😧 我靠！Google 用 vLLM patch 把 DiffusionGemma 当 Jev 跑，本地暖机很多任务比云端 Jev 还快，智能基本打平，还自带视觉！
This is a dev post classified by Jev as AI dev tools (a tool drop), kept by the Dev Radar because it carries real work, not commentary.
😧 我靠！Google 用 vLLM patch 把 DiffusionGemma 当 Jev 跑，本地暖机很多任务比云端 Jev 还快，智能基本打平，还自带视觉！ 对比亮点： 🔹暖机 one-read：tickets 410ms vs 590ms，reviews 119ms vs 135ms，PII 641ms vs 931ms 🔹一次 canvas 并行去噪，不是逐 token 生成，Spark 上约 0.2s 🔹双向注意力，选项同时看完整上下文，置信度更稳 🔹继承 Gemma 4 多模态，Jev 目前纯文本 🔹开源可本地跑，26B MoE（激活 3.8B），量化约 18GB 怎么用： 1️⃣克隆 http://github.com/mmastrac/djev-spark 2️⃣下载 nvidia/diffusiongemma-26B-A4B-it-NVFP4 3️⃣docker compose up 4️⃣POST localhost:8011/v1/systemone（Jev 同款 state + questions） #Google #DiffusionGemma #Jev #Gemma #vLLM #开源AI
Posted by SuSu_酥酥👅 (15.6k followers) 18 h ago · 235 likes · 19.1k views · view the original post on X. Kept by the Dev Radar as AI dev tools. Tools mentioned: djev-spark, vllm.
More dev work like this
- 架构设计： — @mylifcc
- Multi-agent prototypes stall when orchestration becomes the project. — @DanKornas
- Rebuilding a site you own shouldn’t mean starting from a blank Next.js app. — @DanKornas
- NVIDIA wants coding agents to stop wasting tokens on repeated work. — @agenticgirl
- Runs 69 product management skills and 42 workflows in Claude Code and Cowork. — @tom_doerr
- Zai bros just opensourced their ZCode harness 🔥🔥 — @zainhas
- Your AI agents keep rediscovering work. This repo gives them a shared folder. — @DanKornas
- We made an easy image loader for Qwen Image 2.1. Now you can generate your images on the… — @Blackfrost_AI
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. 17.6k posts from 4.9k X accounts over the last 21 days, 2k tools, 12 markets. Collected every 5 minutes, fully re-ranked every hour — last update 2026-09-21 02:20 UTC. Full method.