Downloads · 30 days
47
15% of all-time downloads
BaobabAI-v0.2
okaforpascal40/BaobabAI-v0.2
BaobabAI-v0.2 is a machine learning model from okaforpascal40. Use it for the machine learning task on the model card, and read the license before you ship it in a product. The card lists the license as apache-2.0.
Africa's first continental AI model — now with 15 languages and 86,604 training pairs.
All-time downloads
304
Public
Parameters
3.2B
6.4 GB on disk
Likes
0
Public
Repo makeup
Click a slice to open those files.
.safetensors6.4 GB · 100%
At a glance
- License
- apache-2.0
- Model type
- llama
- Access
- Public
- Created
- Mar 7, 2026
- Updated
- Mar 7, 2026
- SHA
- 2e90ab71
Model card
From the Hugging Face model README
🌳 BaobabAI v0.2
Africa's first continental AI model — now with 15 languages and 86,604 training pairs.
Named after the Baobab tree — the Tree of Life across Africa. Lives 2,000+ years. Feeds entire ecosystems. That is the vision.
🚀 What's New in v0.2
- 3x more training data — 86,604 training pairs
- 5x more training steps — 500 steps
- 4 new languages added
- Training loss improved: 2.81 → 2.07
- New data sources: Glot500 + CulturaX + Masakhane
🌍 Languages Supported
| Language | Country | Speakers |
|---|---|---|
| Hausa | Nigeria/Niger/Chad | 70M+ |
| Yoruba | Nigeria/Benin/Togo | 45M+ |
| Igbo | Nigeria | 30M+ |
| Nigerian Pidgin | West Africa | 75M+ |
| Swahili | Kenya/Tanzania | 200M+ |
| Amharic | Ethiopia | 35M+ |
| Somali | Somalia/Kenya | 20M+ |
| Xhosa | South Africa | 27M+ |
| Zulu | South Africa | 27M+ |
| Shona | Zimbabwe | 15M+ |
| Luganda | Uganda | 8M+ |
| Lingala | DRC/Congo | 70M+ |
| Wolof | Senegal/Gambia | 12M+ |
| Oromo | Ethiopia/Kenya | 40M+ |
📊 Training Details
| Parameter | Value |
|---|---|
| Base Model | Llama 3.2 3B Instruct |
| Training Pairs | 86,604 |
| Training Steps | 500 |
| Final Loss | 2.07 |
| Method | QLoRA (r=16) via Unsloth |
| Hardware | NVIDIA Tesla T4 |
| Data Sources | Masakhane, Glot500, CulturaX |
💡 What BaobabAI Can Do
- Summarize African news articles in local languages
- Identify African languages automatically
- Answer questions about African topics
- Process Nigerian Pidgin English natively
- Classify African content by category
🔧 Quick Start
from transformers import AutoModelForCausalLM, AutoTokenizer
tokenizer = AutoTokenizer.from_pretrained("okaforpascal40/BaobabAI-v0.2")
model = AutoModelForCausalLM.from_pretrained("okaforpascal40/BaobabAI-v0.2")
🗺️ Roadmap
- v0.3 — 20 languages, 200,000+ training pairs
- v0.5 — Upgrade to 8B model
- v1.0 — Live REST API, enterprise ready
- v2.0 — Continental dominance, 50+ languages
👨🏿💻 Built By
Pascal Okafor Ogbonna | SabiFlow Technologies Limited 🌍 baobabai.dev · GitHub
📜 License
Apache 2.0 — free to use, modify and build on.
The tree is planted. The continent will feel its shade. 🌳
- Type
- llama
- License
- apache-2.0
- Languages
- ha, ig, pcm, yo, sw, am
- Created
- Mar 7, 2026
- Updated
- Mar 7, 2026