@hiwise_aigc: Yesterday Qwen-image2.1 "de-censored version" post blew up, and a bunch of friends DM'd me asking how to set it up them…
Summary
作者分享了 Qwen-image2.1 "去审查版" 的本地部署教程，称在仅 8G 显存的 RTX 5060 上即可运行，并整理了 5 个常见坑点，提醒大家适度使用。
Similar Articles
@hiwise_aigc: https://x.com/hiwise_aigc/status/2109255757845914013
An introduction to AtomicChat's "decensored" text encoder for Qwen-Image-2.1-Turbo, released on HuggingFace in GGUF quantized form, plus a complete tutorial for running it locally with ComfyUI on 8GB VRAM (RTX 5060). Includes the full file download checklist, node configurations, 5 real-world gotchas with their fixes, and prompt examples.
@Lonely__MH: Hesitantly asking, Qwen-Image-2.1 local deployment How's the image generation speed?!
Discussing the image generation speed of Qwen-Image-2.1 local deployment, congratulating its release, and highlighting its breakthroughs in small parameter counts and high efficiency, positioning it as a potential leader in domestic AI image generation.
@wquguru: Local deployment of models can't do without some acceleration tricks. Simply put, you can switch to models like Turbo; …
作者系统整理了本地部署模型的加速方案，包括在 macOS 上结合 Turbo、MLX、VAE Q8 和前缀 KV cache 让 Qwen-Image 2.1 提速近 5 倍，以及用 Uzu 的投机解码把 Qwen3.5-9B 从 22 tok/s 提升到 92 tok/s、用 TensorRT 让 Qwen 在 RTX 5090 上比 llama.cpp 快 3 倍等案例。
@MinLiBuilds: https://x.com/MinLiBuilds/status/2107988055101685781
The author shares a complete guide for deploying Qwen3.8-Flash-Next (125B-A6B) on 12GB VRAM (RTX 5070) with Strata, achieving nearly 100 tok/s, and lets Ant Group's newly released Ling-3.1-flash Agent handle the entire workflow from scratch — machine check, download, installation, troubleshooting, and benchmarking.
GPT-Image-style photo editing on your own GPU (12 GB+): I built a zero-setup app for Qwen-Image 2.1 on stable-diffusion.cpp
Kokon 是一款零配置的开源桌面应用，让用户在本地 GPU（12GB+）上直接使用 Qwen-Image 2.1 进行 GPT-Image 风格的照片编辑，基于 stable-diffusion.cpp，内置 MCP 服务器和 HTTP API。