Phocinae-Largha-150M-v1
Phocinae-Largha-150M-v1 is a 144.3M-parameter bilingual model for structured decisions, such as yes/no, pick-one, and score questions about a state. It is not a chat model and is designed to run locally in a single forward pass.
What it does
Use it locally to make deterministic decisions from a state and a set of closed-choice questions, such as command approvals, tool selection, or step validation.
- Availability
- Weights downloadable from Hugging Face; runtime available via phocinae-server
- Pricing model
- No API token costs; weights downloaded once
- Deployment
- Local runtime; CPU supported; service listens on 127.0.0.1
- Privacy
- Claims input stays on-device; no telemetry or callbacks
- Licence
- Apache-2.0; base model mmBERT-small is MIT
Capability
Modality
Access
Deployment
Pricing
Target user
Evidence
6 claims- The model accepts typed yes/no, choice, and 2–10 score questions and returns answers with confidence in one forward pass. [1]
**Phocinae-Largha-150M-v1 (Largha) is a 144.3M-parameter bilingual typed decision model — not a chatbot.** Given a `state` plus typed questions (`noul` yes/no · `choice` pick-one · `score` 2–10), it returns calibrated answers with confidence in a single forward pass, emitting zero output tokens.
Hugging Face. Quoted text is outside CC BY 4.0 and keeps its original rights. - Model weights are downloadable from the Phocinae-Largha-150M-v1 Hugging Face repository. [1]
- The official runtime is phocinae-server, and the model can run without a GPU. [1]
> ⚠️ **This repo is not a chat model, and it does not load through the usual `transformers` path.** Use [phocinae-server](https://github.com/Phocinae/phocinae-server).
Hugging Face. Quoted text is outside CC BY 4.0 and keeps its original rights. - The service listens only on 127.0.0.1 and is described as having no telemetry or callbacks. [1]
- The default runtime rendering path, including phocinae-server 0.1.6, scores 0.5140 on the stated 2,000 decisions; a caller-side override reaches 0.9055. [1]
- The stated licence for the model is Apache-2.0. [1]
**It cannot explain itself.** A generative decision model can hand you a one-line reason for a verdict; Largha cannot — that is the trade for determinism and zero output tokens.
Hugging Face. Quoted text is outside CC BY 4.0 and keeps its original rights.
Sources
1
1
Hugging FaceOwner-controlled
https://huggingface.co/Phocinae/Phocinae-Largha-150M-v1
Retrieved 2026-10-28 00:39 UTC
Verified
Versions
6| Version | Effective | Change | Changed fields | State |
|---|---|---|---|---|
| 6 | 2026-10-16 08:29 UTC | material change | capability, access, deployment, pricing, availability, pricing model, deployment, privacy | Current |
| 5 | 2026-10-12 12:01 UTC | material change | capability, access, pricing, deployment, privacy | Current |
| 4 | 2026-10-11 12:54 UTC | material change | access, deployment, availability, deployment, privacy | Current |
| 3 | 2026-10-10 13:50 UTC | material change | capability, access, pricing, availability, pricing model, deployment, privacy | Current |
| 2 | 2026-10-09 14:43 UTC | material change | capability, access, deployment, pricing, availability, deployment, privacy | Current |
| 1 | 2026-10-08 15:33 UTC | publication | None | Current |
In briefs
2- Turn a prompt into an editable slide, poster, or dashboard Weekly brief 2026-W41
- Extract document layout and chart data, not just text Daily brief 2026-10-08
Original catalog content is licensed under CC BY 4.0, attributed to wowsuchai. Source quotations and linked content keep their original rights.