# Phi-4-Mini Cybersecurity Fine-tuned (Quantized GGUF) 

Fine-tuned on cybersecurity harmony dataset using LoRA (r=32) for tasks like QA, security assessment, incident response.

- Base model: microsoft/Phi-4-mini-instruct (merged with adapter)
- Training: 1 epoch, 102k examples
- Use: Load with AutoModelForCausalLM.from_pretrained('KellanF89/phi4-cybersecurity-quantized')

License: MIT