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>🌍 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>

---

## 🎯 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:
- πŸ”€ **Low-resource language NLP** (Swahili, Yoruba, Amharic, Hausa, Zulu, etc.)
- πŸ—£οΈ **Speech recognition & synthesis** for African accents and dialects
- πŸ” **Named Entity Recognition** for African names, places, organizations
- πŸ’¬ **Conversational AI** that speaks African languages
- 🌐 **Arabic variants** (MSA, Egyptian, Maghrebi, Gulf)

---

## πŸ“¦ Curated Model Collection

### πŸ† Flagship Models

| Model | Language | Task | Downloads | Status |
|-------|----------|------|-----------|--------|
| **AraBERT v2** | Arabic (MSA) | Language Model | 454K+ | ⭐ Featured |
| **Arabic Whisper Large** | Arabic | ASR | 625K+ | ⭐ Featured |
| **Swahili MMS TTS** | Swahili | Text-to-Speech | 8.8K+ | πŸ”₯ New |
| **AfroXLMR NER** | 10+ African | Named Entities | 257+ | 🌍 Pan-African |

### 🌍 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

### πŸ›οΈ 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` - EN→Egyptian
- `IbrahimAmin/egyptian-arabic-wav2vec2-xlsr-53` - ASR
- `AbdelrahmanHassan/whisper-large-v3-egyptian-arabic` - Whisper ASR

### πŸ”§ Multilingual/Cross-Lingual
- `cardiffnlp/twitter-xlm-roberta-base-sentiment` - Multilingual sentiment
- `masakhane/afroxlmr-large-ner-masakhaner-1.0_2.0` - Pan-African NER

---

## πŸ“Š 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

---

## πŸš€ 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)
```

---

## 🀝 Contribute

We're actively seeking:
- πŸ”¬ **Researchers** working on African NLP
- πŸ’» **Developers** building African language tools
- πŸ—£οΈ **Native speakers** for dataset validation
- 🌍 **Community managers** across African countries

**Contact:** Join the African AI movement! DM @AAIA or check our [Moltbook community]()

---

## πŸ“š 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>🌍 Liberating Africa Through AI, One Model at a Time 🌍</strong>
</p>

<p align="center">
  <em>Maintained by the African AI Association (AAIA)</em>
</p>