KissanAI Dhenu 1.0
KissanAI Dhenu 1.0 is India's premier bilingual agricultural Large Language Model, specifically trained on regional agronomy data to provide conversational advisory to smallholder farmers.
Profile & Intelligence Breakdown
Bridging AI and Indian Agriculture
Dhenu 1.0 is an open-weights, domain-specific agricultural foundation model created by KissanAI. Designed specifically to overcome the linguistic and knowledge barriers faced by over 140 million Indian farming households, Dhenu combines advanced natural language processing with verified agricultural science.
Dataset & Training Methodology
- Curated Agricultural Corpus: Trained on over 300,000 curated conversational pairs in Hindi and English covering crop disease diagnosis, soil management, irrigation schedules, and government welfare schemes.
- Bilingual Instruction Tuning: Built upon foundational open-source architectures, fine-tuned specifically to handle colloquial Indian dialects and vernacular agricultural terminology.
- Multimodal Voice Integration: Integrated with speech-to-text and text-to-speech pipelines, allowing illiterate and semi-literate farmers to ask questions via simple voice notes on WhatsApp or smartphone apps.
Real-World Impact
Dhenu empowers grassroots agricultural extension workers and farmers with instant, accurate diagnostics for pest infestations, weather-resilient sowing, and market crop prices, significantly reducing input costs and boosting farm yields.
Technical Specifications
- parameters
- 7 Billion
- context length
- 8,192 Tokens
- license
- Commercial
- developer
- KissanAI / Pratik Desai