Model

(日本語説明は後半にあります)What's new in v4.0Replaced to new VAE : Fixed getting redder when using HiRes-Fix, img2img or ADeteiler.Revised face, eye, composition and painting....

French fashion, often referred to as "la mode française," is renowned worldwide for its elegance, sophistication, and timeless style. 描述: ...

A concept LoRA for viewing from outside window.Suggested weights: 0.5~2.5 depending on your use case.The examples are using clip_skip=2 and highres fix ESRGAN_4x with denoising...

Johnny_Numbnutz Style LoRA!Making models can be time consuming and expensive. If you like what I do consider buying me a borgar?!This artstyle is composed of bits of the hellta...

TW is shogunpuppetRecommended weight of 0.8-1Not gonna lie, this is a picky lora that can do weird things depending on the model and style lora you use. It works best with anim...

Fix File Version 4I was informed by a user that both the Copax Timeless V3 and Copax 2D Anime models have the same hash.Thank you FractalButterfly for your feedback.After check...

更新了玉雕，一样很难用，自己斟酌Updated the jade carving, which is also difficult to use, so consider it yourself————————————————————————————warning!warning!warning!注意！注意...

Common Positive Prompt/共通ポジティブプロンプトmasterpiece,best_qualityCommon Negative Prompt/共通ネガティブプロンプトlow quality, worst quality　　2024.9.13　V3.14th_tail0.5.0...

The model used more than 500 high-quality original pictures for training.The4.0 version of LORA can better restore the hand drawn watercolor style of the artist.Recommendation ...

This is a LoRA to generate Akari from Pokemon Legends: ArceusI suggest using this LoRA at 0.8-1.0 weight for best results.These images were generated and only tested using AOM2...

I have set up a Discord-Server where we can share images /ideas /prompts and help each other out. Also, if you have a LoRA request, feel free to ask me there.Creating LoRA take...

I have prepared a new generation instructions, you just need to follow the diagram, adjust the weight, model, attention vocabulary writing, the rest is default, to get good res...

描述: 训练词语: 名称: RMSDXL_Creative.safetensors ...

更新了更新了 v3 v3 v3!!! Now Human Form is enabled~ Lora trained based on 80+ 1024x1024 Chartreux (British Shorthair/藍貓) Photos.For best results, you may follow the settingSa...

carol oslton anime tomo-chan work with 0.8 weight enjoy it,Lora xl with dpo technology experimental natural language:Experiment with the scores to get the desired details with ...

☕ Donations to my Ko-Fi are appreciatedNo way to generate? try out the model HERE!! ---NOTESTrigger: jolyneAdditional prompts for costume, add/remove as you needed: main outfit...

A LoRa model trained with dropout & block weights.This is a LoRa model that trained by 100+ img generated with NijiV5,and it's the noise version.This model can generate a b...

我发现ai想要画好表情，通常需要好多tag一起使用所以我训练了这种lora便于简单的画出表情数据集不是很好找，所以会略微影响画风但是0.4-0.6的权重就能生效（这样不会影响画风）如果你喜欢请...

These girls next door have grown up. Keep moving on.~~It's confusing again and again that after I set training images larger to 576*768, I thought I would get better results, b...

诛仙动漫版本的陆雪琪 应该可以生成全身，可以搭配面部lora模型使用改善颜值 描述: 1024,1024 训练词语: luxueqi ...