Hi
This is a base SDXL text2image workflow but it has a Noise Invection in the end of generation process for making variable pictures (4 by default).
And thera are Face Detailer node and becouse this wf is good for making beautiful characters.
Open
Download
Nodes Information
Primitive Nodes (7)
CLIPSetLastLayer
CheckpointLoaderSimple
KSamplerAdvanced
RepeatLatentBatch
SaveImage
VAEDecode
VAELoader
Custom Nodes (12)
CLIPTextEncodeSDXL
CR Apply LoRA Stack
CR LoRA Stack
DPCombinatorialGenerator
FaceDetailer
FromBasicPipe_v2
PrimitiveNode
Reroute
SAMLoader
SDXL Empty Latent Image
ToBasicPipe
UltralyticsDetectorProvider