
I wanted tor make a checkpoint to work with my original SD1.5 model.
I like to make images that are semirealistic, with some ability to go photorealistic if needed. There are other models out there, so I thank you for taking a moment to look at my merge.
This merge uses a 50/50 mix between (RealCartoon-XL, thanks to 7whitefire7 for your hard work!) and (SereneXL, thanks to hatchetkilla80328 ). I then bleneded in the WowifierXL LoRA at 0.60 for a little extra Ooomp! (Thanks to mooncryptocow!) I baked in vae-ft-mse-840000-ema, so there should be no need to roll in a seperate VAE (unless you want to)
Usage:
The Sample Images shown have the following settings:
VAE: vae-ft-mse-840000-ema
Clip Skip: 2
Aspect: 768*1216 for portrait
The initial pass through ComfyUI is usually:
Steps: 20
Cfg: 6
Sampler: euler (normal)
The upscale is done at 1.5 times using 4x-Ultrasharp
Steps: 20
Cfg: 8
Sampler: euler (normal)
Denoise: 0.35
描述:
训练词语:
名称: pmamXLPhredMadeAnother_v10.safetensors
大小 (KB): 6776592
类型: Model
Pickle 扫描结果: Success
Pickle 扫描信息: No Pickle imports
病毒扫描结果: Success