Here are a couple simple workflows I made by requests which use PuLID to create portrait drawings from input images.

The checkpoint I have used is Juggernaut XL Version XI https://civarchive.com/models/133005/juggernaut-xl

The lora I have used is Drawing SDXL

https://civarchive.com/models/516701/drawing-sdxl-inspired-by-drawholic

The far superior Pulid version 1.1 for SDXL doesn't work in standard PuLID_ComfyUI as of this writing. There is a pull request to add support from this branch here:
https://github.com/vitrun/PuLID_ComfyUI/tree/feature/v1.1

To install the branch with the Pulid v1.1 patch use the following command in your ComfyUI custom nodes folder:
git clone -b feature/v1.1 https://github.com/vitrun/PuLID_ComfyUI

This v1.1 model goes in your models/pulid folder

https://huggingface.co/Runzy/ip-adapter_pulid/blob/main/ip-adapter_pulidv1.1_sdxl_fp16.safetensors

Remember to follow the other install instructions from the bottom of the PuLID github page. You will need the facexlib dependency as well as insightface's Antelope V2.

The more complicated version of this workflow resizes the input image to 1024x1024 then uses Grounding Dino to separate the subject from the background, then creates a depthmap so that controlnet may cause the output to more closely match the input image. All other nodes should be easily downloaded from node manager: KJNodes, WAS Node Suite, Controlnet_Aux, Segment_Anything