User repository on GitHub
- Forks
- 1
- Last pushed
- Updated Sep 2026
Repo stats from the GitHub API, cached Oct 2026.
CommunityJev Models & Reimplementations
laya-Ascend
Ascend NPU fork of the Laya decision checkpoints that answers Jev-style Choice, Score and Noul questions in a single forward pass on Huawei 910B hardware. It reports 37-47 ms median for a four-question request, 33.8x-70.9x faster than one container CPU thread.
- Stars
- 110
- Language
- Python
- License
- APACHE-2.0
- Published by
- Community
- Last verified
- Oct 2026
About this project
Multilingual, non-autoregressive System 1 decision engine. Typed decisions over 100+ languages in a single forward pass — 33 ms — trained with reinforcement learning against strictly proper scoring rules (RLCD), with a router that picks the right checkpoint per request. Laya is a multilingual, non-autoregressive System 1 decision engine for typed decisions over text, with CPU and Ascend NPU inference support.
Quoted from the project's README. Read the full README.
Used inJev for Software & DevTools
- Languages
- Python
- Capabilities
- Open sourceSelf-hostedRuns locally
Highlights
- Fork of the Apache-2.0 Laya checkpoints adding Huawei Ascend 910B inference support
- 37-47 ms median for a four-question request, 33.8x-70.9x faster than one container CPU thread
- An output-equivalent SDPA decision head avoids torch_npu's CPU fallback on aten::_transformer_encoder_layer_fwd
- Answers choice, score and noul questions over 100+ languages with a router that picks a checkpoint per request
- README-Ascend.md covers CANN environment setup, package installation and model download
Watch out
Apache-2.0. Needs a working Ascend/CANN environment. The repo is a fork of Laya, an independent decision engine rather than TypeSafe's hosted Jev; the shared typed question shapes are the connection.
Questions about this entry
Short answers drawn from the linked project or page.
No. Laya is an independent open-source decision engine whose typed choice, score and noul questions follow the same System One-shaped interface; this fork ports the Laya checkpoints to Ascend NPUs.
It adds NPU inference support for Huawei 910B, including a rewritten SDPA decision head that avoids torch_npu's CPU fallback, with 37-47 ms medians measured for four-question requests.
Python 3.10 or newer, installed with `pip install laya`. The dependency floor comes from huggingface_hub 1.x, transformers 5.x and torch 2.14 all requiring 3.10. Ascend NPU setup is separate: README-Ascend.md covers CANN environment setup, package installation and model download.
Three: laya (ModernBERT-large, 421M, 512 context, English), laya-multilingual (mmBERT-base, 322M, 1024 context, 100+ languages) and laya-typed-decisions (ModernBERT-large, 1024 context). A Router detects script and language in under 0.5 ms before the forward pass and dispatches to the best checkpoint.
The English checkpoint collapses on non-Latin scripts, scoring 0.000 accuracy on Khmer at 0.952 confidence. Because it stays confident while wrong, confidence gating cannot catch that, so the Router detects the script before inference and sends it to the multilingual checkpoint.
At the default max_loaded=1, alternating languages rebuilds a model on every request, measured at a 7.4 s median reload on CPU and 10.3 s on T4. Router(preload=True) keeps checkpoints resident and cuts latency to 32.8 ms on GPU or 193-464 ms on CPU, with no reloads.
The examples directory ships Laya-driven Snake and Tetris demos with terminal playback and web Canvas UIs, runnable on CPU or Ascend NPU. The Router accepts device selection, such as Router(preload=True, device="cuda"), and attach() reuses an already-built agent to avoid duplicate VRAM.
More Jev resources like this
877GitHub stars
A local server that reads next-token probabilities from a fine-tuned Spark model and returns Choice, Score, and Noul answers with zero generated tokens.
Jev Models & Reimplementations#community#python#open-models
Community
Rizzo-AI-Academy/rizzo-flowPython
Rizzo-AI-Academy/rizzo-flowPython130GitHub stars
An open recipe for a 27B multimodal decision model trained with full-weight SFT on 73,000 examples over one H200, serving choice, noul and score on /v1/systemone with per-checkpoint provenance and released calibration plots.
Jev Models & Reimplementations#community#calibration#qwen
Community
denis-pplx/autojevPython
denis-pplx/autojevPython109GitHub stars
Open decision models from a self-improvement loop that publishes every experiment, failures included. They answer Noul, Choice and Score questions on the same POST /v1/systemone wire format as Jev, returning a calibrated probability per option in one forward pass; the 27B scores 65.64 on the Decision Index 0.3 set.
Jev Models & Reimplementations#community#decision-models#system-one
Community
Shanghua-Gao/RSI-JevHTML
Shanghua-Gao/RSI-JevHTML58GitHub stars
A local /v1/systemone server that reads label probabilities from a GGUF chat model, with prompt repetition and a confidence shrink toward uniform.
Jev Models & Reimplementations#community#python#open-models
Community
Mushroom-Systems/lichenPython
Mushroom-Systems/lichenPython17GitHub stars
Jev-compatible decision models with vision input: LoRA fine-tunes of Qwen3.5 at 4B and 9B serving Choice, Score and Noul on the /v1/systemone wire format, with screenshots and photos placed beside text in the state. Runs on your own NVIDIA GPU; weights are non-commercial.
Jev Models & Reimplementations#community#vision#open-weights
Community
Xiaooolong/vevPython
Xiaooolong/vevPython10GitHub stars
A local Jev-compatible server that reads Choice, Score and Noul answers from the logits of any MLX checkpoint or OpenAI-compatible endpoint with no training, serving POST /v1/systemone as a drop-in for the official TypeSafe SDK.
Jev Models & Reimplementations#community#mlx#calibration
Community
bradAGI/rulingPython
bradAGI/rulingPythonFrom the Jev guides
Original write-ups on what this entry is for.
Jev takes text only; Clef-omni adds image, audio, and video input and Liquid d1 adds images, each keeping the same typed-answer shape. Where that matters, where it does not, and the costs the vendors underplay.
Jev Models & Reimplementations#multimodal#open-models#decision-model
Read article
One model in September, a dozen by October. Clef, Liquid d1, Perplexity, and OpenAI all ship the same typed-answer shape now. How to read the rival benchmarks and pick on your own task.
Jev Models & Reimplementations#decision-model#benchmarks#pricing
Read article
Hundreds of repositories carry the system-one tag, but the ecosystem is a dozen ideas in different sizes. A tour of the families, the licenses, and what a download does not include.
Jev Models & Reimplementations#open-models#hugging-face#local
Read article
From the Jev community
Posts from builders shipping with Jev right now.