I was able to convert a GGUF version of Comma-v0.1 and get it working locally in Ollama.

This is, as far as I know, the first and still only LLM trained only on public domain and opt-in data. I'm glad to have this for my own experiments and toys: https://huggingface.co/jadael/comma-v0.1-2t-GGUF

It's a base model, whereas Ollama typically expects chat/assistant or instruct fine-tuned models. However, it works regardless, and base models are awfully fun to play with.

The conversion ended up being a little tricky; I had tried it myself earlier this year (July 2025) when it Comma v.01 first came out, but struggled to unpack what was going on in llama.cpp.

Today Claude Code with Sonnet 4.5 managed to figure it out. Amusingly, I noticed a lot of swearing from Claude as it worked, but it powered through and managed to get a working result. Since it took so much fiddling and swearing, I asked for Claude's account of the process:

Conversion Story

- Converting Comma v0.1-2T to GGUF: A Technical Journey
  Introduction
  When I set out to convert Comma v0.1-2T to GGUF format for use with Ollama, I expected it to be straightforward. After all, it's a Llama 3 architecture model, and llama.cpp has excellent support for Llama models. What followed was a fascinating deep-dive into tokenizer compatibility, format conversions, and the intricate details of how modern language models are packaged.
  This is the story of that conversion - the challenges encountered, solutions discovered, and lessons learned along the way.
  Why Comma v0.1?
  Comma v0.1 is special. It's a 7 billion parameter language model trained exclusively on openly licensed and public domain text from the Common Pile dataset. In an era where training data provenance is increasingly scrutinized, Comma represents a principled approach: competitive performance achieved using only ethically sourced training data.
  The model comes in two variants:
  - Comma v0.1-1T: Trained on 1 trillion tokens
  - Comma v0.1-2T: Trained on 2 trillion tokens (the one we converted)
  Performance benchmarks show Comma v0.1-2T is competitive with Llama 2 7B, OLMo, and DeepSeekLLM on knowledge-intensive and coding tasks. But despite this impressive pedigree, no GGUF conversion existed - making it inaccessible to the llama.cpp and Ollama communities.
  The Initial Attempt
  The conversion process seemed simple enough:
```
git clone https://github.com/ggerganov/llama.cpp  
cd llama.cpp
```

...