Upload .\src\bwenge_ai.egg-info\PKG-INFO with huggingface_hub

.//src//bwenge_ai.egg-info//PKG-INFO

@@ -0,0 +1,178 @@
+Metadata-Version: 2.4
+Name: bwenge-ai
+Version: 0.1.0
+Summary: BwengeAi - Rwandan AI that knows Kinyarwanda
+License: MIT
+Requires-Python: >=3.10
+Description-Content-Type: text/markdown
+Requires-Dist: torch>=2.1.0
+Requires-Dist: transformers>=4.36.0
+Requires-Dist: datasets>=2.16.0
+Requires-Dist: accelerate>=0.25.0
+Requires-Dist: peft>=0.7.0
+Requires-Dist: bitsandbytes>=0.41.0
+Requires-Dist: trl>=0.7.0
+Requires-Dist: huggingface-hub>=0.20.0
+Requires-Dist: sentencepiece>=0.1.99
+Requires-Dist: tokenizers>=0.15.0
+Requires-Dist: safetensors>=0.4.0
+Requires-Dist: beautifulsoup4>=4.12.0
+Requires-Dist: requests>=2.31.0
+Requires-Dist: lxml>=4.9.0
+Requires-Dist: tqdm>=4.66.0
+Requires-Dist: pyyaml>=6.0
+Requires-Dist: python-dotenv>=1.0.0
+Requires-Dist: numpy>=1.24.0
+Provides-Extra: dev
+Requires-Dist: pytest>=7.4.0; extra == "dev"
+Requires-Dist: pytest-cov>=4.1.0; extra == "dev"
+Requires-Dist: ruff>=0.1.0; extra == "dev"
+Requires-Dist: mypy>=1.7.0; extra == "dev"
+
+# BwengeAi - Rwandan AI that knows Kinyarwanda
+
+BwengeAi is a Rwandan AI model trained on Kinyarwanda language data from multiple sources including Huggingface datasets, Kinyarwanda Wikipedia, and Rwandan news content.
+
+## Features
+
+- **Multi-source data collection** from Huggingface, Wikipedia, and Igihe
+- **Kinyarwanda language support** with comprehensive NLP capabilities
+- **Efficient fine-tuning** using LoRA for parameter-efficient training
+- **Evaluation pipeline** with perplexity, ROUGE, BLEU, and accuracy metrics
+
+## Project Structure
+
+```
+BwengeAi/
+├── configs/
+│   └── default.yaml          # Main configuration
+├── src/
+│   ├── data_collection/      # Data collection modules
+│   │   ├── huggingface_collector.py
+│   │   ├── wikipedia_collector.py
+│   │   ├── igihe_scraper.py
+│   │   └── data_processor.py
+│   ├── models/               # Model architecture
+│   │   └── bwenge_model.py
+│   ├── training/             # Training pipeline
+│   │   └── trainer.py
+│   └── evaluation/           # Evaluation metrics
+│       └── metrics.py
+├── scripts/
+│   ├── collect_all.py        # Data collection script
+│   └── train.py              # Training script
+├── data/
+│   ├── raw/                  # Raw collected data
+│   └── processed/            # Processed training data
+├── tests/                    # Test files
+└── docs/                     # Documentation
+```
+
+## Installation
+
+```bash
+# Clone the repository
+git clone <repository-url>
+cd BwengeAi
+
+# Create virtual environment
+python -m venv venv
+source venv/bin/activate  # On Windows: venv\Scripts\activate
+
+# Install dependencies
+pip install -e ".[dev]"
+```
+
+## Data Sources
+
+### Huggingface Datasets
+- **CircuitNotion/kinyarwanda_corpus** - Large Kinyarwanda text corpus (3.52M rows)
+- **mbazaNLP/kinyarwanda_monolingual_v01.1** - Monolingual corpus (1.07M rows)
+- **saillab/alpaca_kinyarwanda_taco** - Instruction tuning (62k rows)
+- **mbazaNLP/Kinyarwanda_English_parallel_dataset** - Translation pairs (55.7k rows)
+- Plus 100+ additional datasets
+
+### Kinyarwanda Wikipedia
+- 9,660 articles with 3M+ words
+- API access for real-time content
+- Dump files for bulk download
+
+### Igihe News
+- Rwanda's leading online news platform
+- Multiple categories: politics, health, sports, technology
+- **Note:** Content is copyrighted - contact info@igihe.com for licensing
+
+## Usage
+
+### 1. Collect Data
+
+```bash
+python scripts/collect_all.py
+```
+
+This will:
+- Download datasets from Huggingface
+- Fetch articles from Kinyarwanda Wikipedia
+- Attempt limited Igihe scraping (with permission)
+- Process and clean all collected data
+
+### 2. Train Model
+
+```bash
+python scripts/train.py
+```
+
+This will:
+- Load the base model (Llama-3 or Llama-2)
+- Set up LoRA for efficient fine-tuning
+- Train on the collected Kinyarwanda data
+- Evaluate the trained model
+
+### 3. Configuration
+
+Edit `configs/default.yaml` to customize:
+- Model selection and parameters
+- Training hyperparameters
+- Data sources and processing options
+- Evaluation metrics
+
+## Model Options
+
+### Base Models
+- `meta-llama/Llama-3.2-1B` (default, small)
+- `almanach/Llama-3-8B-mono-Kinyarwanda` (8B, Kinyarwanda-specific)
+- `almanach/Llama-2-7B-mono-Kinyarwanda` (7B, Kinyarwanda-specific)
+
+### Training Modes
+- **Full fine-tuning**: Train all parameters
+- **LoRA fine-tuning**: Train only adapter layers (recommended)
+- **QLoRA**: Quantized LoRA for memory efficiency
+
+## Evaluation
+
+The evaluation pipeline includes:
+- **Perplexity**: Language model quality
+- **ROUGE**: Text summarization quality
+- **BLEU**: Translation quality
+- **Accuracy**: Exact match accuracy
+
+## Available Kinyarwanda Models on Huggingface
+
+| Model | Type | Size |
+|-------|------|------|
+| almanach/Llama-3-8B-mono-Kinyarwanda | LLM | 8B |
+| almanach/Llama-2-7B-mono-Kinyarwanda | LLM | 7B |
+| RogerB/kinyaRoberta-large-pretrained | Encoder | Large |
+| RogerB/KinyaBERT-small-pretrained | Encoder | Small |
+| leophill/whisper-large-v3-sn-kinyarwanda | ASR | 2B |
+| mbazaNLP/Whisper-Small-Kinyarwanda | ASR | Small |
+
+## License
+
+MIT License
+
+## Contact
+
+For questions about Igihe data licensing:
+- Email: info@igihe.com
+- Phone: +250 788 895 953