File size: 1,772 Bytes
b682563
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
---
license: cc0-1.0
base_model: mlx-community/Qwen2.5-Coder-7B-Instruct-4bit
tags:
- mlx
- cybersecurity
- nist
- security-controls
- compliance
- fine-tuned
language:
- en
---

# HackIDLE-NIST-Coder v1.1 (MLX 4-bit)

**The most comprehensive NIST cybersecurity model** - Fine-tuned on 530,912 examples from 596 NIST publications.

## Model Overview

This is an MLX-optimized 4-bit quantized model fine-tuned specifically for NIST cybersecurity expertise. Version 1.1 includes significant improvements over v1.0:

- **+7,206 training examples** (530,912 total)
- **+28 new documents** (596 NIST publications)
- **CSWP series added**: CSF 2.0, Zero Trust Architecture, Post-Quantum Cryptography
- **Improved quality**: Fixed 6,150 malformed DOI links

## Training Results

- **Training iterations**: 1,000 (+ 200 checkpoint recovery)
- **Best validation loss**: 1.512 (12.5% improvement)
- **Training loss**: 1.420 (final)
- **Trainable parameters**: 11.5M (0.151% of 7.6B total)
- **Training time**: ~5 hours on M4 Max

## Installation

```bash
pip install mlx-lm
```

## Usage

```python
from mlx_lm import load, generate

model, tokenizer = load("ethanolivertroy/HackIDLE-NIST-Coder-v1.1-MLX-4bit")

prompt = "What is Zero Trust Architecture according to NIST SP 800-207?"
response = generate(model, tokenizer, prompt=prompt, max_tokens=500)
print(response)
```

## Other Formats

- **Ollama**: [etgohome/hackidle-nist-coder:v1.1](https://ollama.com/etgohome/hackidle-nist-coder)
- **Dataset**: [ethanolivertroy/nist-cybersecurity-training](https://huggingface.co/datasets/ethanolivertroy/nist-cybersecurity-training)

## License

CC0 1.0 Universal (Public Domain) - All NIST publications are in the public domain.

---

**Version**: 1.1
**Release Date**: October 2025