Gemma2-9B-Swahili-IT

Gemma2-9B-Swahili-IT is a state-of-the-art open variant of Google's Gemma2-9B-IT model, fine-tuned for natural Swahili language understanding and generation. This model demonstrates strong performance across various language tasks while maintaining efficient resource usage.

Model Details

• Developer: Alfaxad Eyembe
• Base Model: google/gemma-2-9b-it
• Model Type: Decoder-only transformer
• Language(s): Swahili
• License: Apache 2.0
• Finetuning Approach: Low-Rank Adaptation (LoRA)

Training Data

The model was fine-tuned on a comprehensive dataset containing:

• 67,017 instruction-response pairs
• 16,273,709 total tokens
• Average 242.83 tokens per example
• High-quality, naturally-written Swahili content

Performance

Massive Multitask Language Understanding (MMLU) - Swahili

• Base Model: 45.61% accuracy
• Fine-tuned Model: 52.63% accuracy
• Improvement: +7.02%

Sentiment Analysis - Swahili

• Base Model: 84.85% accuracy
• Fine-tuned Model: 86.00% accuracy
• Improvement: +1.15%
• Perfect response validity (100%)

Intended Use

This model is designed for:

• Natural Swahili text generation
• Question answering
• Sentiment analysis
• Creative writing
• General instruction following in Swahili

Usage

from transformers import AutoTokenizer, AutoModelForCausalLM
import torch

# Load tokenizer and model
tokenizer = AutoTokenizer.from_pretrained("alfaxadeyembe/gemma2-9b-swahili-it")
model = AutoModelForCausalLM.from_pretrained(
    "alfaxadeyembe/gemma2-9b-swahili-it",
    device_map="auto",
    torch_dtype=torch.bfloat16
)

# Always set to eval mode for inference
model.eval()

# Example usage
prompt = "Eleza dhana ya uchumi wa kidijitali na umuhimu wake katika ulimwengu wa leo."
inputs = tokenizer(prompt, return_tensors="pt").to(model.device)

with torch.no_grad():
    outputs = model.generate(
        **inputs,
        max_new_tokens=500,
        do_sample=True,
        temperature=0.7,
        top_p=0.95
    )

response = tokenizer.decode(outputs[0], skip_special_tokens=True)
print(response)

Training Details

• Fine-tuning Method: LoRA
• Training Steps: 400
• Batch Size: 2
• Gradient Accumulation Steps: 32
• Learning Rate: 2e-4
• Training Time: ~12 hours on A100 GPU

Key Features

• Strong performance on structured tasks
• Natural Swahili language generation
• Balanced technical and conversational capabilities
• Efficient parameter updates through LoRA
• Improved response coherence and completion

Citation

@misc{gemma2-9b-swahili-it,
  author = {Alfaxad Eyembe},
  title = {Gemma2-9B-Swahili-IT: SWahili Variation For Gemma2-9B-IT},
  year = {2025},
  publisher = {Hugging Face},
  journal = {Hugging Face Model Hub},
}

Contact

For questions or feedback, please reach out through:

• HuggingFace: @alfaxadeyembe
• Twitter: @alfxad