Buckets:

---

adab-tech

/

Code-170k-sango-bucket

Dataset Description

Code-170k-sango is a groundbreaking dataset containing 176,999 programming conversations, originally sourced from glaiveai/glaive-code-assistant-v2 and translated into Sango, making coding education accessible to Sango speakers.

🌟 Key Features

• 176,999 high-quality conversations about programming and coding
• Pure Sango language - democratizing coding education
• Multi-turn dialogues covering various programming concepts
• Diverse topics: algorithms, data structures, debugging, best practices, and more
• Ready for instruction tuning of Large Language Models

🎯 Use Cases

• Training Sango-language coding assistants
• Building educational tools for Sango developers
• Researching multilingual code generation
• Creating programming tutorials in Sango
• Supporting low-resource language AI development

Dataset Structure

Data Fields

• conversations: A list of conversation turns, where each turn contains:
  • from: The speaker ("human" or "gpt")
  • value: The message content in Sango

Example

{
  "conversations": [
    {
      "from": "human",
      "value": "[Question in Sango]"
    },
    {
      "from": "gpt",
      "value": "[Answer in Sango]"
    }
  ]
}

Usage

Loading the Dataset

from datasets import load_dataset

# Load the dataset
dataset = load_dataset("michsethowusu/Code-170k-sango")

# Access training data
train_data = dataset['train']

# Example: Print first conversation
for turn in train_data[0]['conversations']:
    print(f"{turn['from']}: {turn['value']}")

Citation

@dataset{code170k_sango,
  title={Code-170k-sango: Programming Conversations in Sango},
  year={2025},
  publisher={Hugging Face},
  url={https://huggingface.co/datasets/michsethowusu/Code-170k-sango}
}

License

This dataset is released under the Apache 2.0 License.

Thank you for using Code-170k-sango to advance programming education in Sango! 🌍✨