🛠️ my-lora-model (Qwen2.5-1.5B Sysadmin Tool-Use)

A lightweight LoRA fine-tune of Qwen2.5-1.5B-Instruct, specialized for fast function calling and tool use in system administration tasks. Designed for local execution on low-resource hardware, CPUs, or edge devices.

---

💡 What it does

The model processes natural language requests in Russian and English and outputs structured tool execution parameters.

Supported Tools out-of-the-box:

• get_gpu_status — Retrieve GPU temperature, utilization, and memory usage.
• execute_terminal_command — Run an arbitrary Bash command.
• read_file — Read the contents of a specified file path.

Example Prompts handled:

• 🇬🇧 "What's the GPU temperature?" ➔ Calls get_gpu_status
• 🇬🇧 "Uptime?" ➔ Calls execute_terminal_command with uptime
• 🇷🇺 "Покажи конфиг config.json" ➔ Calls read_file with path config.json

---

🚀 How to use

Via Ollama (GGUF version)

1. Download the .gguf file from the repository.
2. Create a Modelfile:

   FROM ./my-lora-model-q8_0.gguf
   SYSTEM "You are a helpful system administration assistant capable of using system tools."