Soham308/CyberGuard-Model

Qwen3.5 4B GGUF (Quantized)

This repository provides a GGUF quantized version of the original Qwen3.5 4B model, optimized for efficient local inference using tools like llama.cpp, LM Studio, and similar runtimes.

---

🔗 Base Model

This model is derived from the official:

👉 https://huggingface.co/Qwen/Qwen3.5-4B (by Alibaba / Qwen Team)

Please refer to the original model for full details, training methodology, benchmarks, and licensing terms.

---

⚙️ Quantization Details

• Format: GGUF
• Quantization: Q4_K_XL
• Size: ~2.9 GB
• Architecture: Qwen3.5

This version is designed to balance performance and memory efficiency, making it suitable for local deployments.

---

📦 Quantization Source

This GGUF file is sourced from:

👉 https://huggingface.co/unsloth/Qwen3.5-4B-GGUF

Specifically:

• Qwen3.5-4B-UD-Q4_K_XL.gguf

All credit for quantization goes to the original uploader (Unsloth).

---

🚀 Usage

You can run this model locally using:

llama.cpp

./main -m qwen3.5-4b-q4_k_xl.gguf -p "Explain SQL injection"

Other tools

• LM Studio
• KoboldCpp
• Ollama

---

💡 Example Use Cases

• General-purpose chat
• Coding assistance
• Technical explanations
• Integration into custom AI systems (e.g., agents, tools)

---

🧪 Tested With

• Local inference (CPU/GPU hybrid)
• Integration with external tools (web search, reasoning pipelines)

---

⚠️ Disclaimer

• This is not an original model.
• Behavior and capabilities are inherited from the base Qwen3.5 model.

---

📜 License

• Please follow the license of the original Qwen model.

---

🙌 Acknowledgements

• Qwen Team (Alibaba) — Base model
• Unsloth — GGUF quantization
• llama.cpp — GGUF runtime support

---

🌐 Related Project

This model is used in:

👉 CyberGuard AI (Cybersecurity assistant system)

• Hosted on huggingface spaces

---