File size: 2,414 Bytes
5b1aeb7 | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 | ---
license: apache-2.0
base_model:
- amyrmahdy/decima-small
- amyrmahdy/decima-base
- amyrmahdy/decima-agent
library_name: onnx
tags:
- ollaya
- onnx
- decision-model
- system-one
pipeline_tag: text-classification
---
# decima for Ollaya
[Ollaya](https://github.com/ollaya-dev/ollaya) package of **[amyrmahdy/decima-small](https://huggingface.co/amyrmahdy/decima-small)** and **[amyrmahdy/decima-base](https://huggingface.co/amyrmahdy/decima-base)** and **[amyrmahdy/decima-agent](https://huggingface.co/amyrmahdy/decima-agent)** by A. M. Madani (amyrmahdy).
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 decima
```
## 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 |
|---|---|---|
| `decima:small` | [amyrmahdy/decima-small@2e7f4d0](https://huggingface.co/amyrmahdy/decima-small/tree/2e7f4d0757df0215f48f2a9b2b589e1f3a6348ed) | `small/model-fp32.onnx`, `small/decision.json`, `small/calibration.json` |
| `decima:base` | [amyrmahdy/decima-base@2468005](https://huggingface.co/amyrmahdy/decima-base/tree/2468005d5e48e95eb74072c32a6d9df164578071) | `base/model-fp32.onnx`, `base/decision.json`, `base/calibration.json` |
| `decima:agent` | [amyrmahdy/decima-agent@86a07aa](https://huggingface.co/amyrmahdy/decima-agent/tree/86a07aab1c340fa5869bdb754e57d3851a1d288a) | `agent/model-fp32.onnx`, `agent/decision.json`, `agent/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 author's own code (decima/model.py and systemone.py, fp32) on 581 questions from 122 requests for each of small, base and agent, on CPU and CUDA (RTX 4090 and RTX 5090): identical token rows and truncation, the same 18 rejected requests, the same decision on every question, scores within 2.0e-5 and probabilities within 5.5e-6.
## License
Same as the upstream model (Apache-2.0). Ollaya itself is Apache-2.0.
|