Update README.md

README.md

@@ -12,7 +12,7 @@ inference: true
 ## Example: 
 ”Negative prompt: (worst quality, low quality:1.4)” is really useful in anywhere
 ## Ekmix-Pastel
-pastel but lines (Merging the loras into the model.)
+pastel but lines(hires fix) (Merging the loras into the model.)
 ~~~
 python networks\merge_lora.py --sd_model .\models\model.safetensors --save_to .\lora\2.safetensors --models  .\lora\MagicLORA.pt  .\lora\Jordan_3.safetensors  .\lora\sttabi_v1.4-04.safetensors  .\lora\xlimo768.pt  .\lora\dpep2.pt  --ratios 0.3 1 0.5 0.6 0.35
 ~~~
@@ -22,6 +22,8 @@ masterpiece,best quality,best quality,Amazing,beautiful detailed eyes,1girl,fine
 Negative prompt: (worst quality, low quality:1.3)
 Steps: 30, Sampler: Euler a, CFG scale: 6, Seed: 191289851, Size: 512x768, Model hash: 0526445f65, Denoising strength: 0.5, Eta: 0.5, Clip skip: 2, ENSD: 31337, Hires resize: 856x1280, Hires steps: 30, Hires upscaler: Latent
 ~~~
+pastel but lines(without hires fix) (better!)
+![]()
 ## Ekmix-gen4
 balance between anime and reality(Merging by block weighted merge.)
 ![0](https://huggingface.co/EK12317/Ekmix-Diffusion/resolve/main/examples/xy_grid-0273-620659051-masterpiece%2Cbest%20quality%2Cbest%20quality%2CAmazing%2Cbeautiful%20detailed%20eyes%2C1girl%2Cfinely%20detail%2CDepth%20offield%2Cextremely%20detailed%20CG%20un.png)