MokolIslam/gemma-3-1b-it-qwen3-tool-template
MokolIslam/gemma-3-1b-it-qwen3-tool-template is a 1 billion parameter instruction-tuned Gemma 3 model developed by Google DeepMind. This variant specifically integrates a Qwen3-style tool-calling template, enabling structured function calls and responses. It is designed for tool-use applications, allowing the model to interact with external functions for enhanced capabilities.
Loading preview...
Overview
This model, MokolIslam/gemma-3-1b-it-qwen3-tool-template, is a specialized variant of Google DeepMind's Gemma 3 1B instruction-tuned model. It retains the core capabilities of the Gemma 3 family, which are lightweight, state-of-the-art open models built from the same research as Gemini models. The key differentiator for this specific model is its Qwen3-style tool-calling template, which has been adapted for Gemma 3's <start_of_turn> / <end_of_turn> format.
Key Capabilities
- Tool Calling: Supports structured tool calls and responses using
<tool_call>and<tool_response>tags, enabling interaction with external functions. - Tool Definitions: Tool definitions can be injected into the first user turn.
- Role Support: Handles system and tool roles (
<start_of_turn>system,<start_of_turn>tool). - Multimodal (Gemma 3 base): The underlying Gemma 3 models are multimodal, capable of handling text and image inputs and generating text outputs, though this specific variant focuses on tool-use for text generation.
- Multilingual Support: The Gemma 3 family supports over 140 languages.
Good for
- Developing AI agents: Ideal for applications requiring the model to interact with external APIs or tools.
- Automated workflows: Can be used to build systems where the model needs to perform actions based on user requests.
- Function calling scenarios: Specifically optimized for scenarios where structured function calls are a primary requirement.