The principle is similar to that of wan2.2:
Image aesthetics generated by SPRO are complemented by wan2.2 low-noise model
SRPO: https://huggingface.co/tencent/SRPO/tree/main
fp8: https://huggingface.co/wikeeyang/SRPO-Refine-Quantized-v1.0/tree/main
flux+SRPO+Wan2.2=Extreme aesthetics and detail
Workflow Preview
Showcases (Image/Video)
Showcases (Image/Video)
Nodes Detail
42 NodesPrimitive Node Types
16- UpscaleModelLoader1
- VAEDecode2
- PreviewImage2
- VAELoader2
- CLIPLoader1
- SaveImage2
- CLIPTextEncode3
- VAEEncode1
- KSampler1
- LoadImage1
Custom Node Types
26- SamplerCustomAdvanced1
- FluxGuidance1
- BasicGuider1
- ModelSamplingSD31
- ImageStitch1
- UNETLoader2
- DualCLIPLoader1
- EmptySD3LatentImage1
- RandomNoise1
- PrimitiveNode2
- BasicScheduler1
- ModelSamplingFlux1
- KSamplerSelect1
- easy cleanGpuUsed1
- ImageScaleToMegapixels2
- ImageSharpen1
- LoraLoaderModelOnly1
- RH_Captioner1
- BetterFilmGrain1
- MarkdownNote1
- ShowText|pysssss1
- RH_Translator1
- Text1