Resource directory
Find what you need for your workflow.
Guides: techniques, setup, and troubleshootingSearch Sweet Tea Pipes, external workflows, models, node packs, and guides in one evidence-linked catalog.
Start with a workflow instead ↗Search results
All resources
18 shown of 30,993
Wan2.2-Fun-Camera-LowNoise-14B-Q5_K_S.gguf
Wan2.1-I2V-14B-480p-Q6_K.gguf
Qwen3.5 2b LoRA Checkpoints
Anima_Style (data03-v1.0)
Version based on data02 with additional partial 2.5D elements. Tends toward a fusion of anime and semi-realistic styles, with stronger sense of volume in characters and natural lighting/shadows.
DonutsDeliveryMix V4 - NSFW Photorealism
Also try my new checkpoint:
Change-A-Character: Naughty Schoolgirl, You Waifu Has Been a Bad Girl!
What does it do you ask?
artist:murakami suigun
更新 一口气更新两个版本,一个是之前说的老画风,一个是新老画风在同一个lora里可以分开调用的多画风,老画风的启动词是@murakami suigun \(2000s\),而新版本是@murakami suigun。 另,我觉得双画风在一个Lora里会很炫酷,你们不觉得很炫酷吗。 以防你们不信以及不会翻到我证明一个丹两个画风,我把对比图放这 anima版本更新 其实anima会这个画师串,但是因为这个画师的画作时间跨度太长,所以取了比较新的时间的训练,这个Lora可以很好稳定在比较近期的风格,比画师串观感好的多,未来我会再炼一个怀旧画风的分支 V3更新 更好的还原度,更好的肢体,更好的兼容性,甚至更好的训练集 权重1.0 又新又好,啊不行我不能嘴角上翘。 V2更新 没什么好说的,全面的比1.0要好,建议直接换成V2.0。画风更像的同时还没有了以前老画风的污染情况,也不用将画师tag调成0.6-0.8了,并且兼容的模型我测下来更多了。 murakami suigun 的画师风格lora 权重建议0.6-0.8,注意,你的murakami suigun这个TAG也要0.6-0.8.原因是noob本身有这个画师TAG,但是学的不像又过拟合,这个lora主要解决tag不像的问题。 不使用启动词时是另一种风格
Wan Video 2.2
Anime Foot Repair for Anima
I trained this LORA with the sole purpose of making the depiction of feet more appealing while maintaining the original style.
幼稚園児コスプレ
幼稚園児のコスプレ衣装です。 安定しないので、まだまだ改良が必要です。 服はいい感じですが、帽子やカバンが微妙。(最初は服も全然ダメでした。)
Voxcpm Phase3
z_image_bf16.safetensors
Wet Clothing (LTX-2.3)
Sylvanas Windrunner (World of Warcraft) - Cinematic Version
Krea 2 version might be a bit rough around the edges, but hey! I'm still learning. Her armor is a bit hard to get rid of sometimes.
Noob SDXL ControlNet Softedge Hed
fair-ai-public-license-1.0-sd
Yufine (Epic Seven)
Yufine from Epic Seven.
Lewm Billiards
> Author: Santosh Jaiswal (@hellojais) > Base architecture: LeWM by Lucas Maes et al. (2025) > Training data: hellojais/billiards-worldmodel > Code: hellojais/le-wm
