rh-zimage-turbo-v2-lora
Description
| Field | Content |
|---|---|
| Model Name | rh-zimage-turbo-v2-lora |
| Model Type | LoRA (text-to-image) |
| Platforms | ComfyUI / RunningHub / Hugging Face |
| Author | RunningHub-@Hao |
| Original | https://www.runninghub.cn/model/public/2027184765341798401 |
| Try online | https://www.runninghub.ai |
| https://www.runninghub.cn/runninghub-api-doc-en/ |
About this model
Z-image_turbo_无滤真实_V2
本lora为写实风格，非网红脸，更偏日常耐看素人感。尽力还原现实拍摄，弱化摄影打光效果，还原现实无滤镜实拍，未p图效果。避免原版摄影光影太强，给作品后期自由调色、磨皮程度、修图留出空间。
很大程度上去除人物一些无法描述的ai味，并且优化了原生Zimage女性都是小胸的问题，作为视频首帧图有不错效果~
测试图均为网图反推生成，已尽可能测试包含【多人群像、日常拍摄、现实景点、专业棚拍、舞台歌舞、cos漫展、婚纱汉服】等各种场景和风格，封面均为本lora在0.6~0.9权重下直出图片。
可搭配部分脸模使用，自行测试。也可出图后自行高清+换脸。
权重范围：
权重0.7-0.9，皮肤纹理、真实感越强烈、女性胸围越大（男性不影响）
权重0.6时，z的美颜摄影和现实感接近平衡。
权重0.5及以下，风格靠近Zimage原始的美颜摄影风格，略微保留现实感。
步数：10-15
采样器：euler a、res系列等均可
调度器：simple、sgm_uniform、beta、FlowMatch等均可，自行搭配测试
分辨率：建议1960，小等于1536肢体容易出现bug
风格词可不写，玩得愉快，多多包含~
This repository provides rh-zimage-turbo-v2-lora weight files. You can load them on RunningHub.
Finetuned from: Z-image-turbo
Files
| File | Size | Purpose |
|---|---|---|
无滤真实_V2.safetensors |
162 MiB | LoRA weights |
License
Published by RunningHub on behalf of the author. Copyright remains with the author. Follow the original project or upstream license.
Training at RunningHub
Train models on RunningHub: https://www.runninghub.ai/page-model
Run Seedance 2.5 with RunningHub API: https://www.runninghub.ai/call-api/api-detail/2133100000000700025
- RunningHub：https://www.runninghub.ai
- RunningHub 中国站：https://www.runninghub.cn
- RunningHub API 文档：https://www.runninghub.cn/runninghub-api-doc-cn/
- RunningHub API Documentation：https://www.runninghub.cn/runninghub-api-doc-en/
- Original project: https://www.runninghub.cn/model/public/2027184765341798401
- Author page: https://www.runninghub.cn/user-center/1900423919516700673