lyuyiqi/open-jev-fast

Faster inference backend for Open-Jev-27B: fused CUDA kernels, prefix tree, CUDA Graphs (B300, bf16)

Language modelsCuda

1.3 Steady Magnitude out of 10 — how fast interest is growing, not a quality score. Star growth looks organic. Data as of September 30, 2026, 16:35 UTC.

Open on Seismograph Open on GitHub

About the project

A fast inference backend for Open-Jev-27B that replaces the model forward pass with hand-written CUDA kernels, a prefix tree, and CUDA Graphs, acting as a drop-in replacement for Open-Jev's own local server.

Useful for

Speed up a local Open-Jev-27B server by swapping PyTorch inference for CUDA kernels with the same accuracy
Run JevBench task batches with lower latency on a single B300 GPU
Benchmark stock Open-Jev against the specialized kernels on your own requests

README summarized by DeepSeek V4.1 Flash. Details may be inaccurate.

Why it’s trending

4 stars so far today, about 5 expected by the end of the day.
The spike has held for 2 days in a row — not a one-off blip.
The repository is 3 days old and already has 87 stars. With less than two weeks of history, there's no usual pace to compare the spike against yet.
Top new repositories this week: #196.

Stars per day

Bars are daily stars, the line is the usual pace. Red marks spike days.

Numbers

Total stars: 87
Today: 4 · ≈ 5 by evening
Forks: 6
Issues and pull requests: 0
Watchers: 0
Language: Cuda
License: MIT
Created: September 27, 2026
Last push: September 28, 2026

Star trust

Growth looks organic: forks and discussion are in line with active projects, and stars arrive unevenly, the way people give them.

These are heuristics, not a verdict: we judge by the repository’s behavior, not by a list of stargazers.

Spotted in

September 30, 2026Top new repositories this week: #195

README badge

Paste it into your README — the badge shows the current magnitude and links to this page.

[![Seismograph](https://gitnova.dev/badge/lyuyiqi/open-jev-fast.svg?lang=en)](https://gitnova.dev/en/r/lyuyiqi/open-jev-fast)

More in this category

8.1
Niko1221/Strata
A C++ local inference engine that runs the 125B MoE model Qwen3.8-Flash-Next on a regular PC with one NVIDIA GPU (12-24 GB) and 64 GB RAM, exposing an OpenAI/Anthropic-compatible API on localhost.

BreakoutLanguage modelsC+++618 stars today, ≈ 883 by evening
7.7
VectifyAI/PageIndex
A vectorless RAG engine that builds a hierarchical tree index of a document and retrieves by LLM reasoning over that tree, the way a human navigates a report. Aimed at long professional PDFs such as financial, legal…

BreakoutLanguage modelsPython+350 stars today, ≈ 564 by evening
7.4
firelex/jeff
Small fine-tuned Qwen3.5 and Gemma 4 models for zero-shot classification: given a situation description and a list of options, they return a calibrated probability for each option in a single forward pass.

Early signalLanguage modelsPython+63 stars today, ≈ 128 by evening
7.0
PostHog/jeeves
Jeeves is a reasoning model built on Qwen3.5-9B (LoRA + pointer head), trained with SFT and CISPO, that thinks before answering and returns calibrated probabilities for yes/no, multiple-choice and rating questions via…

Early signalLanguage modelsPython+25 stars today, ≈ 62 by evening
6.9
ninjahawk/livenerf
A benchmark for tracking whether a frontier model quietly degrades after release: it runs a frozen question panel daily and statistically measures drift in accuracy and token counts against the launch-week baseline.

Early signalLanguage modelsPython+264 stars today, ≈ 375 by evening
6.4
deepseek-ai/DeepGEMM-Ascend
A port of DeepGEMM to Huawei Ascend: GEMM kernels (BF16, FP8, FP4, MQA logits, MegaMoE) with an API compatible with DeepGEMM, targeting peak NPU performance.

Early signalLanguage modelsC+++159 stars today, ≈ 226 by evening