Google FunctionGemma: Fine-tune and run locally with Unsloth and LM Studio
Action Required
Developers can now create specialized agents for tool-based workflows and applications, improving efficiency and reducing reliance on cloud-based services.
AI Impact Summary
Google released FunctionGemma, a 270M parameter model designed for task-specific tool calling, and provided a step-by-step guide on how to fine-tune it using Unsloth and run it locally in LM Studio. This capability allows developers to create specialized agents for workflows and applications, leveraging the model's small size for efficient local inference. Users need to follow the provided instructions to fine-tune the model and run it in LM Studio, which requires installing Unsloth and LM Studio.
Affected Systems
- Date
- 23 Dec 2025
- Change type
- capability
- Severity
- high