User repository on GitHub
Jev decision models: Korean Slidev course and hands-on Jupyter workshops
- Forks
- 0
- Last pushed
- Updated Oct 2026
Repo stats from the GitHub API, cached Oct 2026.
Jev directory
CommunityJev Guides & Articles
Jev Decision Models Course (Korean)
A Korean Slidev course — 59 slides, 110 minutes plus deep-dive Q&A — with three standalone Jupyter notebooks that take Jev decision models from NLL/Brier/ECE calibration to training a Qwen-based Jev-compatible model with Unsloth and testing its HTTP API.
- Stars
- 0
- Language
- Jupyter Notebook
- Reading time
- 2 min
- Published by
- Community
- Author
- heury
- Last verified
- Oct 2026
- Languages
- Python
- Capabilities
- Open source
Highlights
- A 59-slide Korean course for deep-learning practitioners with an NVIDIA GPU server, covering concepts, loss, evaluation, training and serving.
- Three notebooks: probabilities and metrics (NLL/Brier/ECE, temperature scaling, deferral policy), Qwen+LoRA decision-head training, and decision-API testing.
- Notebook 02 trains Qwen into a Jev-compatible decision model with Unsloth on a Linux NVIDIA GPU; notebook 01 runs on CPU.
- The README says Jev itself is an API model; the workshop trains a compatible local model instead and documents sources and code revisions in sources.md.
- Run the slides with npm ci and npm run dev at localhost:3030; npm run export builds a PDF after npx playwright install chromium.
Quickstart
bash
npm ci
npm run devWatch out
README reviewed 2026-10-17 from the round-eleven capture, which is truncated at a package-version note: exact dependency pins, notebook outputs, the referenced pytorch.kr article and the repository's licence (none declared on GitHub) are outside the digest. The material is in Korean.
Questions about this entry
Short answers drawn from the linked project or page.
A Korean-language study pack for Jev decision models: a 59-slide Slidev course plus three independently runnable Jupyter notebooks covering probabilities, calibration, training a compatible model and testing the decision API.
Deep-learning practitioners who read Korean and have access to an NVIDIA GPU server, since the training notebook targets Linux CUDA. Notebooks 01 and 03 run on CPU.
Install Node.js 22.12 or newer, run npm ci and npm run dev, then open http://localhost:3030; the presenter view is at /presenter/. npm run build produces a static site.
It trains Qwen into a Jev-compatible decision model using Unsloth with a LoRA and a decision head on real data, then calibrates and verifies saving and reloading, per the README's file table.
Yes. Notebook 01 covers synthetic logits, NLL/Brier/ECE metrics, temperature calibration and deferral policy on CPU, and notebook 03 smoke-tests the decision API's response contract and latency in Korean.
The README says Jev is an API model and that the workshop trains a Jev-compatible model locally with Unsloth based on a referenced article and official material, so this is a compatible-model exercise rather than the hosted API.
The material is Korean-only, notebook 02 requires a Linux NVIDIA GPU, and no licence is declared on GitHub. The digest truncates at a package-version note, so outputs and exact pins are unverified.
More Jev resources like this
A free animated course on System One decision models, covering TypeSafe Jev alongside OpenAI Decisions, Cloudflare Clef and AWS Strands Decider: 22 short lessons backed by 199 sources and 14,318 API calls the authors ran themselves.
Jev Guides & Articles#community#github#course
Community
Haple/system-oneJavaScript
Haple/system-oneJavaScriptEleven small runnable projects for learning Jev and other System One decision models, each lesson pairing a one-command web demo with a whiteboard and README, from the three question types through jobs sorting and contract review to Jev inside a SQL query. Every demo makes live API calls.
Jev Demos & Experiments#community#github#python
Community
aianytime/jev-crash-coursePython
aianytime/jev-crash-coursePythonA one-hour, six-step workshop built around a fighting game that teaches TypeSafe's Discriminative (Jev) decision model and puts it next to Gemini in a Google ADK workflow. It runs in Cloud Shell or anywhere signed into gcloud, with hosted Jev, a local DiffusionGemma VM, or a rehearsal stand-in.
Jev Guides & Articles#community#github#workshop
The community hub for System One decision models, live at systemone.dev. It teaches building with models that answer typed questions about a program's state using calibrated probabilities instead of generated text, and its samples work unchanged with Clef and Jev.
Jev Guides & Articles#community#github#docs
Community
systemonedev/systemone.devPython
systemonedev/systemone.devPythonChinese-language Jev cookbook with runnable Jupyter notebooks covering the three primitives, 18 official recipes, evaluation method, voice smart-home demo, and local Laya fine-tuning.
Jev Guides & Articles#community#jevbest#catalog-import
A community curriculum of Jupyter notebook recipes that teach typed decisions with Jev, TypeSafe's System One model. Each recipe takes one bounded Choice, Noul or Score decision over text or image state, spanning first classification through closed-loop control.
Jev Cookbooks#community#github#python
From the Jev guides
Original write-ups on what this entry is for.
From the Jev AI model community
Posts from builders shipping with Jev right now.