phattrandeveloper/functiongemma-270m-function-calling

Model Overview

phattrandeveloper/functiongemma-270m-function-calling is a compact 0.3 billion parameter model, derived from Google's functiongemma-270m-it. It has been specifically fine-tuned on a Vietnamese banking tool-calling dataset, making it highly specialized for function-calling tasks in this domain.

Key Capabilities

• Function Calling: Designed to interpret natural language requests and generate structured tool calls, adhering to the FunctionGemma-style chat template.
• Vietnamese Banking Domain: Optimized for scenarios involving Vietnamese banking operations, leveraging its specialized training dataset.
• Chat Template Integration: Intended for use with tokenizer.apply_chat_template by passing tool schemas, facilitating seamless integration into function-calling workflows.
• MediaPipe/Flutter Compatibility: Includes tokenizer.model for conversion to MediaPipe / Flutter Gemma .task formats, enabling deployment in mobile and edge environments.
• Extended Context: Features a substantial context length of 32768 tokens, allowing for more complex and detailed function-calling scenarios.

Good For

• Developing Function-Calling Agents: Ideal for creating AI agents that need to interact with external tools or APIs based on user input.
• Vietnamese Banking Applications: Particularly well-suited for applications requiring automated responses or actions within the Vietnamese banking sector.
• Mobile/Edge Deployment: The inclusion of tokenizer.model makes it a strong candidate for applications requiring on-device inference via MediaPipe or Flutter.