---
license: apache-2.0
base_model:
- Mapika/decider-4b
- Mapika/decider-2b
- Mapika/decider-2b-vision
- Mapika/decider-0.8b
library_name: onnx
tags:
- ollaya
- onnx
- decision-model
- system-one
pipeline_tag: text-classification
---

# decider for Ollaya

[Ollaya](https://github.com/ollaya-dev/ollaya) package of **[Mapika/decider-4b](https://huggingface.co/Mapika/decider-4b)** and **[Mapika/decider-2b](https://huggingface.co/Mapika/decider-2b)** and **[Mapika/decider-2b-vision](https://huggingface.co/Mapika/decider-2b-vision)** and **[Mapika/decider-0.8b](https://huggingface.co/Mapika/decider-0.8b)** by Mapika.
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 decider
```

## What is in this repository

This repository holds only the files Ollaya derives, with no weights. Each graph is an ONNX export of the
original model whose weights **reference the authors' own weight files by byte offset**,
so `ollaya pull` downloads the weights from the upstream repositories, unmodified and pinned to a
commit, and verifies their sha256.

| Tag | Upstream | Files |
|---|---|---|
| `decider:4b` | [Mapika/decider-4b@eb5fbdf](https://huggingface.co/Mapika/decider-4b/tree/eb5fbdfc9448473ec25e399882912863afbdb70e) | `4b/model-fp32.onnx`, `4b/decision.json`, `4b/calibration.json` |
| `decider:2b` | [Mapika/decider-2b@9839cc9](https://huggingface.co/Mapika/decider-2b/tree/9839cc9d908be16c5988c0d041034b5fdf82c7a2) | `2b/model-fp32.onnx`, `2b/decision.json`, `2b/calibration.json` |
| `decider:2b-vision` | [Mapika/decider-2b-vision@863e290](https://huggingface.co/Mapika/decider-2b-vision/tree/863e290863655f1d6b69324d77d09ac972d21609) | `2b-vision/model-fp32.onnx`, `2b-vision/vision-fp32.onnx`, `2b-vision/decision.json`, `2b-vision/calibration.json` |
| `decider:0.8b` | [Mapika/decider-0.8b@a0a01d6](https://huggingface.co/Mapika/decider-0.8b/tree/a0a01d6f8135298f400a8c856b355793012ae971) | `0.8b/model-fp32.onnx`, `0.8b/decision.json`, `0.8b/calibration.json` |

Each tag has an fp32 graph, used on CPU and GPU. Each tag also has `decision.json` (sequence layout, special tokens) and
`calibration.json` (temperatures).

## Parity

Ollaya's Rust runtime matches the Python reference exactly on 479 questions (902 rows) per model. The token ids and answer-slot positions are identical, and so is the decision on every question. Probabilities are within 6.3e-6, on CPU and CUDA.

## License

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