Scaling Indic Pretraining: 2.8 Trillion Tokens and an Open Data Generation Pipeline
NVIDIA and AI4Bharat have released Nemotron‑pretraining‑Indic‑v1, a 2.8‑trillion‑token corpus covering 22 Indian languages, along with an open, GPU‑accelerated pipeline for generating and quality‑checking synthetic multilingual data. The resource aims to close the data gap for Indic NLP, offering native‑script and romanized text, and is freely available for developers to adapt and expand.