Browse Datasets

AfriInstruct

Instruction Tuning for African Languages

instruction-tuning fine-tuning llm Millions of tokens CC-BY-4.0
Access Dataset

Description

Dataset specialized for instruction-tuning multiple African languages. Bridges the gap between high-resource and African language LLM performance.

Quick Start

Python 
from datasets import load_dataset

dataset = load_dataset("llama-lang-adapt/AfriInstruct-Data", split="train")
print(dataset['instruction'][:3], dataset['output'][:3])

Project Ideas

Hausa Instruction Following

Fine-tune Llama-2-7B to follow instructions in Hausa

Advanced

Cross-lingual Tax

Evaluate cross-lingual instruction tuning tax

Advanced

Ethical Considerations

  • Derived from other datasets; inherits their biases