TianyuCodings/NanoJev
A nano replica of Jev: parallel decisions, dynamic candidates, and an end-to-end training pipeline.
Older snapshot: check the original source for current information.
From the publisher
README & documentation
Source previewRead the project’s overview, installation instructions and usage examples. The original README is the source of truth.
English | 简体中文 A 0.6B parallel decision model: states and questions in, complete probability distributions out. Zero output-token decoding. Play ViZDoom · Maze & Snake · Model · Dataset Now playing ViZDoom: one shared checkpoint handles Basic aiming and Predict Position's moving-target rocket shots, alongside Maze and Snake. 4 tasks · 18,760 decision questions per data variant · 896 Predict Position expert episodes September 20, 2026 — One model, four games. Real browser replays of Jev, NanoJev and Untuned Qwen. These animations loop automatically; click either one to open its interactive player. All four demos use the same current NanoJev checkpoint. The interactive development site currently requires access; all recordings can also be played locally using the commands below. [](https://nanojev-dev.tianyuchen99.chatgpt.site/?autoplay=1) Move into position, line up the target, fire. NanoJev eliminates the target with one shot in 1.40 s; Jev and Untuned Qwen each fire 19 shots without an elimination before…
Read the full README ↗ · Preview checked 2026-10-28T22:24:40.749Z
Inside the original README — Document outline
- NanoJev — A nano replica of Jev
- What's new
- Three models, side by side
- ViZDoom Basic · Aim, then fire
- Find the exit · 50×50 Maze
- What NanoJev does
- Held-out gameplay
- Dataset scale and model
- Release version and training run
- Quick start
- Development notes
- Roadmap
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: MIT. A listing is not a grant of reuse or training rights. Confirm the document’s own terms at the source.
What this repository does
A nano replica of Jev: parallel decisions, dynamic candidates, and an end-to-end training pipeline.
Repository facts
- Owner
- TianyuCodings
- Primary language
- Python
- Stars
- 2,558
- Forks
- 263
- Open issues + pull requests
- 9
- License
- MIT
- Archived
- No
- Default branch
- main
- Created
- 2026-09-17T16:08:30.000Z
- Last push
- 2026-10-16T01:00:52.000Z
Topics and intended use
No topics were included in the latest source metadata.
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.
Source and freshness
Source: GitHub. Metadata observed 2026-10-17T12:10:02.740Z. Daily imports are snapshots, not real-time monitoring.
Popularity and source listings do not establish security, suitability, licensing rights or benchmark performance.
Related repositories
EbookFoundation/free-programming-books
donnemartin/system-design-primer