User repository on GitHub
when a gopher met Jev
- Forks
- 0
- Last pushed
- Updated Oct 2026
Repo stats from the GitHub API, cached Oct 2026.
CommunityJev Tools & CLIs
Go System One
Native Go runtime for Gemma 4 12B finite-choice decisions, ordered scores and yes/no probabilities, with an NVIDIA path that loads hand-tuned PTX kernels without CGo or CUDA toolkit.
- Stars
- 9
- Language
- Go
- License
- MIT
- Published by
- Community
- Author
- rcarmo
- Last verified
- Oct 2026
About this project
Go System One uses Gemma 4 12B for finite-choice decisions, ordered scores and yes/no probabilities. We started with a fixed Gemma baseline, built a llama.cpp prototype to define the expected results and speed, then replaced it with a native Go runtime.
Quoted from the project's README. Read the full README.
Works forJev for Classification
Used inJev for Software & DevTools
- Languages
- Go
Highlights
- NVIDIA path loads hand-tuned PTX kernels from Go with no CGo, llama.cpp or CUDA toolkit.
- Pinned RTX 3060 request median of 72.80 ms; two-field batches reach 12.01 entries/s.
- Reports 196/231 public-subset JevBench v1.4.0 accuracy with no failed requests.
- Serves Noul, Choice and Score at /go-system-one with a Batch decisions interface.
Watch out
Model files download separately under the Gemma licence and require Go 1.26.2 or the go.mod version. Copy reviewed against the repository README on 2026-10-04, not from running it.
Questions about this entry
Short answers drawn from the linked project or page.
The prototype defined expected results and speed, then was replaced. The current runtime is native Go, with the NVIDIA path calling the Driver API directly and no llama.cpp at run time.
They are pinned-workload figures from the documented benchmarks: 72.80 ms median request and 12.01 entries/s for automatic two-field batches at 100 entries. Your GPU, revision and precision settings will change them.
No. The digest reports 196/231 public-subset accuracy with one fixed runtime revision and no failed requests. It explicitly says this is not the official sealed-set composite score or leaderboard rank.
You need Go 1.26.2 or the go.mod version, downloaded model files, then make setup, artifacts-download with licence acceptance, and make run BACKEND=nvidia LISTEN=127.0.0.1:8080.
Gemma 4 12B. The runtime is MIT; the model files download separately under the Gemma licence, so make artifacts-download needs ACCEPT_GEMMA_LICENSE=1 and an HF_TOKEN.
Noul, Choice and Score questions at /go-system-one on the address passed to make run BACKEND=nvidia LISTEN=127.0.0.1:8080, plus a Batch decisions interface for the boolean and enum context.
84.85% overall, 196 of 231: easy 48/48, original 71/72 and hard 77/111, with no failed requests on one fixed runtime revision. The README says that is public accuracy, not the sealed-set composite or a leaderboard rank.
More Jev resources like this
5GitHub stars
Go gateway exposing the Jev POST /v1/systemone contract over Ollama Cloud, llama.cpp, vLLM, SGLang, and OpenAI-compatible servers by scoring one label token per question.
Jev Tools & CLIs#community#awesomejev#catalog-import
Community
ziozzang/hearimGo
ziozzang/hearimGo3GitHub stars
Go utilities for TypeSafe with a standard-library-only client plus qgrep for files and qlog for Kubernetes pod logs.
Jev Tools & CLIs#community#awesomejev#catalog-import
Community
sd109/typesafe-goGo
sd109/typesafe-goGo3GitHub stars
A self-hosted gateway that pools several official TypeSafe accounts behind one base URL, adding per-user keys with RPM and token quotas, an outbound proxy pool, and automatic account switching on upstream errors.
Jev Tools & CLIs#community#awesomejev#catalog-import
Community
antTing/jev-accounts-hubGo
antTing/jev-accounts-hubGo38GitHub stars
Python memory filter that judges every stored memory against the current request with one calibrated yes/no per memory in a single Jev request, keeping everything above a threshold.
Jev Tools & CLIs#community#jevbest#catalog-import
Community
samdotmak/jev-recallTypeScript
samdotmak/jev-recallTypeScript46GitHub stars
Browser app for systematic-review extraction that opens a trial report with supplements and returns verbatim quotes with file and place, stored locally as projects holding studies.
Jev Tools & CLIs#community#jevbest#catalog-import
Community
choxos/jev-reviewerJavaScript
choxos/jev-reviewerJavaScript34GitHub stars
Pure-HTTP automation for the TypeSafe register, magic-link login, onboarding, and API-key creation flow, with pluggable temp-mail backends, concurrent batch runs, and key verification exports.
Jev Tools & CLIs#community#awesomejev#catalog-import
Community
2951461586/Jev-Register-ToolPython
2951461586/Jev-Register-ToolPythonFrom the Jev guides
Original write-ups on what this entry is for.
The request shape barely changes across surfaces; the surrounding contract does. When to call the API directly and when a gateway, a Worker binding, or a router entry earns its extra hop.
Jev Tools & CLIs#gateway#vercel#cloudflare
Read article
Taking a decision checkpoint out of Python: what the GGUF and browser WebGPU exports preserve, what they change, and the parity evidence the model cards publish.
Jev Tools & CLIs#open-models#local#onnx
Read article
Jev labels a record that already exists. Code moves, flags, or holds it. How the document, mail, and transfer tools keep the action out of the model.
Jev Tools & CLIs#classification#email#security
Read article
From the Jev community
Posts from builders shipping with Jev right now.