Add v1.1 model card

README.md

@@ -0,0 +1,198 @@
+---
+license: cc0-1.0
+base_model: mlx-community/Qwen2.5-Coder-7B-Instruct-4bit
+tags:
+- gguf
+- cybersecurity
+- nist
+- security-controls
+- compliance
+- fine-tuned
+- llama-cpp
+language:
+- en
+quantized_by: ethanolivertroy
+---
+
+# HackIDLE-NIST-Coder v1.1 (GGUF)
+
+**The most comprehensive NIST cybersecurity model** in GGUF format - Compatible with llama.cpp, Ollama, LM Studio, and text-generation-webui.
+
+## Model Overview
+
+Fine-tuned on 530,912 examples from 596 NIST publications. Version 1.1 includes:
+
+- **+7,206 training examples** (530,912 total)
+- **+28 new documents** (596 NIST publications)
+- **CSWP series**: CSF 2.0, Zero Trust Architecture, Post-Quantum Cryptography
+- **Improved quality**: Fixed 6,150 malformed DOI links, 0 broken link markers
+
+## Available Quantizations
+
+| Quantization | Size | Use Case | Description |
+|--------------|------|----------|-------------|
+| **F16** | ~14 GB | Reference Quality | Full precision, best quality |
+| **Q8_0** | ~7.5 GB | High Quality | Minimal quality loss |
+| **Q5_K_M** | ~5.1 GB | Balanced | Good quality/size trade-off |
+| **Q4_K_M** | ~4.4 GB | Recommended | Best speed/quality balance |
+
+**Recommended**: Start with **Q4_K_M** for best overall performance.
+
+## Training Data (v1.1)
+
+**Dataset**: [ethanolivertroy/nist-cybersecurity-training](https://huggingface.co/datasets/ethanolivertroy/nist-cybersecurity-training)
+
+**Coverage:**
+- **FIPS**: Cryptographic standards
+- **SP 800**: Security guidelines and controls
+- **SP 1800**: Practice guides
+- **IR**: Technical reports
+- **CSWP**: White Papers (CSF 2.0, Zero Trust, PQC, IoT, Privacy) ✨ NEW
+
+**Stats**: 530,912 examples • 596 documents • 61,480 working references
+
+## Installation
+
+### Ollama
+
+```bash
+# Pull from Ollama registry
+ollama pull etgohome/hackidle-nist-coder:v1.1
+
+# Or create from GGUF
+ollama create hackidle-nist-coder -f Modelfile
+```
+
+### LM Studio
+
+1. Open LM Studio
+2. Search for "hackidle-nist-coder"
+3. Download Q4_K_M or Q5_K_M quantization
+4. Load and chat
+
+### llama.cpp
+
+```bash
+# Clone llama.cpp
+git clone https://github.com/ggerganov/llama.cpp
+cd llama.cpp && make
+
+# Download model (Q4_K_M recommended)
+wget https://huggingface.co/ethanolivertroy/HackIDLE-NIST-Coder-v1.1-GGUF/resolve/main/hackidle-nist-coder-v1.1-q4_k_m.gguf
+
+# Run inference
+./llama-cli -m hackidle-nist-coder-v1.1-q4_k_m.gguf -p "What is Zero Trust Architecture?"
+```
+
+### text-generation-webui
+
+1. Place GGUF file in `models/` directory
+2. Select model in UI
+3. Load and chat
+
+## Usage Examples
+
+### Ollama
+
+```bash
+ollama run etgohome/hackidle-nist-coder:v1.1 "Explain the CSF 2.0 GOVERN function"
+```
+
+### Python (llama-cpp-python)
+
+```python
+from llama_cpp import Llama
+
+llm = Llama(
+    model_path="hackidle-nist-coder-v1.1-q4_k_m.gguf",
+    n_ctx=4096,
+    n_threads=8
+)
+
+response = llm("What are the core principles of Zero Trust Architecture in SP 800-207?",
+               max_tokens=500)
+print(response['choices'][0]['text'])
+```
+
+## Model Capabilities
+
+Trained on comprehensive NIST content:
+
+✅ **Security Controls** (SP 800-53)
+✅ **CSF 2.0** with GOVERN function
+✅ **Zero Trust Architecture** (SP 800-207)
+✅ **Risk Management Framework** (RMF)
+✅ **Cloud Security** (SP 800-145, 800-146)
+✅ **FIPS Cryptography** standards
+✅ **Post-Quantum Cryptography** migration
+✅ **Privacy Engineering**
+✅ **Supply Chain Risk Management**
+✅ **IoT Cybersecurity**
+
+## What's New in v1.1
+
+**Added Content:**
+- CSF 2.0 (Cybersecurity Framework 2.0)
+- Zero Trust Architecture planning guidance
+- Post-Quantum Cryptography recommendations
+- IoT security and labeling
+- Privacy Framework v1.0
+- Supply chain risk management case studies
+
+**Quality Improvements:**
+- Fixed 6,150 malformed DOI links
+- Removed 202 broken link markers
+- Validated 124,946 total links
+- Clean training data
+
+## System Requirements
+
+| Quantization | RAM Required | CPU/GPU |
+|--------------|-------------|---------|
+| Q4_K_M | 6 GB | CPU or GPU |
+| Q5_K_M | 7 GB | CPU or GPU |
+| Q8_0 | 10 GB | CPU or GPU |
+| F16 | 16 GB | GPU recommended |
+
+## Other Formats
+
+- **MLX**: [ethanolivertroy/HackIDLE-NIST-Coder-v1.1-MLX-4bit](https://huggingface.co/ethanolivertroy/HackIDLE-NIST-Coder-v1.1-MLX-4bit) (Apple Silicon)
+- **Ollama**: [etgohome/hackidle-nist-coder](https://ollama.com/etgohome/hackidle-nist-coder)
+
+## Limitations
+
+- Training data current as of October 2025
+- May not reflect NIST publications released after training
+- 54.2% of references are broken links (cataloged for recovery)
+- Optimized for NIST-specific cybersecurity questions
+
+## Citation
+
+```bibtex
+@misc{hackidle-nist-coder-v1.1-gguf,
+  title={HackIDLE-NIST-Coder: NIST Cybersecurity Expert Model},
+  author={Troy, Ethan Oliver},
+  year={2025},
+  version={1.1},
+  format={GGUF},
+  url={https://huggingface.co/ethanolivertroy/HackIDLE-NIST-Coder-v1.1-GGUF}
+}
+```
+
+## License
+
+CC0 1.0 Universal (Public Domain) - All NIST publications are in the public domain.
+
+## Acknowledgments
+
+- NIST Computer Security Resource Center
+- Qwen2.5-Coder base model (Alibaba Cloud)
+- llama.cpp quantization (Georgi Gerganov)
+- MLX framework (Apple)
+
+---
+
+**Version**: 1.1
+**Release Date**: October 2025
+**Training Dataset**: [nist-cybersecurity-training v1.1](https://huggingface.co/datasets/ethanolivertroy/nist-cybersecurity-training)
+**Format**: GGUF (compatible with llama.cpp ecosystem)