ethanolivertroy/HackIDLE-NIST-Coder-GGUF

HackIDLE-NIST-Coder (GGUF)

This is the first GGUF build of HackIDLE-NIST-Coder, a NIST-focused local model built from Qwen2.5-Coder-7B-Instruct and fine-tuned on a NIST cybersecurity corpus.

This repo is kept for reproducibility. For new testing, start with the v1.1 GGUF build:

• ethanolivertroy/HackIDLE-NIST-Coder-v1.1-GGUF

Use this model as a helper. Do not treat it as a source of truth for exact control names, RMF step lists, or reference-architecture component names without checking the source publication.

Training data

This first build used 523,706 examples from 568 NIST cybersecurity documents.

Training dataset:

• ethanolivertroy/nist-cybersecurity-training

Current eval status

The dated smoke eval from April 22, 2026 was run against the Ollama latest tag, which matched the v1.1 line in the local install used for that check. I have not rerun that exact eval against this older GGUF build.

The v1.1 result matters for this older build too because it sets the right expectation for the model family: the model can stay in-domain while still missing exact NIST structure.

Be careful with:

• exact control names
• exact RMF step ordering
• exact SP 800-207 component naming
• source-level answers that need to be right on the first pass

Available quantizations

Quantization	Approx. size	Use case
F16	14 GB	Full precision reference build
Q8_0	7.5 GB	Higher quality local inference
Q5_K_M	5.1 GB	Balanced size and quality
Q4_K_M	4.4 GB	Small local default for most machines

Run with llama.cpp

git clone https://github.com/ggerganov/llama.cpp
cd llama.cpp
make

wget https://huggingface.co/ethanolivertroy/HackIDLE-NIST-Coder-GGUF/resolve/main/hackidle-nist-coder-q4_k_m.gguf

./llama-cli \
  -m hackidle-nist-coder-q4_k_m.gguf \
  -p "Which NIST docs would you start with for contractor remote access?" \
  -n 500

License

The base model is Qwen2.5-Coder-7B-Instruct, released under Apache 2.0. The NIST source publications used for the dataset are public domain U.S. government works. This model card uses Apache 2.0 for the model artifact and documents the NIST data source separately.