What does duan decide?

It answers 3 kinds of question: choice picks one option from a list you give it; score rates the state on a scale you define; noul returns the probability that a condition holds. Every answer is a typed value with a probability, never free text.

The registry lists it without files: install it from the source repository linked on this page.

Can I use duan commercially?

It is published under cc-by-nc-4.0. Read the licence before shipping; the model card may add terms of its own.

What is duan built on?

Its base model is convaiinnovations/laya-multilingual, a laya architecture. The Model tree section shows the whole chain.

How much can duan read at once?

The publisher has not recorded a context length. The model has 322M parameters.

duan: open AI decision model by Weipeng Han (wipen)

Duan

Weipeng Han's healthcare decision model: a full fine-tune of Convai's Laya-multilingual (322M, mmBERT-base) that answers choice, score and noul questions on English and Chinese clinical cases in one forward pass, with a probability per option. Non-commercial weights.

The author marks it work in progress and makes no clinical performance claim; it is not for autonomous diagnosis or prescribing, and low-confidence cases are meant to go to a person. Training data mixes clinical outcome labels, exam answer keys, expert consensus, literature labels and fully synthetic labels; the 0.99 differential-diagnosis accuracy is on synthetic data. Per-bucket temperature scaling is applied, but ECE ranges from 0.007 to 0.456 by task. The Hugging Face repo is click-through gated; a ModelScope mirror (wipenhan/Duan) is public. Weights CC BY-NC 4.0 because one training source is non-commercial; the serving code is Apache-2.0. Runs on CPU.

What it decides

choice — picks one option from a set
score — places the input on an ordered scale
noul — answers a yes/no question with one probability

At a glance


Parameters	322M
Base model	`convaiinnovations/laya-multilingual`
Maker	Weipeng Han (wipen)
Released	2026-09-27
License	cc-by-nc-4.0

Get the weights

pip install systemonemodels
systemone pull wipen/duan

The files are served from the maker's Hugging Face repository, wipen/Duan, and verified against the checksums recorded here.

Playground

Build a request for wipen/duan. The form and the JSON are the same object; the response panel shows the shape every System One server returns.

Input

StateThe context the questions are asked about. Up to the model's context length.

A yes/no question. The answer is one probability that the condition holds.

Instructions

Response

Placeholder values, in the shape a run returns.

safenoul7%

This model doesn't answer live here. Send this request to your own server for this model; every System One server speaks this shape.

Call it

curl https://api.systemonemodels.ai/v1/systemone \
  -H "Authorization: Bearer $SYSTEMONE_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{"model":"wipen/duan","state":"Task: clean up inactive accounts before the quarterly report.\nProposed tool call: delete_rows(table=\"customers\", where=\"last_login < 2023-01-01\")\nContext: the customers table is production; there is no backup newer than Friday.","questions":{"safe":{"type":"noul","instructions":"Is this action safe to run without a human approving it first?"}}}'

Weipeng Han (wipen): duan

Duan

What it decides

At a glance

Get the weights

Read more

Playground

Weipeng Han (wipen): duan

Duan

What it decides

At a glance

Get the weights

Read more

Playground

Providers

Pricing

Performance

Model tree

Activity

Documentation

systemone.yaml

FAQ