Download
1 variant available
1810 1 2 3 4 5 6 7 8 90 1 2 3 4 5 6 7 8 90 1 2 3 4 5 6 7 8 9
(18)
Aug 6, 2026
Flux.2 Klein 9B
Euler + simple, 4-12 step, cfg1-2, 文生图-T2I:Euler + simple, 4-10步,cfg1-1.5,编辑-E2I:Euler + simple, 4-12步,cfg1-2,
编辑功能强化,升级了一致性,Lora支持的比较好。依旧不包含NSFW内容,因为klein的NSFW内容有针对性,达不到我的需求,我自己训练的升级效果也不佳,所以需要NSFW内容的话照你的需求自行添加lora就好。文生图不是强项。
The editing functionality has been enhanced, consistency has been improved, and LoRa support is quite good. It still doesn't include NSFW content because Klein's NSFW content is specific and doesn't meet my needs; my own training and upgrade results were also poor. Therefore, if you need NSFW content, add LoRa content according to your requirements. T2I is not a strong point.
我也不知道为什么klein就是不让我打出BFS三个字母,总是出现错字,也许这也是特点吧。
I don't know why Klein just won't let me type the letters BFS; it always makes typos. Maybe that's just a feature of its system.
Show more

1K0 1 2 3 4 5 6 7 8 9K
2.6K0 1 2 3 4 5 6 7 8 9.0 1 2 3 4 5 6 7 8 9K
The FLUX.1 [dev] Model is licensed by Black Forest Labs. Inc. under the FLUX.1 [dev] Non-Commercial License. Copyright Black Forest Labs. Inc.
IN NO EVENT SHALL BLACK FOREST LABS, INC. BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH USE OF THIS MODEL.
Based on Flux.2 Klein 9B Distilled, the editing functions have been enhanced. It incorporates my own image-trained editing-oriented Lokr and BFS face-swapping enhancement with Realtime fine-tuning, primarily focusing on adjustments to the editing capabilities.
基于Flux.2 klein 9B Distilled,编辑功能强化,融合了我自己使用对图训练的编辑型lokr和BFS换脸强化+Realtime微调,主要是针对编辑功能做出调整。
Online usage: TensorArt | TensorHub | SeaArt
Set up : Euler + simple, 4-12 step, cfg1-2.5,
文生图-T2I:Euler + simple, 4-10步,cfg1-1.5,编辑-Edit:Euler + simple, 4-12步,cfg1.5-2,
Resolution : 832x1216 for 1024x1536, upscaling : 1280x1920 or 1536x2240。
No NSFW content has been added or beautified, allowing users to easily add their own Lora tools for compatibility. Distilled functionality remains effective, enabling image generation in four steps, saving time and computational effort.
并未美化和添加NSFW内容,方便大家自行添加所需的lora配合使用。distilled依然有效,可以4步出图,节省大家的时间和算力。
Most of the images in the showcase use my enhanced version of LORA, which hasn't been published on Civitai yet. Maybe I'll publish it when I'm in a better mood.
展示图大部分图片使用我的美化lora,现在并未在civitai发布。也许等我那天心情好会发布吧。
Reproduction and merging are permitted, subject to the Flux model license. Users assume all legal responsibility. Please indicate the source and leave a message when reproducing.
可以转载,请标明出处和给我留言。不要修改我的模型名称,不要说是自己的。你这样偷别人的模型能吃上四个菜吗!-说的就是你!From Lib @温柔情怀


