Sarvam AI releases Vāgartha: 217,959 Sanskrit verses with detailed explanations on Hugging Face
selfawareatom · x · 2026-09-04
Sarvam AI has released Vāgartha on Hugging Face — a dataset of 217,959 Sanskrit verses, each paired with a detailed explanation. The name means "united as word and meaning."
- Scale & format: 100K–1M text entries, parquet format, CC-BY-4.0 license, covering Devanagari and English
- Tasks: text generation, translation, question answering
- Content: lexicographical-style verses defining words with part of speech, multiple senses, and cultural context (e.g., Śobhā meaning mirage, splendor, or desire)
A valuable high-quality corpus for low-resource language modeling, classical Indic NLP, and translation research.
More from Research
- Apodex launches TRACES, a benchmark measuring whether AI can rigorously explore the unknown — eyishazyer · 2026-09-04
- Diffusion-augmented LLM Uno: 8B beats 26B DiffusionGemma with 3x lossless speedup — iScienceLuvr · 2026-09-04
- Scal3R Uses Multi-Reference Pose Queries to Cut Drift in Online 3D Reconstruction — Chin-Yang Lin · 2026-09-04
- AI BioDesign accelerator launches with UW Medicine and Fred Hutch to let AI design biology — AllThingsApx · 2026-09-04
- ORNL's AI system arranges molecules atom-by-atom, working 25+ hours unattended — Scobleizer · 2026-09-04
- zackmdavis pitches embryo selection as the no-GPUs path to smarter humans, citing Gwern — zackmdavis · 2026-09-04