Spaces:
Configuration error
Configuration error
Update README.md
Browse files
README.md
CHANGED
|
@@ -1,10 +1,115 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
---
|
| 2 |
-
|
| 3 |
-
|
| 4 |
-
|
| 5 |
-
|
| 6 |
-
|
| 7 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 8 |
---
|
| 9 |
|
| 10 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
# π¦ SynhalaAI
|
| 2 |
+
|
| 3 |
+
**Sinhala. Intelligence. Future.**
|
| 4 |
+
|
| 5 |
+
*Powered by skAI β SKY PRODUCTION*
|
| 6 |
+
|
| 7 |
---
|
| 8 |
+
|
| 9 |
+
## What is SynhalaAI?
|
| 10 |
+
|
| 11 |
+
SynhalaAI is Sri Lanka's first dedicated Artificial Intelligence initiative built exclusively for the Sinhala language.
|
| 12 |
+
|
| 13 |
+
This is not a single model. This is not a one-time experiment.
|
| 14 |
+
|
| 15 |
+
SynhalaAI is a long-term commitment to building the complete AI infrastructure that Sinhala has always deserved β models, datasets, benchmarks, tools, and open resources for every developer, researcher, and creator who wants to build with us.
|
| 16 |
+
|
| 17 |
+
> *"AI should work for everyone β in every language. SynhalaAI is our commitment to making that true for Sinhala."*
|
| 18 |
+
> β SKY PRODUCTION
|
| 19 |
+
|
| 20 |
---
|
| 21 |
|
| 22 |
+
## Why Sinhala?
|
| 23 |
+
|
| 24 |
+
**22 million people** speak Sinhala β yet the language has almost no dedicated support in modern AI technology.
|
| 25 |
+
|
| 26 |
+
Most of the world's leading AI systems are built for high-resource languages: English, Mandarin, Spanish. Languages like Sinhala receive little to no dedicated investment, creating a growing digital divide where entire communities are left behind as the world accelerates forward with AI.
|
| 27 |
+
|
| 28 |
+
SynhalaAI exists to close that gap.
|
| 29 |
+
|
| 30 |
+
---
|
| 31 |
+
|
| 32 |
+
## What We're Building
|
| 33 |
+
|
| 34 |
+
| Project | Status | Description |
|
| 35 |
+
|---|---|---|
|
| 36 |
+
| **SynhalaAI-Tharushi-3B** | π Coming Soon | Fine-tuned LLaMA 3.2 3B for Sinhala β runs offline on mobile |
|
| 37 |
+
| **SynhalaEes** | π In Progress | Sri Lanka's first dedicated Sinhala AI benchmark |
|
| 38 |
+
| **Sinhala Dataset** | π In Progress | 10Kβ100K instruction/response pairs for Sinhala NLP |
|
| 39 |
+
| **Sinhala GGUF Models** | π Coming Soon | Quantized models (q4_k_m, q5_k_m, q8_0) for mobile & edge devices |
|
| 40 |
+
|
| 41 |
+
---
|
| 42 |
+
|
| 43 |
+
## SynhalaAI-Tharushi-3B β First Model
|
| 44 |
+
|
| 45 |
+
Our first release under SynhalaAI is a fine-tuned **LLaMA 3.2 3B** model optimized for the Sinhala language using **Unsloth + QLoRA**.
|
| 46 |
+
|
| 47 |
+
### Key Facts
|
| 48 |
+
|
| 49 |
+
- π±π° **First offline Sinhala AI model** that runs on a mobile phone β no internet required
|
| 50 |
+
- π± **Mobile-friendly** β GGUF format for llama.cpp / Ollama
|
| 51 |
+
- π€ **Open source** β free for everyone under the SynhalaAI Open License
|
| 52 |
+
- π **Evaluated on SynhalaEes** β Sri Lanka's first Sinhala AI benchmark
|
| 53 |
+
|
| 54 |
+
### Capabilities
|
| 55 |
+
|
| 56 |
+
- β
Sinhala chat & instruction following
|
| 57 |
+
- β
Sinhala text generation
|
| 58 |
+
- β
Sinhala Q&A
|
| 59 |
+
- β
English β Sinhala translation
|
| 60 |
+
- β
Fully offline β no cloud, no internet required
|
| 61 |
+
|
| 62 |
+
|
| 63 |
+
## SynhalaEes Benchmark
|
| 64 |
+
|
| 65 |
+
**SynhalaEes** is Sri Lanka's first dedicated Sinhala language AI benchmark to measure and improve AI performance on the Sinhala language.
|
| 66 |
+
|
| 67 |
+
Because you can't improve what you don't measure.
|
| 68 |
+
|
| 69 |
+
> Benchmark results and methodology will be published alongside the first model release.
|
| 70 |
+
|
| 71 |
+
---
|
| 72 |
+
|
| 73 |
+
## License
|
| 74 |
+
|
| 75 |
+
All models and datasets released under the **SynhalaAI Open License v1.0**
|
| 76 |
+
|
| 77 |
+
```
|
| 78 |
+
β
Free to use β personal, research, and commercial
|
| 79 |
+
β
Free to modify and build upon
|
| 80 |
+
β
Attribution required β cite "SynhalaAI by SKY PRODUCTION"
|
| 81 |
+
β Cannot harm or misrepresent the Sinhala language or culture
|
| 82 |
+
β Cannot be rebranded as a new base model without disclosure
|
| 83 |
+
β SKY PRODUCTION branding cannot be used without permission
|
| 84 |
+
```
|
| 85 |
+
|
| 86 |
+
---
|
| 87 |
+
|
| 88 |
+
## Roadmap
|
| 89 |
+
|
| 90 |
+
- [x] Project launch β SynhalaAI brand & initiative
|
| 91 |
+
- [ ] SynhalaAI-Tharushi-3B model release (safetensors + GGUF)
|
| 92 |
+
- [ ] SynhalaEes benchmark publication
|
| 93 |
+
- [ ] Sinhala instruction dataset release
|
| 94 |
+
- [ ] SynhalaAI API for developers
|
| 95 |
+
- [ ] SynhalaAI v2 β expanded capabilities
|
| 96 |
+
- [ ] Community dataset contributions open
|
| 97 |
+
|
| 98 |
+
---
|
| 99 |
+
|
| 100 |
+
## Connect With Us
|
| 101 |
+
|
| 102 |
+
| Platform | Link |
|
| 103 |
+
|---|---|
|
| 104 |
+
| π GitHub | [github.com/SynhalaAI](https://github.com/SynhalaAI)
|
| 105 |
+
|
| 106 |
+
|
| 107 |
+
---
|
| 108 |
+
|
| 109 |
+
<div align="center">
|
| 110 |
+
|
| 111 |
+
**SynhalaAI** β Sinhala. Intelligence. Future.
|
| 112 |
+
|
| 113 |
+
*Powered by skAI β SKY PRODUCTION* π¦π±π°
|
| 114 |
+
|
| 115 |
+
</div>
|