---
language:
- fr
- en
license: apache-2.0
library_name: gguf
base_model: Qwen/Qwen2.5-1.5B-Instruct
tags:
- cybersecurity
- iso27001
- gguf
- quantized
- ollama
- llama-cpp
pipeline_tag: text-generation
---

# ISO27001-Expert-1.5B-GGUF

**GGUF quantized versions** of [AYI-NEDJIMI/ISO27001-Expert-1.5B](https://huggingface.co/AYI-NEDJIMI/ISO27001-Expert-1.5B) for use with [Ollama](https://ollama.ai), [llama.cpp](https://github.com/ggerganov/llama.cpp), [LM Studio](https://lmstudio.ai), and other GGUF-compatible inference engines.

## Model Description

This is a fine-tuned Qwen2.5-1.5B-Instruct model specialized in **ISO 27001 information security management**. It can answer questions about ISO 27001 controls, implementation guidance, risk assessment, compliance requirements, and security best practices in both French and English.

Part of the **AYI-NEDJIMI Cybersecurity AI Portfolio**:
- [AYI-NEDJIMI/CyberSec-AI-Portfolio](https://huggingface.co/collections/AYI-NEDJIMI/cybersec-ai-portfolio-6850da55c1b0578430f1f553) — Full collection

## Available Quantizations

| Filename | Quant Type | Size | Description |
|---|---|---|---|
| `iso27001-expert-1.5b-Q4_K_M.gguf` | Q4_K_M | 941 MB | **Recommended** — Best balance of quality and size (~33% of F16) |
| `iso27001-expert-1.5b-Q5_K_M.gguf` | Q5_K_M | 1.07 GB | Higher quality, slightly larger (~38% of F16) |
| `iso27001-expert-1.5b-Q8_0.gguf` | Q8_0 | 1.57 GB | Near-lossless quantization (~54% of F16) |

### Quantization Format Details

- **Q4_K_M**: 4-bit quantization with k-quant medium quality. Excellent for resource-constrained environments. Minimal quality loss for most tasks.
- **Q5_K_M**: 5-bit quantization with k-quant medium quality. Good middle ground between Q4 and Q8.
- **Q8_0**: 8-bit quantization. Near-original quality with ~50% size reduction from F16.

## How to Use

### Ollama

Create a `Modelfile`:

```
FROM ./iso27001-expert-1.5b-Q4_K_M.gguf

TEMPLATE """<|im_start|>system
{{ .System }}<|im_end|>
<|im_start|>user
{{ .Prompt }}<|im_end|>
<|im_start|>assistant
"""

SYSTEM "You are an ISO 27001 expert assistant. You provide detailed, accurate guidance on information security management systems (ISMS), ISO 27001 controls, risk assessment, and compliance. You respond in the same language as the user's question."

PARAMETER temperature 0.7
PARAMETER top_p 0.8
PARAMETER top_k 20
PARAMETER stop "<|im_end|>"
```

Then run:

```bash
ollama create iso27001-expert -f Modelfile
ollama run iso27001-expert
```

### llama.cpp

```bash
# Interactive chat
./llama-cli -m iso27001-expert-1.5b-Q4_K_M.gguf \
  -p "You are an ISO 27001 expert assistant." \
  --chat-template chatml \
  -cnv

# Server mode
./llama-server -m iso27001-expert-1.5b-Q4_K_M.gguf \
  --host 0.0.0.0 --port 8080
```

### LM Studio

1. Download the desired GGUF file
2. Open LM Studio and load the model from your downloads
3. Select the **ChatML** chat template
4. Set the system prompt to: "You are an ISO 27001 expert assistant."
5. Start chatting!

### Python (llama-cpp-python)

```python
from llama_cpp import Llama

llm = Llama(model_path="iso27001-expert-1.5b-Q4_K_M.gguf", n_ctx=4096)

response = llm.create_chat_completion(
    messages=[
        {"role": "system", "content": "You are an ISO 27001 expert assistant."},
        {"role": "user", "content": "Quels sont les contrôles clés de l'Annexe A de la norme ISO 27001?"}
    ],
    temperature=0.7,
    top_p=0.8,
    top_k=20,
)
print(response["choices"][0]["message"]["content"])
```

## Related Models

| Version | Link |
|---|---|
| Merged (SafeTensors) | [AYI-NEDJIMI/ISO27001-Expert-1.5B](https://huggingface.co/AYI-NEDJIMI/ISO27001-Expert-1.5B) |
| LoRA Adapter | [AYI-NEDJIMI/ISO27001-Expert-1.5B-Adapter](https://huggingface.co/AYI-NEDJIMI/ISO27001-Expert-1.5B-Adapter) |
| GGUF (this repo) | [AYI-NEDJIMI/ISO27001-Expert-1.5B-GGUF](https://huggingface.co/AYI-NEDJIMI/ISO27001-Expert-1.5B-GGUF) |
| Portfolio Collection | [AYI-NEDJIMI/CyberSec-AI-Portfolio](https://huggingface.co/collections/AYI-NEDJIMI/cybersec-ai-portfolio-6850da55c1b0578430f1f553) |

## Technical Details

- **Base Model**: Qwen/Qwen2.5-1.5B-Instruct
- **Fine-tuning**: QLoRA (4-bit) with LoRA adapters merged back
- **Architecture**: Qwen2ForCausalLM
- **Context Length**: 4096 tokens
- **Chat Template**: ChatML
- **Converted with**: llama.cpp (convert_hf_to_gguf.py)