Portrait Master Model Comparer
5.0
0 reviewsDescription
Uses the Portrait Master nodes to create the prompt and that feeds into 4 different models; Kwai-Kolors, Flux, SD and SD3.5.
For Kolors, personally I like the Open Kolors model as I find the results look more natural compared to the base Kolors model. Â https://civitai.com/models/602580/kolors-openkolors-v24-multiple-style-general-kolors-model?modelVersionId=965529
The group also uses Perturbed Attention Guidance and Auto CFG nodes.
The Flux group uses the Flux Sampler Paramenters node and has an optional group for Plot Sampler Parameters. If you've not come across this before, it allows you to generate a batch with different settings like samplers, schedulers, guidance, shift etc. to be able to compare and find which works best. For more info on these see this video: Â https://youtu.be/WtmKyqi_aFM?si=X-UEp67BwjFWjT0g
The SD group uses PAG, Auto CFG and AYS.
The final group is setup for the new SD3.5-Large model.
Small update, meant to add Power LoRA loader node to the SD3.5 groups as LoRAs are starting to appear but forgot, so added that.
Discussion
(No comments yet)
Loading...
Reviews
No reviews yet
Versions (3)
- latest (a year ago)
- v20241028-011658
- v20241027-203724
Node Details
Primitive Nodes (27)
DF_Text_Box (4)
EmptySD3LatentImage (1)
Fast Groups Muter (rgthree) (1)
FluxResolutionNode (1)
FluxSamplerParams+ (1)
MZ_ChatGLM3Loader (1)
MZ_ChatGLM3_V2 (2)
MZ_KolorsUNETLoaderV2 (1)
ModelSamplingSD3 (1)
OverrideCLIPDevice (2)
PlotParameters+ (1)
PortraitMasterBaseCharacter (1)
PortraitMasterMakeup (1)
PortraitMasterSkinDetails (1)
PortraitMasterStylePose (1)
Power Lora Loader (rgthree) (3)
SD3NegativeConditioning+ (1)
SamplerSelectHelper+ (1)
SchedulerSelectHelper+ (1)
TripleCLIPLoader (1)
Custom Nodes (47)
- CR Text Concatenate (4)
ComfyUI
- VAEDecode (4)
- PerturbedAttentionGuidance (2)
- VAELoader (2)
- LoraLoaderModelOnly (1)
- DualCLIPLoader (1)
- RandomNoise (1)
- SaveImage (4)
- CheckpointLoaderSimple (2)
- UNETLoader (1)
- CLIPAttentionMultiply (1)
- EmptyLatentImage (3)
- KSamplerSelect (1)
- SamplerCustom (1)
- PreviewImage (4)
- KSampler (2)
- CLIPTextEncode (5)
- AlignYourStepsScheduler (1)
- Automatic CFG (2)
- ShowText|pysssss (2)
SDXLResolutionPresets
- SDXLResolutionPresets (3)
Model Details
Checkpoints (2)
01. SDXL\juggernautXL_juggXIByRundiffusion.safetensors
03. SD3.5\sd3.5_large.safetensors
LoRAs (1)
03. Kolors\kolors-flux-2000.safetensors