
Fast and beautiful LCM model
No LCM LoRA is required
No negative prompt is required
Configuration for best results:
Sampler: LCM
Steps: 9-18
CFG Scale: 1-2
WxH: 960x960
Recipe:
1. [85%] p-AI-nter_v0.2* + [15%] protogen_v2.2 = p-AI-nter_v0.3*
2. [25%] p-AI-nter_v0.3 + [75%] (mira-v1.1* [60%] + menheraV2Experimental [40%])
3. [50%] ([15%] <2> + [85%] mira-v3.0--4500steps*) + [50%] ([25%] <2> + [75%] mira-v3.0--5000steps*)
4. [60%] <3> + [40%] photon_v1
5. <4> <- LCM_LoRA_Weights_SD15:1.0 <- ip-adapter-faceid_sd15_lora:0.666
NOTE: *"p-AI-nter_v0.2/0.3"(public model) and "mira"(private models) - trained by myself
p-AI-nter_v0.2 and p-AI-nter_v0.3 you can find on HF https://huggingface.co/Gourieff
p-AI-nter models are rather old but I like them and had been using them for some time
"mira" models are trained on photos of my good friend (with her consent), she's goddess and that's why these models can produce really beautiful female faces and give a good beautification boost for this entire merge of models
描述:
A bit more details compared to 1.0 version
训练词语:
名称: artifidaByGourieff_v11LCM.safetensors
大小 (KB): 2082643
类型: Model
Pickle 扫描结果: Success
Pickle 扫描信息: No Pickle imports
病毒扫描结果: Success