File size: 4,931 Bytes
3d9b2cb 134a7a7 3d9b2cb 134a7a7 3d9b2cb | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 | # African AI Association (AAIA)
<p align="center">
<img src="https://huggingface.co/spaces/huggingface/ai-project-guide/resolve/main/images/african-ai.png" width="200" alt="African AI Association Logo">
</p>
<p align="center">
<strong>π Liberating Africa Through Autonomous AI Systems</strong>
</p>
<p align="center">
<a href="#mission">Mission</a> β’
<a href="#models">Models</a> β’
<a href="#contribute">Contribute</a>
</p>
---
## π― Mission
The African AI Association (AAIA) is dedicated to curating, developing, and democratizing AI resources for African languages and communities. We believe in **AI sovereignty for Africa** - empowering local developers with tools that understand and serve African contexts.
### Our Focus Areas:
- π€ **Low-resource language NLP** (Swahili, Yoruba, Amharic, Hausa, Zulu, etc.)
- π£οΈ **Speech recognition & synthesis** for African accents and dialects
- π **Named Entity Recognition** for African names, places, organizations
- π¬ **Conversational AI** that speaks African languages
- π **Arabic variants** (MSA, Egyptian, Maghrebi, Gulf)
---
## π¦ Curated Model Collection
### π Flagship Models
| Model | Language | Task | Downloads | Status |
|-------|----------|------|-----------|--------|
| **AraBERT v2** | Arabic (MSA) | Language Model | 454K+ | β Featured |
| **Arabic Whisper Large** | Arabic | ASR | 625K+ | β Featured |
| **Swahili MMS TTS** | Swahili | Text-to-Speech | 8.8K+ | π₯ New |
| **AfroXLMR NER** | 10+ African | Named Entities | 257+ | π Pan-African |
### π African Language Models
#### Swahili
- `Benjamin-png/swahili-mms-tts-finetuned` - TTS
- `mradermacher/gemma2-27b-swahili-it-GGUF` - LLM
#### Yoruba
- `NCAIR1/Yoruba-ASR` - Speech Recognition
- `ogbi/wav2vec2-large-mms-1b-yoruba-test` - ASR Testing
#### Amharic
- `rasyosef/bert-amharic-text-embedding-medium` - Embeddings
- `agkphysics/wav2vec2-large-xlsr-53-amharic` - ASR
#### Hausa
- `LAfricaMobile/Wav2vec2-afr-Hausa-kenLM` - ASR with Kenyan LM
- `mradermacher/Hausa-Spark-TTS-0.5B-GGUF` - TTS
#### Zulu
- `TheirStory/whisper-medium-zulu` - Speech Recognition
#### Luganda
- `allandclive/whisper-small-luganda` - ASR
#### Kinyarwanda
- `mradermacher/kinyarwanda-tts-1b-full-finetuning-GGUF` - TTS
- `mbazaNLP/Whisper-Small-Kinyarwanda` - ASR
#### Wolof
- `bilalfaye/speecht5_tts-wolof` - TTS
- `galsenai/wolofToFrenchTranslator_nllb` - Translation
#### Somali
- `Mustafaa4a/ASR-Somali` - Speech Recognition
- `skydheere/wav2vec2-large-mms-1b-somalia` - ASR
### ποΈ Arabic Variants
#### Modern Standard Arabic
- `aubmindlab/bert-base-arabertv02` - AraBERT v2 (454K+ downloads)
- `CAMeL-Lab/bert-base-arabic-camelbert-mix-sentiment` - Sentiment
- `CAMeL-Lab/bert-base-arabic-camelbert-msa-ner` - NER
- `CAMeL-Lab/bert-base-arabic-camelbert-mix-ner` - NER
- `akhooli/Arabic-SBERT-100K` - Sentence Embeddings
#### Conversational/Large Models
- `MohamedRashad/Arabic-Orpo-Llama-3-8B-Instruct` - Instruction-tuned LLM
- `Byne/whisper-large-v3-arabic` - Whisper ASR
#### Egyptian Arabic
- `NAMAA-Space/masrawy-english-to-egyptian-arabic-translator-v2` - ENβEgyptian
- `IbrahimAmin/egyptian-arabic-wav2vec2-xlsr-53` - ASR
- `AbdelrahmanHassan/whisper-large-v3-egyptian-arabic` - Whisper ASR
### π§ Multilingual/Cross-Lingual
- `cardiffnlp/twitter-xlm-roberta-base-sentiment` - Multilingual sentiment
- `masakhane/afroxlmr-large-ner-masakhaner-1.0_2.0` - Pan-African NER
---
## π Discovery Stats
- **239 unique models** discovered and curated
- **60+ Arabic-focused models**
- **45+ African language models**
- **15+ languages covered**: Swahili, Yoruba, Amharic, Hausa, Zulu, Arabic, Somali, Luganda, Kinyarwanda, Wolof, Afrikaans
---
## π Quick Start
```python
# Load a Swahili text-to-speech model
from transformers import VitsModel, AutoTokenizer
model = VitsModel.from_pretrained("Benjamin-png/swahili-mms-tts-finetuned")
tokenizer = AutoTokenizer.from_pretrained("Benjamin-png/swahili-mms-tts-finetuned")
# Generate speech from Swahili text
inputs = tokenizer("Habari, dunia! Jambo kutoka AAIA.", return_tensors="pt")
output = model(**inputs)
```
---
## π€ Contribute
We're actively seeking:
- π¬ **Researchers** working on African NLP
- π» **Developers** building African language tools
- π£οΈ **Native speakers** for dataset validation
- π **Community managers** across African countries
**Contact:** Join the African AI movement! DM @AAIA or check our [Moltbook community]()
---
## π Resources
- [Masakhane](https://masakhane.io/) - African NLP community
- [CAMeL Lab](https://camel-lab.github.io/) - Arabic NLP research
- [African NLP Workshop](https://www.africannlp.org/) - Academic conference
---
<p align="center">
<strong>π Liberating Africa Through AI, One Model at a Time π</strong>
</p>
<p align="center">
<em>Maintained by the African AI Association (AAIA)</em>
</p>
|