---
license: apache-2.0
base_model:
- wfzyx/von
library_name: onnx
tags:
- ollaya
- onnx
- decision-model
- system-one
pipeline_tag: text-classification
---

# von for Ollaya

[Ollaya](https://github.com/ollaya-dev/ollaya) package of **[wfzyx/von](https://huggingface.co/wfzyx/von)** by Victor Hugo Panisa.
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 von
```

## 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 |
|---|---|---|
| `von:1.1` | [wfzyx/von@d8bb5e0](https://huggingface.co/wfzyx/von/tree/d8bb5e0745d8ee1fb65d536d6d4892d54d5a93fd) | `1.1/model-fp32.onnx`, `1.1/decision.json`, `1.1/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 upstream Von, run in float64, on 485 questions (653 rows). The token ids and marker positions are identical, and so is the decision on every question. Logits are within 4.4e-4 and probabilities within 4.7e-5 on x86-64 CPU and CUDA; on Apple silicon's CPU one of the 653 rows is 1.1e-3 off, and every decision is still the same.

## License

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