Internet Celebrity MAKER/网红制作机
5.0
1 reviewsDescription
基于sdxl的pony模型和flux模型的强大能力搭建了这套工作流用于生成更加真实的人物
增加了photomaker 用于更加贴近目标//如果不想使用pm我也放了无pm版本在夸克里
在保证人物美感的同时生成结果细节拉满,放大甚至可以看见皮肤细节
工作流采用的推理为florence 但joy-captain效果也许会更好
总得来说大致就是搭好以后放图就可以生成了 显存不大的兄弟可以把unet 换成fp8的flux模型
所有东西应该都打包好放在夸克了有需要自取,这里要特别感谢TTP大佬的放大插件,没有环境需求,速度快效果好,
Based on the powerful capabilities of sdxl's pony model and flux model, this workflow is built to generate more realistic characters.
Photomaker is added to get closer to the target. if u donnot want to use pm plz remove it by yourself
The details of the generated results are full while ensuring the beauty of the characters. You can even see the skin details when zooming in. The reasoning used in the workflow is florence, but joy-captain may have better results.
In general, it is roughly built and then the image can be generated. Brothers with small video memory can replace unet with fp8's flux model. Everything should be packaged and put in Quark. If you need it, pick it up yourself. Here I would like to thank TTP for the zoom plug-in, which has no environmental requirements, fast speed and good results.
我用夸克网盘分享了「xhs屠版流」,点击链接即可保存。打开「夸克APP」,无需下载在线播放视频,畅享原画5倍速,支持电视投屏。
链接:https://pan.quark.cn/s/64075da8ff81
Discussion
(No comments yet)
Loading...
Reviews
No reviews yet
Versions (1)
- latest (a year ago)
Node Details
Primitive Nodes (28)
Anything Everywhere (1)
DownloadAndLoadFlorence2Model (2)
Florence2Run (2)
FluxGuidance (2)
Image Comparer (rgthree) (1)
Image Inset Crop (rgthree) (1)
ModelSamplingFlux (1)
Note (4)
PhotoMakerInsightFaceLoader (1)
PhotoMakerLoaderPlus (1)
PrimitiveNode (1)
Reroute (6)
TTP_CoordinateSplitter (1)
TTP_Image_Tile_Batch (1)
TTP_Tile_image_size (1)
TTP_condsetarea_merge (1)
TTP_condtobatch (1)
Custom Nodes (85)
- CR Prompt Text (1)
- CR SDXL Aspect Ratio (1)
ComfyUI
- PreviewImage (2)
- MaskToImage (1)
- VAEEncode (3)
- ControlNetLoader (1)
- KSampler (2)
- ImageScaleToTotalPixels (1)
- CLIPTextEncode (3)
- SamplerCustomAdvanced (3)
- BasicGuider (2)
- DualCLIPLoader (1)
- ConditioningAverage (1)
- ImageSharpen (1)
- ImageScaleBy (1)
- RandomNoise (2)
- KSamplerSelect (2)
- LoraLoaderModelOnly (1)
- UNETLoader (1)
- SplitSigmasDenoise (1)
- DisableNoise (1)
- VAELoader (2)
- ControlNetApplyAdvanced (1)
- LoadImage (4)
- ImageBatch (2)
- KSamplerAdvanced (2)
- CLIPSetLastLayer (1)
- CheckpointLoaderSimple (1)
- VAEDecode (4)
- SaveImage (5)
- LatentUpscaleBy (1)
- BasicScheduler (2)
- VAEDecodeTiled (1)
- ImpactImageBatchToImageList (1)
- UltralyticsDetectorProvider (1)
- SAMLoader (1)
- FaceDetailer (1)
- MeshGraphormer-DepthMapPreprocessor (1)
- PhotoMakerStyles (1)
- PhotoMakerEncodePlus (1)
- IPAdapterUnifiedLoaderFaceID (1)
- IPAdapterFaceID (1)
- IPAdapterNoise (1)
- PrepImageForClipVision (4)
- TextCombinerTwo (2)
- TextInput_ (2)
- Get Image Size (1)
- ShowText|pysssss (2)
- smZ Settings (1)
- smZ CLIPTextEncode (2)
- Seed Everywhere (1)
- Constant Number (1)
- Text Concatenate (1)
Model Details
Checkpoints (1)
sdxl2\real\whitePonyDiffusion4_whitepony4lite.safetensors
LoRAs (1)
FLUX\F-real\墨幽-Flux-Lora-网图_v1.safetensors