---
license: apache-2.0
base_model:
- ggml-org/gemma-4-12B-it-GGUF
tags:
- ollaya
- gguf
- llama.cpp
- decision-model
- system-one
pipeline_tag: text-classification
---

# cygnet for Ollaya

[Ollaya](https://github.com/ollaya-dev/ollaya) package of **[ggml-org/gemma-4-12B-it-GGUF](https://huggingface.co/ggml-org/gemma-4-12B-it-GGUF)** by blockbrain-ai (recipe) and Google DeepMind (Gemma 4).
Ollaya runs open decision models locally, the way Ollama runs LLMs: typed questions in,
calibrated answers out, behind a TypeSafe-compatible API.

```sh
ollaya run cygnet
```

## What is in this repository

This repository holds only the files Ollaya derives, with no weights. The model is the authors' own GGUF
file: `ollaya pull` downloads it from their repository, unmodified and pinned to a commit, verifies its
sha256, and Ollaya runs it on llama.cpp.

| Tag | The authors' GGUF | Files |
|---|---|---|
| `cygnet:12b` | [ggml-org/gemma-4-12B-it-GGUF@e3e6817](https://huggingface.co/ggml-org/gemma-4-12B-it-GGUF/tree/e3e681731089efaa3f0917336944ac64752db8ba) `gemma-4-12B-it-Q8_0.gguf` | `12b/decision.json`, `12b/calibration.json` |

Each tag has `decision.json` (the prompt, the option labels Ollaya reads and llama.cpp's settings) and
`calibration.json` (temperatures).

## Parity

Ollaya's runner matches stock llama-server of the pinned build (b11146) on the same GGUF, CUDA (RTX 4090): 502 questions, every decision the same, option logits within 7.7e-6 and probabilities within 4.1e-7. The user messages are identical to Cygnet's own shim on 1,364 test prompts.

## License

Same as the upstream model (Apache-2.0). Ollaya itself is Apache-2.0.