5-mini	77.8	reported
openai/gpt-5.2	80.0	reported

MMLU 10-sample

Model	MMLU Acc.	Source
anthropic/claude-opus-4.6	75.0	reported
anthropic/claude-sonnet-4.6	87.5	reported
deepseek/deepseek-v3.2-exp	75.0	reported

SIB-200 10-sample

Model	Classification Acc.	Source
anthropic/claude-sonnet-4.5	100.0	reported
google/gemini-2.5-pro	100.0	reported
meta-llama/llama-3.1-70b-instruct	100.0	reported

📝 Notes

Observations

Major language in Ghana (Akan macrolanguage includes Twi and Fante)
Ghana has active UNICEF NLP work

Data

GhanaNLP Parallel Corpora — Includes 5,001 Fante-English sentence pairs (Fante is an Akan variety). Multi-domain, dialect-aware (Coastal vs Inland Fante). CC BY-NC-SA 4.0. Datasets on HuggingFace but not tagged with language IDs: https://huggingface.co/collections/Ghana-NLP/data-collection-project

ASR Models (21)

Model	Size	Downloads	Likes
MahmoudAshraf/mms-300m-1130-forced-aligner	0.3B	4.0M	83
facebook/mms-1b-all	1.0B	2.0M	195
facebook/mms-1b-fl102	1.0B	8.7K	29
UBC-NLP/Simba-S	2B	2.2K	4
facebook/mms-1b-l1107	1.0B	280	12
GiftMark/akan-whisper-model	0.2B	221	9
fiifinketia/whisper-large-v3-turbo-akan	—	77	3
UBC-NLP/Simba-W	2B	44	0
UBC-NLP/Simba-M	—	33	0
bc7ec356/heep-universal	0.8B	31	0
onnx-community/mms-300m-1130-forced-aligner-ONNX	—	25	0
johntsi/ZeroSwot-Medium_asr-cv_en-to-200	0.4B	17	0
johntsi/ZeroSwot-Large_asr-cv_en-to-200	0.4B	16	1
UBC-NLP/Simba-SLID-49	94.6M	15	0
johntsi/ZeroSwot-Medium_asr-mustc_en-to-200	0.4B	14	0
johntsi/ZeroSwot-Large_asr-mustc_en-to-200	0.4B	12	0
UBC-NLP/Simba-H	—	5	0
romara-labs/mms-300m-1130-forced-aligner-ONNX	—	4	0
UBC-NLP/Simba-X	—	3	0
VERUMNNODE/OS	—	0	0

View on HuggingFace: ak: 20 | aka: 1

TTS Models (6)

Model	Downloads	Likes
satzkumar/1232	0	0
antoniomae1234/voice-clona-text	0	2
Dapor/Dapson	0	4
cewaw/3D	0	0
jimmykis/unicornbanshee	0	0
Takerpilot/FLOXY	0	0

View on HuggingFace: ak: 6

MT Models (31)

Model	Size	Downloads	Likes
facebook/nllb-200-distilled-600M	—	1.1M	880
google/madlad400-3b-mt	3B	329.0K	192
facebook/nllb-200-distilled-1.3B	—	147.0K	181
facebook/nllb-200-3.3B	—	81.6K	443
jbochi/madlad400-3b-mt	3B	76.4K	154
facebook/nllb-200-1.3B	—	31.4K	71
Xenova/nllb-200-distilled-600M	—	4.4K	50
JustFrederik/nllb-200-distilled-600M-ct2-int8	—	925	1
facebook/nllb-moe-54b	—	557	123
JustFrederik/nllb-200-1.3B-ct2-int8	—	200	0
JustFrederik/nllb-200-distilled-600M-ct2-float16	—	157	0
winstxnhdw/nllb-200-distilled-1.3B-ct2-int8	—	141	6
JustFrederik/nllb-200-distilled-1.3B-ct2-int8	—	116	2
JustFrederik/nllb-200-distilled-600M-ct2	—	65	2
JustFrederik/nllb-200-3.3B-ct2-float16	—	52	3
michaelfeil/ct2fast-nllb-200-distilled-1.3B	—	47	1
michaelfeil/ct2fast-nllb-200-3.3B	—	43	14
vonjack/opus-mt-mul-en-big	0.3B	39	4
ychenNLP/nllb-200-3.3B-easyproject	—	36	8
JustFrederik/nllb-200-distilled-1.3B-ct2-float16	—	30	1

View on HuggingFace: ak: 30 | aka: 1

LLM Models (39)

Model	Size	Downloads	Likes
bigscience/bloomz-560m	0.6B	1.0M	137
bigscience/bloom-560m	0.6B	164.0K	371
bigscience/bloom-1b7	2B	49.4K	122
bigscience/bloom-3b	3B	14.1K	94
bigscience/bloomz-3b	3B	13.3K	81
bigscience/bloom-7b1	7B	10.4K	202
bigscience/bloom-1b1	1B	7.0K	66
bigscience/bloom	176B	5.9K	5.0K
bigscience/bloomz-7b1	7B	4.2K	147
bigscience/bloomz-7b1-mt	7B	4.1K	144
bigscience/bloom-7b1-intermediate	7B	3.7K	0
bigscience/bloomz-7b1-p3	7B	3.5K	6
bigscience/bloomz	176B	3.5K	513
bigscience/bloom-1b1-intermediate	1B	3.4K	0
bigscience/bloomz-mt	176B	3.3K	32
bigscience/bloomz-1b1	1B	2.9K	33
bigscience/bloomz-1b7	2B	2.8K	23
UBC-NLP/toucan-base	0.6B	2.5K	5
bigscience/bloom-1b7-intermediate	2B	2.4K	0
bigscience/bloom-560m-intermediate	0.6B	2.2K	1

View on HuggingFace: ak: 30 | aka: 9

Datasets

ASR Datasets (29)

Dataset	Rows	Downloads	Likes
google/WaxalNLP	2.6M	12.3K	219
sarulab-speech/yodas2_sidon	1.5M	6.0K	60
claudefitz/WaxalNLP	2.6M	3.6K	0
BolajiJsAI/WaxalNLP	2.6M	2.5K	1
ngoloan/WaxalNLP	2.0M	2.1K	0
novelwolde36/WaxalNLP	2.5M	2.0K	1
Phsntom/WaxalNLP	2.6M	1.9K	0
Ephraimmm/WaxalNLPr	2.6M	1.8K	0
fiifinketia/WaxalNLP	2.5M	1.7K	1
zoelaenskii/zinny	2.6M	1.4K	1
0xzanuee/WaxalNLP	1.9M	1.4K	0
suryatmodulus/yodas2_sidon	—	1.1K	0
UBC-NLP/SimbaBench_dataset	748.0K	975	0
WueNLP/sib-fleurs	94.4K	260	11
bdallhrajh371/WaxalNLP	2.6M	250	0
YuRiVeRTi/V1Q	186.0K	201	3
ghananlpcommunity/twi-speech-sota-240hrs	132.0K	147	0
Groovy-123/QuantumAI	—	27	0
Groovy-123/deep-think	—	23	1
jdapaah/asante-twi-bible	60	17	1

View on HuggingFace: ak: 15 | aka: 14

TTS Datasets (28)

Dataset	Rows	Downloads	Likes
google/WaxalNLP	2.6M	12.3K	219
sarulab-speech/yodas2_sidon	1.5M	6.0K	60
claudefitz/WaxalNLP	2.6M	3.6K	0
BolajiJsAI/WaxalNLP	2.6M	2.5K	1
ngoloan/WaxalNLP	2.0M	2.1K	0
novelwolde36/WaxalNLP	2.5M	2.0K	1
Phsntom/WaxalNLP	2.6M	1.9K	0
Ephraimmm/WaxalNLPr	2.6M	1.8K	0
fiifinketia/WaxalNLP	2.5M	1.7K	1
zoelaenskii/zinny	2.6M	1.4K	1
0xzanuee/WaxalNLP	1.9M	1.4K	0
suryatmodulus/yodas2_sidon	—	1.1K	0
UBC-NLP/SimbaBench_dataset	748.0K	975	0
WueNLP/sib-fleurs	94.4K	260	11
bdallhrajh371/WaxalNLP	2.6M	250	0
YuRiVeRTi/V1Q	186.0K	201	3
ghananlpcommunity/twi-speech-sota-240hrs	132.0K	147	0
Groovy-123/QuantumAI	—	27	0
hci-lab-dcug/bibletts-asante-twi-max29secs-total9hrs-sr22050	3.2K	26	1
Groovy-123/deep-think	—	23	1

View on HuggingFace: ak: 16 | aka: 12

Translation Datasets (41)

Dataset	Rows	Downloads	Likes
facebook/flores	—	43.4K	97
openlanguagedata/flores_plus	893.0K	15.8K	124
gtak1/panlex-meanings	64.1M	6.3K	0
google/smol	798.0K	2.8K	94
Muennighoff/flores200	—	1.6K	20
ayymen/Weblate-Translations	11.7M	383	18
Helsinki-NLP/opus_ubuntu	37.4K	354	4
mteb/FloresBitextMining	2.0K	303	1
visheratin/laion-coco-nllb	894.0K	255	44
cointegrated/panlex-meanings	157.0M	255	13
YuRiVeRTi/V1Q	186.0K	201	3
slone/nllb-200-10M-sample	10.0M	168	13
bible-nlp/biblenlp-corpus	—	164	37
mteb/biblenlp-corpus-mmteb	3.5M	146	2
Helsinki-NLP/qed_amara	—	114	5
mteb/flores	2.0K	96	0
FrancophonIA/gatitos	—	61	4
vpermilp/nllb-200-distilled-600M-rust	—	51	1
robinhad/long_flores	—	49	0
Svngoku/xP3x-Kongo	1.2M	30	4

View on HuggingFace: ak: 30 | aka: 11