MohitM2/sruti-1.5b
MohitM2/sruti-1.5b is a 1.5 billion parameter language model, fine-tuned from Qwen2.5-1.5B, specifically designed as a knowledgeable guide to Hindu, Sanātana, Vedic, and Vedāntic philosophy. It excels at answering questions in clear English, often providing original Sanskrit (Devanāgarī with IAST transliteration) with translation and source. This model is optimized for philosophical inquiry, drawing from a curated 19M-token corpus of primary Hindu texts and hand-authored Q&A.
Loading preview...
Śruti 1.5B: A Guide to Hindu Philosophy
MohitM2/sruti-1.5b is a specialized 1.5 billion parameter language model, built upon the Qwen2.5-1.5B architecture. It functions as an expert guide to Hindu, Sanātana, Vedic, and Vedāntic philosophy, covering a wide range of texts including the Vedas, Upaniṣads, Bhagavad Gītā, Purāṇas, and various darśanas.
Key Capabilities
- Philosophical Inquiry: Provides clear English answers to questions on Hindu philosophy.
- Sanskrit Integration: Where relevant, it quotes original Sanskrit in Devanāgarī with IAST transliteration, accompanied by translation and source.
- Extensive Knowledge Base: Fine-tuned with LoRA on hand-authored Q&A and a ~19M-token corpus of primary sources, including ancient scriptures and modern expositors.
Training Details
The model was fine-tuned using LoRA (r=64, α=128) on all attention and MLP projections of the Qwen/Qwen2.5-1.5B base model. Training utilized fp16 precision on a single NVIDIA T4/P100, focusing on instruction-based Q&A with Sanskrit citations.
Limitations
As a 1.5B-parameter model, Śruti may occasionally misquote, oversimplify complex doctrinal points, or hallucinate citations. Users should verify Sanskrit quotations and sources for scholarly use, as it is not a substitute for direct study of primary texts or guidance from qualified teachers.