Snon - A large-scale open-source agricultural model from Nanjing Agricultural University

What is the Minister of Agriculture?

Sinon (Sinon Large Language Model) is a large-scale vertical model for the agricultural field launched by Nanjing Agricultural University. Leveraging the university's disciplinary strengths, the model has collected over 4 billion tokens covering multiple agricultural disciplines, including books, papers, and policies. Trained through synthetic data construction, instruction fine-tuning, and reinforcement learning, the model possesses a powerful ability to understand agricultural knowledge. Sinon is available in open-source 8B and 32B versions, supporting multi-agent retrieval.

The main functions of the Minister of Agriculture

• Agricultural Knowledge Q&AThe model can provide accurate and detailed answers to agriculture-related questions, covering multiple fields such as planting, breeding, and agricultural economics.
• Multi-agent retrieval enhancementBy optimizing the search framework, relevant information can be quickly extracted from massive amounts of agricultural literature, thereby improving the efficiency of knowledge acquisition.
• Mind chain and context understandingIt supports both thought-chain question answering (COT-QA) and context-referenced question answering (Incontext-QA), enabling in-depth reasoning and analysis based on context.
• Understanding and Execution of Directives in the Agricultural SectorThrough fine-tuning of instructions, it can accurately understand and execute complex instructions in the agricultural field, assisting agricultural decision-making.
• Data-driven decision supportBy combining agricultural big data, we can provide data support and decision-making suggestions for agricultural production, management, and investment.

Sinon's project address

• GitHub repositoryhttps://github.com/njauzzx/Sinong

Sinong's core advantages

• Domain focusSinong focuses on the agricultural field, deeply optimizing its solutions for sub-disciplines such as animal science, plant protection, and agricultural economics. It can accurately understand and handle agricultural-related issues and provide professional and accurate solutions.
• Large-scale high-quality dataLeveraging the disciplinary strengths of Nanjing Agricultural University, Sinofarm has collected over 4 billion tokens of agricultural data, covering various types such as books, papers, policies, and patents. The data is abundant and has undergone rigorous screening and organization to ensure a solid foundation for model training.
• Multimodal fusion technologyBy combining visual models and large language models, Sinong can process various data formats such as images and text, realize the fusion of multimodal information, and improve the ability to understand and analyze complex agricultural scenarios.
• Instruction fine-tuning and reinforcement learningThrough synthetic data construction, instruction fine-tuning, and reinforcement learning, Sinon excels in instruction understanding and execution in the agricultural field, and can quickly adapt to diverse tasks in agriculture, such as question answering, reasoning, and decision support.
• Multi-agent retrieval enhancementSinon introduces a multi-agent retrieval enhancement framework, which optimizes knowledge base construction and retrieval efficiency, enabling the rapid extraction of relevant information from massive amounts of agricultural literature and improving the efficiency of knowledge acquisition and application.

Application scenarios of Sinon

• Agricultural production and plantingIt provides farmers with planting plans, pest and disease control, and agricultural technology consultation to help them with scientific planting and efficient management.
• Livestock farmingIt assists farmers in breeding management, animal health diagnosis, and breeding environment optimization, thereby improving breeding efficiency.
• Agricultural Economics and Market AnalysisAnalyze market trends, interpret agricultural policies, provide advice to farmers and businesses, and assist in agricultural economic decision-making.
• Agricultural research and educationThe model supports researchers in literature retrieval, experimental design, and the generation of educational resources, thereby promoting the development of agricultural research and education.
• Smart agriculture and digital transformationBy combining the Internet of Things (IoT) to achieve intelligent monitoring and early warning, analyze agricultural big data, and optimize agricultural supply chain management.