Vyakyarth - Multilingual Sentence Embedding Model

About Model

Vyakyarth is a multilingual sentence embedding model designed for Indic languages. Developed by Krutrim AI Labs, it is fine-tuned from stsb-xlm-r-multilingual, leveraging sentence-transformers to encode sentences and paragraphs into 768-dimensional embeddings. The model supports a wide range of NLP tasks, including semantic textual similarity, semantic search, paraphrase mining, text classification, and clustering. Built on XLM-RoBERTa, Vyakyarth ensures high-quality representations across multiple Indian languages, making it ideal for applications in search engines, recommendation systems, and multilingual NLP research.