Simon-Liu/gemma-4-e2b-kubectl-mcp-server-mcp-sft-it

VISIONConcurrent Unit Cost:1Model Size:5.1BQuant:BF16Context Size:32kTool Calling:SupportedPublished:Jul 17, 2026License:apache-2.0Architecture:Transformer Open Weights Featherless Exclusive Cold

Simon-Liu/gemma-4-e2b-kubectl-mcp-server-mcp-sft-it is a 5.1 billion parameter Gemma-based model, fine-tuned for tool-calling within a Kubernetes Multi-Cluster Platform (MCP) environment. Developed by the Agent Tools Fine-Tuning Platform, it excels at generating single tool calls for a wide array of kubectl and MCP server operations. This model is specifically designed for automated Kubernetes cluster management and operations, offering enhanced accuracy in function name and parameter generation for tool-use scenarios.

Loading preview...

Overview

This model, Simon-Liu/gemma-4-e2b-kubectl-mcp-server-mcp-sft-it, is a 5.1 billion parameter Gemma-based language model that has undergone a full fine-tuning process. It is specifically designed for tool-calling within a Kubernetes Multi-Cluster Platform (MCP) context. The fine-tuning was conducted by the Agent Tools Fine-Tuning Platform using a "Teacher reverse data generation → quality filtering → SFT" methodology, building upon the google/gemma-4-E2B-it base model.

Key Capabilities

  • Extensive Tool Integration: Supports a comprehensive set of 100+ tools from the MCP server, covering a wide range of kubectl and Kubernetes management operations.
  • Enhanced Tool-Calling Accuracy: Achieves a 96.7% accuracy for function names and 80.7% for full name + parameter correctness in tool calls, significantly improving upon the base model's performance.
  • Automated Kubernetes Operations: Trained to output a single <tool_call> when a user's request requires tool interaction, enabling automated execution of cluster management tasks.

Good For

  • Automated Kubernetes Management: Ideal for applications requiring automated interaction with Kubernetes clusters, such as deployment scaling, helm chart management, pod diagnostics, and resource monitoring.
  • Agent-Based Systems: Suitable for integration into AI agents designed to perform complex operations within a Kubernetes environment by accurately generating tool calls.
  • Operational Efficiency: Improves the reliability and precision of tool-calling for developers and operators managing multiple Kubernetes clusters.

Finetunes
1 models