CyberCoder-0.5B

🤖 CyberCoder 0.5B (GGUF)

🚀 Size: < 500 MB | ⚡ Format: GGUF (Q4_K_M) | 🛡️ Task: Clean Coding & Ethical Hacking

CyberCoder 0.5B is a highly optimized, ultra-lightweight language model engineered specifically for clean, secure code generation, real-time debugging, and authorized ethical hacking & penetration testing methodologies. Fine-tuned on specialized coding and cybersecurity datasets, this model maintains strict adherence to defensive guardrails and educational boundaries.

---

📊 Technical Specifications & Hardware Targets

Thanks to its compact architecture and aggressive quantization, this model is built to run flawlessly on low-end hardware, edge configurations, and mobile devices.

• Base Architecture: Qwen2.5-0.5B-Instruct
• Quantization Format: GGUF (Q4_K_M - 4-bit)
• Model File Size: < 500 MB
• Context Length: 2048 Tokens
• RAM Requirement: < 4GB (Runs perfectly on standard smartphones and older laptops)

---

💬 Prompt & Chat Format (ChatML)

This model utilizes the ChatML template configuration for structured messaging. Ensure your interface uses <|im_start|> and <|im_end|> delimiters to maintain maximum performance.

<|im_start|>system
You are a helpful, authorized coding and cybersecurity assistant. You adhere strictly to defensive guidelines.<|im_end|>
<|im_start|>user
Write a python script to scan open ports on a local network IP address.<|im_end|>
<|im_start|>assistant

---

📱 How to Use on Mobile (PocketPal AI)

You can run this model completely offline on your smartphone using PocketPal AI.

• Download and install PocketPal AI from your device's app store.

• Download the CyberCoder-0.5B-Q4.gguf file from this repository to your phone's storage.

• Open PocketPal AI and import/load the local .gguf file.

• Configure the Chat Template settings in the app to use ChatML.

• Set your system prompt to enforce safe, authorized parameters, and start chatting offline!

---

💻 How to Use on Desktop (Ollama & LM Studio)

Option 1: Run instantly with Ollama (Recommended) Ollama allows you to stream and run models directly from Hugging Face repositories without downloading files manually. Run the following command in your terminal or command prompt:

• Bash ollama run hf.co/Nitishsharma9/super-lite-model-upload:Q4_K_M

• What this does:

• ollama run: Initializes and executes a local model instance.

• hf.co/...: Directs Ollama to fetch the model from Hugging Face's registry.

• Nitishsharma9/super-lite-model-upload: Specifies your specific repository path.

• :Q4_K_M: Targets the precise 4-bit optimized quantization file.

Note: The first time you run this command, Ollama will automatically pull the file down from Hugging Face. Subsequent loads will launch instantly offline.

Option 2: LM Studio (Visual GUI Experience) Launch LM Studio on your desktop.

• In the search bar, type your repository identifier: Nitishsharma9/super-lite-model-upload.

• Locate and download the Q4_K_M GGUF file from the files list.

• Go to the Chat view, select the model from the top dropdown list, and ensure the Preset format is set to ChatML.

---

🛡️ Responsible Use Disclaimer

-- This model is intended solely for educational purposes, defensive security research, authorized penetration testing, and software development optimization. Users are legally obligated to comply with local laws and must obtain proper, written authorization before utilizing this tool against any networks, systems, or digital infrastructure.