Jev 发布还不到一周，OpenJEV 也开始扎堆了。
This is a dev post classified by Jev as Other (a list of tools), kept by the Dev Radar because it carries real work, not commentary.
Jev 发布还不到一周，OpenJEV 也开始扎堆了。 我整理了 8 个目前比较有代表性的项目： 1. SemIf 最省事的一派。直接拿现成 Qwen，在模型准备回答 A / B / C 时截住，不让它继续生成文字，只读取每个选项的分数，再换算成概率。 不用重新训练模型，更像是把传统大模型换了一种用法。 https://github.com/TheoLeeCJ/SemIf 2. Simple Jev 思路和 SemIf 类似，但进一步省掉重复计算。 同一段材料只让模型读一次，后面的多个问题共用这部分结果，再分别算答案概率。 作者也说得很清楚，它复刻的是 Jev 的接口和用法，不代表准确率、速度和概率校准已经做到 Jev 水平。 https://github.com/featherless-ai/simple-jev 3. Laya 直接放弃「会写字的大模型」，改用更像传统分类器的编码模型，再加专门的决策头。 模型只有几亿参数，可以做得非常快，但复杂任务还有明显差距。比如一次从 77 个选项里选答案，Laya 准确率 42.5%，Jev 是 87.0%。 https://github.com/NandhaKishorM/laya 4. Von 也是从头做一个专门的决策模型。 Von 只有约 4 亿参数，不生成文字，一次前向直接处理 Choice、No
Posted by 思维怪怪 (7.8k followers) 13 h ago · 86 likes · 6.1k views · view the original post on X. Kept by the Dev Radar as Other. Tools mentioned: semif, simple-jev, laya, von, verdict-open-jev, kev, nimble, openjev.
More dev work like this
- You know whyyy this is crazy — @ItsmeAjayKV
- 🔥 ZastTranslate 1.21 is officially LIVE — powered by local Qwen-Image-2.1 in Pinokio… — @zast57
- it is mind blowing that you can now run an Adobe Photoshop-grade AI image model on an… — @cryptopunk7213
- google shipped 2 nano bananas this year and a 7b model just put both of them behind it. — @sudoingX
- Resolvi iniciar mais 3 novos pequenos projetos de retrogames: — @AkitaOnRails
- WTH, Qwen just dropped a 7B open-weight image model that beats Google’s Nano Banana 2. — @choblin29
- We now have an open-weight 7B model that outperforms Nano Banana 2.0. Just let that sink… — @kimmonismus
- Impressive flex of image editing robustness from Qwen — @teortaxesTex
Every post is read and classified by Jev (TypeSafe): what it is, which market it belongs to, and whether the link is a real tool. 16.1k posts from 4.9k X accounts over the last 21 days, 1.8k tools, 12 markets. Collected every 5 minutes, fully re-ranked every hour — last update 2026-09-20 19:24 UTC. Full method.