Google-backed Cropin Technology launches open-source AI model for climate smart agriculture

Bengaluru-based agtech firm Cropin Technology recently unveiled 'akṣara', the first micro language model (µ-LM) tailored specifically for climate-smart agriculture.

The tool, which was developed on top of Mistral’s foundational model, is open-sourced under the Apache 2.0 License.

'Akṣara' is designed to transform agricultural practices by empowering developers, researchers, and other stakeholders to create scalable AI solutions. These solutions aim to enhance food security, adapt to climate challenges, and foster regenerative agricultural practices, particularly in the Global South.

The design and utility of akṣara

The micro model, 'akṣara', is initially equipped to handle data related to nine critical crops, including paddy, wheat, maize, and soybean, covering five countries in the Indian subcontinent.

It utilizes a compression technique from 16-bit to 4-bit using Quantization and Low-Rank Adapters (QLoRA) to minimize its environmental footprint. This technique significantly reduces the resources needed for computation and storage, making it both cost-effective and energy-efficient. 

Krishna Kumar, Founder & CEO of Cropin, said, "In an era where Large Language Models are reshaping jobs, businesses, and customer interactions, the spotlight is now on industry-specific models trained on niche and comprehensive domain data as the ‘next big thing.’ These models can potentially transform agriculture, paving the way for a new era of tech-driven farming in a sector that has traditionally seen limited technological advancement."

"Domain-specific AI models for agriculture are expected to attract significant investments, offering a practical and economically viable approach to food systems transformation. Akṣara reinstates our commitment to leading the tech-driven agricultural movement in the years ahead, significantly impacting small-scale farmers’ lives," Kumar added.

Aligning with responsible AI principles

Emphasizing the importance of ethical AI, Cropin ensures that 'akṣara' aligns with responsible AI principles by incorporating feedback from discussions with Google’s Responsible AI team.

The model is designed to be frugal, scalable, and devoid of biases, promoting sustainable practices such as soil and water conservation. This open-source initiative aims to democratize access to AI technologies, enabling a wide range of users, from agronomists to local farmers, to harness its capabilities.

Extending capabilities to multiple languages

Currently, 'akṣara' supports only English, but plans are in place to extend its capabilities to multiple languages and additional crops over the next six months.

This progressive enhancement is expected to widen the model's applicability, allowing it to reach a broader audience and have a more substantial impact on global agriculture. Furthermore, Cropin is actively seeking collaboration with industry players and academia to accelerate innovation in digital agriculture.