The largest open, multi-domain Bangla text dataset, combining 801,645 samples from 15 different sources — newspapers, social media, education, reviews, QA, medical, poetry, and more. Ready for Bangla LLM pretraining, fine-tuning, and downstream NLP tasks.
Bangla (Bengali) is the 7th most spoken language in the world with 230M+ speakers, but high-quality Bangla NLP… See the full description on the dataset page:
https://huggingface.co/datasets/nahidstaq/bangla-llm-data.
