metadata
license: apache-2.0
Amharic Sentences Corpus V1.0
Source: Telegram
This 1.6 million Amharic sentences corpus reflects current Amharic usage as of December 20, 2025, and is designed for anyone interested in:
- Training Amharic-based LLMs
- Fine-tuning NLP models
- Building search, summarization, or generative systems in Amharic
The dataset is heavily cleaned and normalized, but like any serious LLM dataset, it still needs proper tokenization for pre-training.
I recommend using an Amharic-specific tokenizer such as: https://pypi.org/project/amharic-tokenizer/0.2.6
Another useful Amharic Sentences corpus by @rasyosef & @Addis AI
How did I create this text corpus?
- I vibe coded a simple yet powerful script that creates a sentences from a .Json file. If you Want to create your own text corpus, feel free to use my script. https://github.com/ababiyaworku/HuggingFace_dataset_creator
This project is fully open-source & community-driven. If you’re building in NLP, AI research, or language technology, this is for you.
Use it. Improve it. Build on top of it.