Heman10x-NGU/openJev-verdict-2.0

Calibrated 151M Non-Autoregressive Decision Engine beating TypeSafe Jev & Laya on LocalLLaMA/typed-decisions (77.10% acc, 0.0636 Brier, 0.0144 ECE)

Open original source ↗

From the publisher

README & documentation

Source preview

Read the project’s overview, installation instructions and usage examples. The original README is the source of truth.

[](https://huggingface.co/heman10x/openJev-verdict-2.0) [](https://github.com/Heman10x-NGU/openJev-verdict-2.0) [](#the-openjev-verdict-20-benchmark-breakthrough) [](#dual-channel-calibration) [](#single-pass-efficiency) [](#in-browser-webgpu-engine) [](LICENSE) This repository contains code and references for two distinct models: These are inference fixes, not a retrain. The weights are byte-identical to the published checkpoint. Measured on the 231 public JevBench tasks. The update addresses three defects in the inference engine: Model weights and artifacts are hosted on Hugging Face at heman10x/rlcd-modernbert-151m. Full benchmark details and leaderboards are available at Benchmark Heaven Jev Models. An open-source 149.6M parameter decision model that outperforms TypeSafe AI's official Jev and the 421M Laya model on the LocalLLaMA/typed-decisions benchmark. On 2,000 held-out enterprise decisions, openJev-verdict-2.0 delivers 77.10% accuracy (ahead of Laya's 76.60% and Jev's 72.70%), achieves a 0.0636 Brier score (best overall), and drops calibration error to 1.44% ECE on its dedicated confidence head. It achieves this at 2.8x fewer parameters, fine-tuned in 8.8 hours on a budget consumer laptop GPU. If you find this model, benchmark,…

Heman10x-NGU/openJev-verdict-2.0 on GitHub A short preview, not the full document.

Read the full README ↗ · Preview checked 2026-09-21T07:27:40.284Z

Inside the original README — Document outline

Headings are captured from the source. Links open the publisher’s document, not a locally hosted copy.

Links from the README

References supplied by the publisher, not independently verified endorsements. Check the destination before downloading files or entering credentials.

Documentation belongs to its respective authors. Reported project/model license: Other. A listing is not a grant of reuse or training rights. Confirm the document’s own terms at the source.

What this repository does

Calibrated 151M Non-Autoregressive Decision Engine beating TypeSafe Jev & Laya on LocalLLaMA/typed-decisions (77.10% acc, 0.0636 Brier, 0.0144 ECE)

Repository facts

Owner: Heman10x-NGU
Primary language: Python
Stars: 212
Forks: 29
Open issues + pull requests: 1
License: Other
Archived: No
Default branch: main
Created: 2026-09-19T12:14:46.000Z
Last push: 2026-09-20T14:54:40.000Z

Topics and intended use

Owner-supplied topics: agentic-ai, brier-score, calibration, decision-engine, deep-learning, encoder, jev, machine-learning, modernbert, non-autoregressive, onnx, openjev, rlcd, selective-prediction, structured-outputs, system-one, text-classification, typesafe-ai, uncertainty-quantification, webgpu

Review the README for scope, installation, examples and limitations. We do not run repository code or certify it.

Evaluate before installing

Review licensing and dependencies, inspect recent commits and unresolved issues, and test in an isolated environment before production use. Stars and forks alone cannot answer these questions.

README and project files

Issues and maintenance discussion

Releases and changelog

Source and freshness

Source: GitHub. Metadata observed 2026-09-21T06:41:42.808Z. Daily imports are snapshots, not real-time monitoring.

Popularity and source listings do not establish security, suitability, licensing rights or benchmark performance.

Open original source

Related repositories

public-apis/public-apis

EbookFoundation/free-programming-books

donnemartin/system-design-primer

vinta/awesome-python

practical-tutorials/project-based-learning

NousResearch/hermes-agent