超高还原度图片仿写 Image replication
5.0
0 reviews
Description
你觉得那张是真图那张又是改写的？
这个爆款仿写工作流已经做到了真正分辨不出来的程度。
这个工作流不像可图ipadapter风格迁移那样，他出图质量特别取决于参考图的质量，参考图需要高清些
混合了ACE lora模型和redux模型使用
redux默认是1，强度上调可以用提示词控制图像变化
running hub已上传应用可以直接体验
https://www.runninghub.cn/ai-detail/1905284747202072578
---------------------------------------------
Which one do you think is the original picture and which one is the rewritten version?
This popular imitation workflow has reached a level where it's truly impossible to tell the difference.
The default value of Redux is 1. You can control the changes in the image through prompt words when increasing its intensity.
The application has been uploaded to Running Hub, and you can directly experience it.
Discussion
(No comments yet)
Loading...
Reviews
No reviews yet
Versions (2)
- latest (9 months ago)
- v20250310-091951
Node Details
Primitive Nodes (19)
Anything Everywhere3 (1)
CLIPTextEncodeFlux (1)
Fast Groups Bypasser (rgthree) (1)
Florence2ModelLoader (1)
Florence2Run (1)
FluxGuidance (2)
Image Comparer (rgthree) (1)
ImageConcatMulti (1)
LongCLIPTextEncodeFlux (1)
ModelSamplingFlux (1)
Note (1)
PrimitiveNode (1)
ReduxAdvanced (1)
Reroute (1)
TTP_Image_Assy (1)
TTP_Image_Tile_Batch (1)
TTP_Tile_image_size (1)
easy imageBatchToImageList (1)
Custom Nodes (68)
- CR Latent Batch Size (1)
ComfyUI
- SamplerCustomAdvanced (3)
- BasicScheduler (2)
- VAEDecodeTiled (1)
- UpscaleModelLoader (1)
- LoadImage (1)
- SolidMask (4)
- MaskToImage (4)
- LoraLoaderModelOnly (3)
- UNETLoader (2)
- DualCLIPLoader (1)
- VAELoader (1)
- StyleModelLoader (1)
- CLIPVisionLoader (1)
- BasicGuider (3)
- VAEDecode (2)
- DisableNoise (1)
- RandomNoise (2)
- KSamplerSelect (2)
- SplitSigmasDenoise (1)
- VAEEncode (2)
- ImageUpscaleWithModel (1)
- ImageScale (1)
- ImageScaleToTotalPixels (1)
- KSampler (1)
- PreviewImage (5)
- InpaintModelConditioning (1)
- CLIPTextEncode (1)
- SaveImage (4)
- ConditioningZeroOut (1)
- ImageToMask (2)
- easy cleanGpuUsed (2)
- easy imageConcat (3)
- MaskPreview+ (1)
- GetImageSize+ (1)
- ImpactInt (2)
- LayerUtility: CropByMask (1)
- LayerUtility: ImageScaleByAspectRatio V2 (1)
Model Details
Checkpoints (0)
LoRAs (3)
ace\comfyui_subject_lora16.safetensors
f.1摄影模拟\Kodak Vision3 500T v2.safetensors
flux\flux_realism_lora.safetensors