- By Ashish Singh
- Tue, 16 Apr 2024 02:45 PM (IST)
- Source:JND
Google-backed Agritech firm Cropin Technologies announced the first open-source micro-language model for climate-smart agriculture on Tuesday, aimed at underprivileged farmers in the Global South. Google-backed Cropin created the "Aksara" AI model with the goal of lowering knowledge gaps so that anyone in the agriculture ecosystem can develop scalable and reasonably priced AI solutions for the sector.
'Aksara's' initial version will cover nine crops for five nations in the Indian subcontinent: paddy, wheat, maize, sorghum, barley, cotton, sugarcane, soybean, and millets. 'Aksara' has been painstakingly compressed from 16-bit to 4-bit by Cropin, who is aware of the environmental impact of operating big language models (LLMs).
"Domain-specific AI models for agriculture are expected to attract significant investments, offering a practical and economically viable approach to food systems transformation," Krishna Kumar, founder and chief executive officer of Cropin, stated.
These models have the potential to revolutionise agriculture and usher in a new era of tech-driven farming in an industry that has historically seen little progress in terms of technical improvement. The 'Aksara' AI model was refined using over 5,000 superior question-answer pairs related to agriculture and over 160,000 tokens within the context.
"These numbers are expected to increase as we add more crops, geographic locations and use cases," the business stated. Unpredictable heatwaves, excessive or irregular rainfall, and a rise in pest and disease attacks all have an impact on farmers' operations and lower agricultural production, profitability, and yield.
According to Cropin, the goal of "Aksara" is to close this gap by utilising GenAI to deliver precise data, agricultural advice, and insights into contemporary farming methods. According to the company, the open-source project is to assist agronomists, agri-scientists, field personnel, and extension agents while progressively expanding the services to farmers in several languages.
(With Agency Inputs)