The model that can generate images with simple prompts.
The following settings are recommended:
Sampling method: DPM++ 2M SDE
Schedule type: Karras
Sampling steps: 15
CFG Scale: 2.25
Negative prompts:
embedding:BeyondSDXLv3 embedding:unaestheticXL_Alb2
If you want more details
Sampling steps: 30
CFG Scale: 2.25~8.0
If you are using LoRA, the following settings are recommended:
Sampling method: Euler
Schedule type: Automatic(Exponential)
Sampling steps: 30
CFG Scale: 4
I merged the following two models
Thanks to the model providers
国风4 GuoFeng4 XL v1.2 56%
Atomix Anime XL v1.0 44%